Social Psychology Issue 1, 2019 by Hogrefe

Volume 50 / Number 1 / 2019

Social Psychology

Editor-in-Chief Kai Epstude Associate Editors Anna Baumert Marco Brambilla Adam Fetterman Ilka Gleibs Michael Häfner Hans J. IJzerman Ulrich Kühnen Toon Kuppens Ruth Mayo Kim Peters

The latest knowledge on how to tackle the complexities of hoarding disorder “If you wish to help those who suffer with the debilitating problem of hoarding, get this book and learn from these experienced scientist–practitioners.” Michael A. Tompkins, PhD, ABPP, Co-Director, San Francisco Bay Area Center for Cognitive Therapy; Assistant Clinical Professor, University of California at Berkeley

Gregory S. Chasson / Jedidiah Siev

Hoarding Disorder (Series: Advances in Psychotherapy – Evidence-Based Practice – Volume 40) 2019, viii + 76 pp. US $29.80 / € 24.95 ISBN 978-0-88937-407-2 Also available as eBook Hoarding disorder, classified as one of the obsessive-compulsive and related disorders in the DSM-5, presents particular challenges in therapeutic work, including treatment ambivalence and lack of insight of those affected. This evidence-based guide written by leading experts presents the latest knowledge on assessment and treatment of hoarding disorder. The reader gains a thorough grounding in the treatment of choice for hoarding – a specific form of CBT interweaved with psy-

www.hogrefe.com

choeducational, motivational, and harm-reduction approaches to enhance treatment outcome. Rich anecdotes and clinical pearls illuminate the science, and the book also includes information for special client groups, such as older individuals and those who hoard animals. Printable handouts help busy practitioners. This book is essential reading for clinical psychologists, psychiatrists, psychotherapists, and practitioners who work with older populations, as well as students.

Social Psychology

Volume 50/Number 1/2019

Editor-in-Chief

Kai Epstude, University of Groningen, Department of Psychology, Grote Kruisstraat 2/1, 9712 TS Groningen,The Netherlands, Tel. +31 50 363-7632, Fax + 31 50 363-4581, E-mail k.epstude@rug.nl

Editorial Ofﬁce

Wim Meerholz, University of Groningen, Department of Psychology, Grote Kruisstraat 2/1, 9712 TS Groningen,The Netherlands, Tel. +31 50 363-6393, Fax + 31 50 363-4581, E-mail SocialPsych.EditorialOfﬁce@gmail.com MichaelHäfner,UniversitätderKünsteBerlin,Germany Anna Baumert, Max-Planck Institute for Hans IJzerman, Université Grenoble Alpes, France Collective Goods, Germany Ulrich Kühnen, Jacobs University, Germany Marco Brambilla, University of Milano-Bicocca, Toon Kuppens, University of Groningen, Italy The Netherlands Adam Fetterman, University of Texas, USA Ruth Mayo, Hebrew University of Jerusalem, Israel Ilka Gleibs, London School of Economics, Kim Peters, University of Queensland, Australia United Kingdom

Associate Editors

Consulting Editors

Susanne Abele (Oxford, OH, USA) Andrea Abele-Brehm (Erlangen-Nürnberg, Germany) Anja Achtziger (Friedrichshafen, Germany) Herbert Bless (Mannheim, Germany) Gerd Bohner (Bielefeld, Germany) Oliver Christ (Hagen, Germany) Paul Conway (Tallahassee, FL, USA) Katja Corcoran (Graz, Austria) Olivier Corneille (Louvain, Belgium) Amanda Diekman (Oxford, OH, USA) Andrew Elliott (Rochester, NY, USA) Birte Englich (Cologne, Germany) Bertram Gawronski (Austin, TX, USA) Guido Gendolla (Geneva, Switzerland) Jessica Good (Davidson, NC, USA) Tobias Greitemeyer (Innsbruck, Austria) Bettina Hannover (Berlin, Germany) Nina Hansen (Groningen, The Netherlands) Nicole Harth (Jena, Germany) S. Alexander Haslam (Brisbane, Australia) Ying-Yi Hong (Hong Kong, ROC) Roland Imhoff (Mainz, Germany) Eva Jonas (Salzburg, Austria) Franciska Krings (Lausanne, Switzerland)

Publisher

Hogrefe Publishing, Merkelstr. 3, 37085 Göttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-111, E-mail publishing@hogrefe.com North America: Hogrefe Publishing, 7 Bulﬁnch Place, Suite 202, Boston, MA 02114, USA, Tel. (866) 823-4726, Fax (617) 354-6875, E-mail publishing@hogrefe.com

Production

Regina Pinks-Freybott, Hogrefe Publishing, Merkelstr. 3, 37085 Göttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-111, E-mail production@hogrefe.com

Subscriptions

Hogrefe Publishing, Herbert-Quandt-Str. 4, D-37081 Göttingen, Germany, Tel. +49 551 99950-900, Fax +49 551 90050-998

Advertising/Inserts

Melanie Beck, Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, Tel. +49 551 99950-423, Fax +49 551 99950-111, E-mail marketing@hogrefe.com

ISSN

ISSN-L 1864-9335, ISSN-Print 1864-9335, ISSN-Online 2151-2590

Ó 2019 Hogrefe Publishing. This journal as well as the individual contributions and illustrations contained within it are protected under international copyright law. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microﬁlming, recording or otherwise, without prior written permission from the publisher. All rights, including translation rights, reserved.

Publication

Published in 6 issues per annual volume. Social Psychology is the continuation of Zeitschrift für Sozialpsychologie (ISSN 0044-3514), the last annual volume of which (Volume 38) was published in 2007.

Subscription Prices

Calendar year subscriptions only. Rates for 2019: Institutions – from US $454.00/ 348.00 (print only; pricing for online access can be found in the journals catalog at hgf.io/journals2019); Individuals – US $223.00/ 159.00 (print & online). Postage and handling – US $24.00/ 18.00. Single copies – US $79.50/ 62.50 + postage and handling.

Payment

Payment may be made by check, international money order, or credit card, to Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany. US and Canadian subscriptions can also be ordered from Hogrefe Publishing, 7 Bulﬁnch Place, Suite 202, Boston, MA, 02114, USA.

Daniel Lakens (Eindhoven, The Netherlands) Thomas Morton (Exeter, UK) Corinne Moss-Racusin (Saratoga Springs, NY, USA) Roland Neumann (Trier, Germany) Esther Papies (Glasgow, UK) Susanne Quadﬂieg (Bristol, UK) Marc-André Reinhard (Kassel, Germany) Toni Schmader (Vancouver, BC, Canada) Manfred Schmitt (Landau, Germany) Thomas W. Schubert (Oslo, Norway) Beate Seibt (Oslo, Norway) Frank Siebler (Tromsø, Norway) Monika Sieverding (Heidelberg, Germany) Dagmar Stahlberg (Mannheim, Germany) Fritz Strack (Würzburg, Germany) Rolf van Dick (Frankfurt/Main, Germany) Harm Veling (Nijmegen, The Netherlands) Tobias Vogel (Mannheim, Germany) Ulrich Wagner (Marburg, Germany) Eva Walther (Trier, Germany) Michaela Wänke (Mannheim, Germany) Michael Wohl (Ottawa, ON, Canada) Bogdan Wojciszke (Warsaw, Poland) Vincent Yzerbyt (Louvain-la-Neuve, Belgium)

Electronic Full Text

The full text of Social Psychology is available online at http://econtent.hogrefe.com and in PsycARTICLES.

Abstracting Services

Abstracted/indexed in Current Contents/Social and Behavioral Sciences (CC/S&BS), Social Sciences Citation Index (SSCI), PsycINFO, PASCAL, PSYNDEX, ERIH, Scopus, and EM Care. 2017 Impact Factor 1.200, 5-year Impact Factor 2.157, Journal Citation Reports (Clarivate Analytics, 2018)

Social Psychology (2019), 50(1)

Ó 2019 Hogrefe Publishing

Contents Focal Article

A Reﬂection on Crucial Periods in 50 Years of Social Psychology (in Germany) Katja Corcoran, Michael Häfner, Mathias Kauff, and Stefan Stürmer

Original Articles

Intended, Unintended, and Unknown Consequences of Contact: The Role of Positive-Negative Contact on Outgroup Attitudes, Collective Action Tendencies, and Psychological Well-Being Sabahat C. Bagci and Abbas Turnuklu

Ó 2019 Hogrefe Publishing

Meta-Accuracy and Perceived Reciprocity From the Perception-Meta-Perception Social Relations Model Ben Porter, Camilla S. Øverup, Julie A. Brunson, and Paras D. Mehta

You Can Leave Your Glasses on: Glasses Can Increase Electoral Success Alexandra Fleischmann, Joris Lammers, Janka I. Stoker, and Harry Garretsen

Attribution of Egoistic Versus Altruistic Motives to Acts of Helping: The Role of the Helper’s Status and the Act’s Intended Publicity Birte Siem and Stefan Stürmer

Social Psychology (2019), 50(1)

Focal Article

A Reflection on Crucial Periods in 50 Years of Social Psychology (in Germany) Katja Corcoran1,2, Michael Häfner3, Mathias Kauff4, and Stefan Stürmer4 1

Institute of Psychology, Karl-Franzens Universität Graz, Austria

BioTechMed-Graz, Austria

Universität der Künste Berlin, Germany

Fakultät für Psychologie, FernUniversität in Hagen, Germany

Abstract: In this article, we reflect on 50 years of the journal Social Psychology. We interviewed colleagues who have witnessed the history of the journal. Based on these interviews, we identified three crucial periods in Social Psychology’s history, that are (a) the early development and further professionalization of the journal, (b) the reunification of East and West Germany, and (c) the internationalization of the journal and its transformation from the Zeitschrift für Sozialpsychologie to Social Psychology. We end our reflection with a discussion of changes that occurred during these periods and their implication for the future of our field. Keywords: History of Social Psychology, Germany, Zeitschrift für Sozialpsychologie, Fachgruppe Sozialpsychologie

In case you haven’t noticed, you are currently reading the 50th volume of Social Psychology! We are convinced that this jubilee and important milestone in the history of the journal should be valued appropriately. Accordingly, we are grateful to the Editor-in-Chief, Kai Epstude, for giving us the opportunity to pause for a moment and regard the past decades of Social Psychology. We will reflect on the transformations the journal has gone through and present our view on what the social psychology community can learn from these crucial changes. Why us? Social Psychology was published in German as Zeitschrift für Sozialpsychologie until 2007. In the editorial to the journal’s first issue, the founding editors, Hubert Feger, Carl-Friedrich Graumann, Klaus Holzkamp, and Martin Irle – leading German social psychologists at the time – stated that the Zeitschrift für Sozialpsychologie was dedicated to pursuing two main goals: “First, it will be – for German-speaking countries – a unique forum for the rapidly increasing amount of social psychological research” (Feger, Graumann, Holzkamp, & Irle, 1970, p. 1). The editors emphasized a broad definition of social psychology, including work in the areas of social psychology, language psychology, and political psychology, which at the time was dispersed across different and more specialized journals. The Zeitschrift’s second main goal was to stimulate 1

and provoke critical reflections, scientific discussions, and controversies about social psychology’s antecedents, theories, methods, research questions, and objectives (Feger et al., 1970, p. 1). In the following years, the Zeitschrift für Sozialpsychologie became the flagship journal of Germanspeaking and psychologically oriented social psychologists (sociologically oriented social psychologists tended to publish in sociological journals). Further, after the Social Psychology Division of the German Society for Psychology (Fachgruppe Sozialpsychologie der Deutschen Gesellschaft für Psychologie) was founded in 1986, the Zeitschrift für Sozialpsychologie and the Fachgruppe became closely linked. In fact, there were times when it appeared as if the Zeitschrift was the official journal of the Fachgruppe (which it was not, even if the Fachgruppe has had a say, for instance, in proposing the editorial board). Due to these connections, we – the current members of the Social Psychology Division’s Council – feel obliged to take the opportunity to commemorate Social Psychology’s past and discuss potential implications for the future. For this purpose, we interviewed and consulted with colleagues who have witnessed critical historical periods, published on related topics, served as former editors-in-chief of Social Psychology, and/or influenced and still influence the development of this journal and of our field.1 What we discussed in our

We would like to thank all interview partners, namely Gerd Bohner, Stefan Busse, Kai Epstude, Wolfgang Frindte, Gerold Mikula, Amélie Mummendey, Lily Silny, Fritz Strack, and Wolfgang Stroebe.

Ó 2019 Hogrefe Publishing

Social Psychology (2019), 50(1), 1–6 https://doi.org/10.1027/1864-9335/a000372

K. Corcoran et al., Learning From History

interviews with these colleagues can be condensed into one simple question: What can we learn from the history of Social Psychology? Based on our understanding of the history of Social Psychology as well as on the results of our interviews, we identified three critical historical periods, all of which had some implications for the journal. The first significant period was the professionalization of West German social psychology as an academic discipline from the 1960s to the late 1980s. During this period, the Zeitschrift für Sozialpsychologie played a leading role in defining the field. The second period we focus on is the reunification of East and West Germany. As a consequence of World War II and the subsequent Cold War, in 1949 two politically and geographically separated German states were created: The Federal Republic of Germany (“West Germany”) and the socialist German Democratic Republic (“East Germany”). The differing realities in both German states led to the development of distinct academic traditions in East German and West German social psychology. We will reflect upon the impact of the reunification of East and West Germany on the development of German social psychology and the journal’s role in this context. Finally, we address the period of the internationalization of German social psychology, in which the German Zeitschrift für Sozialpsychologie became the English-language and international outlet Social Psychology.

Early Development and Further Professionalization In order to understand why we labeled the first big step in the development of Social Psychology in West Germany “further professionalization,” it is worth going back a bit in time. After two wars, not only science in Germany but the country itself were struggling to find new ways. The Nazi ideology had to be overcome, new structures (also at the universities) had to be built, and answers to new questions had to be found. Flight from the old regime, new family configurations, and the upcoming mass media were important issues that were picked up by West German social psychologists at this time (see also Lück, 2002). These topics and the inward looking, problem-centered focus led, however, to a relatively strong diversification in social psychology. There was nothing like the German social psychology, but there were many different social psychologies centering around the problems they aimed to solve. This diversity and the focus on concrete problems may, however, have eclipsed the view on international developments in the field. As a consequence, postwar German 2

social psychology was not as well connected to the leading Anglo-American social psychology as close neighbors were, for instance, the Netherlands. It was not until the 1970s that German social psychology began to prosper, which manifested itself, among others, in the foundation of its own journal, the Zeitschrift für Sozialpsychologie. Certainly, the creation of an own outlet did not only help to spread the word, but mainly helped to develop an own identity with West German social psychology. Despite this unifying effort, a look at the first editorial team reveals how diverse West German social psychology still was at the time. On the one hand, information theory and decision-making became increasingly dominant and had a large influence on Graumann’s and Irle’s research program (Sonderforschungsbereich 24). On the other hand, Holzkamp’s critical psychology was strongly influenced by Marxist ideas, mainly in a dialectic and hermeneutic tradition. Furthermore, there was a strong sociological tradition prevailing, for instance, around René König in Cologne. In light of this diversity, it is even more astonishing how fruitful the following one and a half decades of the Zeitschrift became. The editors did a great job in publishing good and important articles and in promoting a professional identity. Nevertheless, the Anglo-American (social) psychology was still steps ahead, and we are not only talking about topics or methodology but also about the professional standards shared within the discipline. Leading (Anglo-American) journals in the field relied on peer review and were backed up by large professional associations, such as the American Psychological Association (APA). In parallel to these developments, German social psychology and again, the identity carved by the journal certainly also helped opened up and connected more closely to the Anglo-American and European (social) psychology. The latter was also strengthened by the foundation of the European Association of Social Psychology (EASP)2 in 1966, with the explicit goal of tying together (East and West) European social psychology. In this phase of internationalization, it also became ever more common for young scholars to go abroad as doctoral or post-doctoral students. As a matter of fact, it was then a group of such young and international-minded scholars, many of them returning from the US, who in the mid-1980s sought to also professionalize German social psychology. They fought for a special division for social psychologists (the Fachgruppe Sozialpsychologie) within the German Psychological Society (DGPs) and linked the journal to this association, thereby creating a professional background. The journal was no longer an endeavor of a few leading figures, but an outlet for a new generation of German social psychologists. Moreover, by adhering to international professional standards

Back then, it was the European Association of Experimental Social Psychology.

Social Psychology (2019), 50(1), 1–6

Ó 2019 Hogrefe Publishing

K. Corcoran et al., Learning From History

such as peer review and changing editorial teams, more and more social psychologists from the German-speaking countries could identify with the journal. Much of what was then institutionalized can still be seen in the journal today, testimony to the importance of this first big step in the evolution of the journal and German social psychology.

West and East German Social Psychology and Germany’s Reunification Differing political, economic, and societal realities critically shaped the development of social psychology in West and East Germany. Whereas postwar West German social psychology developed through an increasing scientific exchange with social psychologists in North America and Western Europe, the situation in East Germany was different. To begin with, due to a stronger orientation of East German psychology toward the natural rather than the social sciences, there were far fewer social psychology positions at East than at West German universities (the universities of Jena and Leipzig were the two main centers of East German social psychology). Second, social psychological research in East Germany was embedded in the ideological framework of a socialist political system in which “psychological methodology and theory are directly and indirectly related to the axioms and principles of dialectic and historical materialism” (Schmidt, 1987, p. 232). In this system, social psychologists had to develop approaches to bridge and reconcile diverging scientific, societal, and political demands (see, for instance, the Jena program of a Marxist social psychology by Hiebsch & Vorwerg, 1976). Third, and as an outcome of this particular situation, East German social psychology focused on a rather circumscribed number of applied issues, such as the development of social trainings in the context of production work groups. Although there was some form of communication between East and West German social psychologists during the period of the German division (e.g., meetings and visits initiated and/or supported by EASP), overall, scientific exchange was limited. The Zeitschrift für Sozialpsychologie, for instance, was received in East Germany, but it did not serve as an outlet for an ongoing East–West exchange. With the collapse of socialism in 1989, and the unification of East and West Germany in 1990, the history of East German social psychology came to an abrupt end, while West Germany’s social psychology prevailed. One important reason for this was a (politically induced) replacement of virtually the entire academic elite in social psychology at East German universities. West German professors were appointed to the leading social psychology chairs in Jena and Leipzig, and new positions in social psychology were filled by West Germans. For East German scholars, these Ó 2019 Hogrefe Publishing

transformations were biographically dramatic and often personally hurtful. Many West German social psychologists were aware of the tremendous implications of the reunification for German society at large, and they used social psychological concepts and theories to describe and to explain them (see, for instance, the Zeitschrift’s special issue on “Deutschland Ost and Deutschland West,” published in 1993). From today’s perspective, and with hindsight, it seems surprising, however, that the impact of the reunification on the community of East and West German social psychologists itself received relatively little attention as a potential topic of scientific discussion and research. From today’s understanding, the reunification of East and West German social psychology provided unique opportunities to engage in controversies and critical discussion about, for instance, the value bases of social psychological research or social psychology’s role in different political systems. Although such discussion seemed to have met one of the journal’s key goals (see Feger et al., 1970), our interviews revealed that they were rather rare. There were some attempts (mostly initiated by East German psychologists) to discuss the implications of the societal transformations for East German psychology and the biographies of psychological researchers (e.g., Busse, 1996). However, these discussions mainly took place outside the social psychological mainstream publications.

Internationalization After the reunification of East and West Germany and its impact on German social psychology, the next big step was the internationalization of the journal. The decision to change the official language from German to English and to rename it Social Psychology in 2008 was not an easy one. In fact, it took decades to come to this point. Back in the days during the transformation into a professional journal with international standards, language was already an issue. During the last quarter of the 20th century, there was a lively debate on which language should be used in the scientific community (“Sprachenstreit,” see e.g., Lienert, 1977, and Traxel, 1975). The resistance to accept English as the dominant language might partly have been due to the former glory and influence of German. At the end of the 19th century and well into the first-quarter of the 20th, publications in German by German-speaking psychologists (like Wilhelm Wundt, Oswald Külpe, Kurt Koffka, and Kurt Lewin) were much more widely received and honored in Anglo-American regions than was the case after World War II (see also Krampen, Montada, Schui, & Lindel, 2002). For some, publishing in English was not a way to increase the international reputation of social psychology from German-speaking countries, but a sign of marginalization. Furthermore, publishing in a foreign language makes Social Psychology (2019), 50(1), 1–6

it more difficult for German scholars to express their thoughts with the finesse and subtle articulateness of their mother tongue. Thus, in 1987 the efforts to ensure the international standing of the journal by implementing professional and internationally accepted standards (such as periodically changing editors, a wide editorial board, peer review) stopped shortly before switching to English. German remained the official language. Even though it was possible to publish articles in English, this option was hardly used. In the last five years of the Zeitschrift für Sozialpsychologie (2003–2007), only 7 out of 103 articles were published in English. By the turn of the century, the debate about the preferred scientific language was over. Publishing in English became even more important, and by now, the numbers of publications in English, peer-reviewed journals served as a major indicator for scientific success. This development is true for psychology in general, though it was more pronounced in social psychology than in some other subdisciplines. In 2004, Zeitschrift für Sozialpsychologie published an article by Krampen, Schui, and Montada depicting the “International dissemination of social psychology from the Germanspeaking countries.” This article points out that from the mid-1980s on, the proportion of English publications was nearly twice as high in social psychology than in the rest of psychology. This was especially true for the research area of social perception and cognition, in which, by 2001, up to 50% of all publications of German-speaking social psychologists were in English. Only very few other subdisciplines, such as cognitive psychology, biological psychology, and neuropsychology, reached such a high proportion. If social psychology was internationally oriented and many researchers were happy to publish their research in English journals anyway, why keep German as the official language for so many years? One answer might be that Zeitschrift für Sozialpsychologie was more than an outlet to communicate research results. As outlined above, in the beginning it had the additional function of defining the field and strengthening the professional community of social psychologists in Germany and neighboring countries. In the early 1970s, it was the journal for German social psychologists – especially if one preferred an experimental approach. To be recognized in the field, it was important to publish in this journal. However, with growing internationalization, broader communities and other journals grew in importance. Early on, the EASP played a major role in the development of social psychology in Germany. With it, publishing in the European Journal of Social Psychology or other international journals became more important. By the beginning of the 21st century, the discipline of social psychology was well established, research had become a global affair, and publications in other outlets predominantly US journals were far more important for individual Social Psychology (2019), 50(1), 1–6

K. Corcoran et al., Learning From History

careers than an article in Zeitschrift für Sozialpsychologie. Far from being the journal, the editors were confronted with dwindling submission rates: In 2007, only 12 articles were submitted to the journal. This made it difficult to hold up high-quality standards. Something had to be done. Thus, the language switch from German to English was a long overdue step in line with the international focus of social psychologists in German-speaking countries. It immediately helped to connect this outlet to the broader scientific community. This is clearly visible in the impact factor (IF) of the journal Social Psychology (2017: 5-year IF = 2.16) in comparison with Zeitschrift für Sozialpsychologie (2007: 5-year IF = 0.33). Today, Social Psychology is a truly international journal with a diverse editorial board and with submissions from around the globe. Some might look back with nostalgic feelings to a time when the journal was a prime outlet for German social psychologists. Today, it is one among many. However, the journal is surviving and striving, along with it social psychological research in German-speaking countries.

Lessons Learned As the number of interviews conducted for the present paper is limited, the conclusions that we offer in the following are, of course, preliminary. Moreover, the conclusions are subjective – not only in a sense that they are based on the interviewees’ individual recollections but also because they reflect what we personally believe to be important implications for the future of social psychology (and Social Psychology). We are convinced that we have learned something about our discipline that we did not know before and that is worth sharing with a broader international readership. The period of professionalization illustrates the importance of individual commitment and action for the social identity of a (scientific) community. The Zeitschrift für Sozialpsychologie and its founding editors helped West German social psychology to grow as a field. At the same time, it was important to be open, to let go at some point, and to allow changes for the better – such as the professionalization in the mid-1980s. The proactive efforts of some young scholars to professionalize German social psychology were an important step toward revealing the potential of the field. The internationalization and the transformation from the Zeitschrift für Sozialpsychologie to Social Psychology is another good example of the significance of openness to change and the necessity of proactively addressing challenges. The decision to change the journal’s language came not a moment too soon. The decrease in submissions and the related pressure by the publishers made this step unavoidable. In fact, at that time it was not unlikely that Ó 2019 Hogrefe Publishing

K. Corcoran et al., Learning From History

the publication of the journal might have been terminated if nothing had changed. Again, the journal advanced thanks to the determination of certain individuals. However, what is noteworthy is that this change came at a cost. It is fair to say that Social Psychology today can no longer be regarded as determining the social identity of social psychology in the German-speaking countries. Furthermore, English helps to reach a bigger audience within the scientific community, but it might decrease the impact of social psychological research on non-scientific audiences in German-speaking countries. Another interesting observation concerns the period of German reunification. German reunification provided a unique opportunity to test social psychological theories – an opportunity that many social psychologists used to advance both scientific and social progress. The Zeitschrift’s special issue on “Deutschland Ost and Deutschland West” (1993) set an example for concerted efforts among social psychologists to address big societal challenges. In light of pressing social issues (e.g., climate change, aging societies, rising nationalism), one would like to see more similar efforts today. From a different perspective, however, German reunification also offered the potential for selfreflection upon crucial questions pertaining to the philosophy of social psychological science: The issues of explicit and implicit values guiding social psychological research, the multiple scientific identities of social psychological science, and their stabilizing or destabilizing roles in society. With the benefit of hindsight, and from the standpoint of an uninvolved observer, it occurs to us that the potential for critical self-reflection (including the application of our theories in our own field) remained relatively underexploited at this time. It is our belief that future systematic research on this particular period of social psychology’s history, with a balanced focus on both East and West German social psychology, is specifically worthwhile. Such research has a strong potential, we believe, to advance our understanding of how contextual, political, and ideological factors shape and constrain social psychological theorizing and research. Where do these observations leave us with regard to the future of our field? First, we should continue to be open to change and encourage each other to rethink and question established structures and approaches. The current Open Science movement can be regarded as a model example of this idea. Social Psychology, by the way, was among the first journals of the field to address issues related to Open Science and replicability (e.g., Klein et al., 2014). Second, we should try to proactively address future challenges. Individual or group efforts can make a difference! One future challenge will be to increase the accessibility of research. Social Psychology’s professionalization and internationalization affected those who became able to contribute to, access,

Ó 2019 Hogrefe Publishing

and make use of the published research. Today, many researchers and practitioners are still unable to afford and/ or process research published in journals such as Social Psychology. We need to think about ways to address this shortcoming. One interesting move in this direction is actually currently in progress. Only recently, a number of national research foundations (e.g., the Dutch Science Foundation, NWO) signed “Plan S,” an initiative for openaccess publishing in science. On the one hand, this initiative seems to solve the problem of unequal access to publications. On the other hand, however, this change in the publication system is also very challenging. Who is to pay for the fees and will everybody be able to pay these fees? Furthermore, the availability of endless publication space might change criteria for publication and rejection and, in the end, not necessarily lead to a better (quality of the) system. This brings us to another central learning: Good editorial teams, who take time to work on the articles with their authors and who dare to make (quick) decisions, are key to the success and the impact of scientific journals. Luckily, Social Psychology has been and still is blessed with many of them.

References Busse, S. (1996). Psychologie im Realsozialismus. DDR-Psychologen im Interview. Übergänge - Aspekte gesellschaftlichen Wandels. Band 1 [Psychology in real socialism. GDR Psychologists in interviews. Transitions–aspects of societal change. Vol 1]. Pfaffenweiler, Germany: Centaurus. Feger, H., Graumann, C.-F., Holzkamp, K., & Irle, M (1970). Vorwort der Herausgeber [Editorial]. Zeitschrift für Sozialpsychologie, 1, 1–2. Hiebsch, H., & Vorwerg, M. (1976). Einführung in die marxistische Sozialpsychologie [Introduction to Marxist social psychology]. Berlin, Germany: VEB Dt. Verlag der Wissenschaften. Klein, R. A., Ratliff, K. A., Vianello, M., Adams, R. B. Jr., Bahník, Š., Bernstein, M. J., . . . Nosek, B. A. (2014). Investigating variation in replicability: A “many labs” replication project. Social Psychology, 45, 142–152. https://doi.org/10.1027/1864-9335/ a000178 Krampen, G., Montada, L., Schui, G., & Lindel, B. (2002). Zur Rezeptionsgeschichte von Fachliteratur aus der deutschsprachigen Psychologie in der amerikanischen Psychologie (et vice versa) [A reception history of scientific literature from German-language psychology by American psychology (et vice versa)]. Psychologie und Geschichte, 10, 84–113. Krampen, G., Schui, G., & Montada, L (2004). Zur internationalen Verbreitung der Sozialpsychologie aus dem deutschsprachigen Bereich [International dissemination of social psychology from the German-speaking countries]. Zeitschrift für Sozialpsychologie, 35, 83–91. https://doi.org/10.1024/0044-3514. 35.2.83 Lienert, G. A. (1977). Über Werner Traxel: Internationalität oder Provinzialismus, zur Frage: Sollten Psychologen in Englisch publizieren? [About Werner Traxel: Internationality or provincialism,

Social Psychology (2019), 50(1), 1–6

the question is: Should psychologist publish in English]. Psychologische Beiträge, 19, 487–492. Lück, H. E. (2002). Empirische Sozialpsychologie in Westdeutschland: Ein verzögerter Neubeginn? [Empirical social psychology in West Germany: A delayed restart?] Psychologie und Geschichte, 10, 323–332. Schmidt, H.-D. (1987). German democratic republic. In A. R. Gilgen & C. K. Gilgen (Eds.), International handbook of psychology (pp. 222–238). New York, NY: Greenwood Press. Traxel, W. (1975). Internationalität oder Provinzialismus? Über die Bedeutung der deutschen Sprache für deutschsprachige Psychologen [Internationality or provincialism? On the importance of the German language for German-speaking psychologists]. Psychologische Beiträge, 17, 584–594.

Social Psychology (2019), 50(1), 1–6

K. Corcoran et al., Learning From History

History Received December 6, 2018 Accepted December 6, 2018 Published online January 29, 2019 Authorship All authors contributed equally and are listed in alphabetical order. Stefan Stürmer FernUniversität in Hagen Fakultät für Psychologie 58084 Hagen Germany stefan.stuermer@fernuni-hagen.de

Ó 2019 Hogrefe Publishing

Original Article

Intended, Unintended, and Unknown Consequences of Contact The Role of Positive–Negative Contact on Outgroup Attitudes, Collective Action Tendencies, and Psychological Well-Being Sabahat C. Bagci1 and Abbas Turnuklu2 1

Department of Psychology, Isik University, Turkey

Dokuz Eylul University, Turkey

Abstract: We investigated the role of positive and negative contact on outgroup attitudes, collective action tendencies, and psychological wellbeing among minority (Kurds) and majority (Turks) group members in a conflict area (N = 527), testing ingroup identification, relative deprivation, and perceived discrimination as potential mediators in these associations. Contrary to recent research studies demonstrating the superiority of negative contact effects, positive contact was generally a stronger determinant of the dependent variables, directly and indirectly, in both groups, although negative contact also had some direct and indirect associations with the outcomes. Findings highlight the need to incorporate the role of positive and negative contact to provide a full understanding of the potential benefits/costs of the contact strategy in conflict settings. Keywords: positive contact, negative contact, outgroup attitudes, collective action, psychological well-being

With increased ethnic diversity, multiculturalism, and globalization throughout the world, intergroup contact theory (Allport, 1954; Pettigrew, 1998) which has consistently shown the positive role of intergroup contact on reducing prejudice and negative intergroup behavior (see Pettigrew & Tropp, 2006 for a meta-analysis) has been one of the prominent theories intergroup relations researchers commonly relied on over the years. Although the effectiveness of the contact strategy has been evidenced by cross-sectional (Brown & Hewstone, 2005; Pettigrew, 1998, 2016; Pettigrew & Tropp, 2006, 2008), longitudinal (Binder et al., 2009; Swart, Hewstone, Christ, & Voci, 2011), and experimental studies (e.g., Mendoza-Denton & Page-Gould, 2008), the last decade of intergroup relations research has witnessed an intensive re-examination of the contact theory focusing on its understudied aspects. The classical theory has been recently criticized regarding the relative lack of studies (a) incorporating positive and negative contact, (b) examining the unintended effects of intergroup contact such as reduced motivation to engage in collective action, (c) providing the perspectives of both minority and majority status participants, and (d) demonstrating the effects of contact in conflict-ridden settings (Pettigrew & Hewstone,

Ó 2018 Hogrefe Publishing

2017; Tropp, Mazziotta, & Wright, 2016). While researchers have now gradually started to consider the role of both positive and negative contact on the so-called intended (social cohesion) and unintended (social change) consequences of contact strategy (e.g., Barlow, Paolini, et al., 2012; Reimer et al., 2017), whether and how positive and negative contact among majority and minority group members relates to group members’ well-being at the individual level is relatively less known. The current study aims to incorporate these distinct dimensions of intergroup contact by testing the role of positive and negative contact on positive outgroup attitudes and collective action tendencies, as well as on psychological well-being – an outcome that has been far less studied in the contact literature. Moreover, we investigate these relationships by examining three critical mechanisms as mediators – ingroup identification, relative deprivation, and perceived discrimination – and, including both majority (Turks) and minority (Kurds) status participants in a conflict area. The Kurdish group constitutes a relatively disadvantaged minority group in Turkey (Çakal, Hewstone, Güler, & Heath, 2016), which has been often considered an oppressed minority group over the years (Bagci & Çelebi,

Social Psychology (2019), 50(1), 7–23 https://doi.org/10.1027/1864-9335/a000355

2017). Although a large portion of Kurdish people reside in Southeastern Turkey forming the majority of the local population, with recent waves of immigration from the East of Turkey to the West, many big cities have now started to include a fair amount of Kurdish group members, allowing intergroup contact to occur more frequently (Çakal et al., 2016). While some neighborhoods represent a relatively higher percentage of Kurdish group members, there is no visible segregation across the communities in school and work settings (Bagci & Çelebi, 2017). Nevertheless, severe and violent military attacks between separatist groups and the Turkish Army in the Southeast parts of Turkey have resulted in the deaths of thousands of people from both groups (Göçek, 2011), reflected on both Turkish and Kurdish group members’ intergroup attitudes and behaviors all over the country. Previous social psychological studies conducted in this setting have demonstrated that both group members are highly identified with their ingroup, display negative outgroup attitudes, and perceive high levels of intergroup conflict (Bagci & Çelebi, 2017, 2018; Bilali, Çelik, & Ok, 2014; Çelebi, Verkuyten, Köse, & Maliepaard, 2014), making the current setting an excellent milieu for the study of positive and negative contact effects among majority and minority group members. Recent research has also shown contact, in the form of cross-group friendships, to be related to positive outgroup attitudes (Bagci & Çelebi, 2017) and reduced motivation for collective action among both advantaged Turks and disadvantaged Kurds (Çakal et al., 2016; Uluğ & Cohrs, 2017). Yet, the differential aspects of positive and negative contact, psychological outcomes of contact at the individual level, as well as potential mediators involved in these relationships need to be further explored in order to disentangle the complex Turkish– Kurdish interethnic relationship.

Intergroup Contact and Outgroup Attitudes Previous research has shown that direct contact with members of different groups is related to more positive and less negative intergroup outcomes across different contexts (e. g., Pettigrew, 1997; Pettigrew & Tropp, 2008), mainly through affective processes such as increased empathy (Finlay & Stephan, 2000; Pettigrew & Tropp, 2008) and decreased intergroup anxiety (Barlow, Louis, & Hewstone, 2009; Paolini, Hewstone, Cairns, & Voci, 2004), as well as cognitive processes such as increased perspective-taking (Pettigrew & Tropp, 2008). As in Allport’s (1954) original formulations of the contact hypothesis, the benefits typically require certain conditions to be met, particularly that participatory cooperative contact takes place with common goals and between approximately equal-status individuals, Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

although these contact conditions were later found to be facilitators, rather than requirements (Pettigrew & Tropp, 2006). Beyond commonly studied outcomes of intergroup contact such as attitudes and prejudice, it is now known that contact is associated with many peace-related outcomes such as increased trust, forgiveness, and reconciliation efforts, even in conflict-ridden and segregated settings like Northern Ireland and South Africa (Hewstone, Cairns, Voci, Hamberger, & Niens, 2006; Tropp et al., 2017). Moreover, although contact effects have been shown to be more influential among majority group members (e.g., Tropp & Pettigrew, 2005), research has shown contact, in particular cross-group friendships, to be also critical for the improvement of intergroup attitudes among minority group members (e.g., Levin, Van Laar, & Sidanius, 2003; PageGould, Mendoza-Denton, & Tropp, 2008; Tropp, 2006), suggesting the positive role of contact among both minority and majority group members, especially in historically intricate intergroup contexts. A recently growing trend in this literature has been looking into the possibility of negative contact effects reducing positive outgroup attitudes and behaviors (e.g., Barlow, Paolini, et al., 2012; Graf, Paolini, & Rubin, 2014; Paolini et al., 2014; Paolini, Harwood, & Rubin, 2010). In his reformulation of the contact theory, Pettigrew (1998) suggested that contact may become a negative experience in intergroup settings characterized by high hostility and conflict. Although relatively less studied than positive contact, negative contact experiences have been shown to undermine positive intergroup attitudes by increasing perceived threat and negative stereotypes (Aberson & Gaffney, 2009; Stephan et al., 2002). More recent research investigating the effects of positive and negative contact simultaneously has shown that the detrimental effects of negative contact may even outweigh the benefits of positive contact (e.g., Barlow, Paolini, et al., 2012; Bekhuis, Ruiter, & Coenders, 2013). In line with this, Paolini et al. (2010) proposed that negative contact effects are likely to be more influential than positive contact effects on attitudes, since negative contact increases category salience which would in turn generalize more easily to outgroup attitudes (Brown & Hewstone, 2005). Nevertheless, in most studies, positive contact has been found to be more frequent than negative contact, which has been suggested to compensate the stronger influence of negative contact (Aberson, 2015; Graf et al., 2014). While other research has found that positive contact was still a stronger determinant of intergroup attitudes than negative contact (e.g., Pettigrew, Tropp, Wagner, & Christ, 2011), a number of studies also showed that both positive and negative contact may be equally likely to affect intergroup outcomes in opposite directions (Árnadóttir, Lolliot, Brown, & Hewstone, 2018; Mazziotta, Rohmann, Wright, Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Tezanos-Pinto, & Lutterbach, 2015; Visintin, Green, Pereira, & Miteva, 2017). Moreover, although relatively understudied, previous research has found this effect to exist among both minority and majority group members (Árnadóttir et al., 2018). Based on these findings, we proposed that positive contact would be associated with more positive outgroup attitudes and negative contact would be associated with less positive outgroup attitudes among both minority and majority group members.

Intergroup Contact and Collective Action Motivation Another recent research avenue in the re-examination of the intergroup contact theory has been investigating contact’s unintended effects. This proposition stemmed from initial suspicion toward the overall benefits of intergroup contact in tackling prejudice at the societal level (e.g., Dixon, Durrheim, & Tredoux, 2005). Further empirical research has consistently demonstrated that contact was related to reduced motivation for social change by creating false expectancies about intergroup equality, which is often referred to as the irony of harmony (Dixon, Tropp, Durrheim, & Tredoux, 2010; Pettigrew & Hewstone, 2017; Saguy, Schori-Eyal, Hasan-Aslih, Sobol, & Dovidio, 2016; Saguy, Tausch, Dovidio, & Pratto, 2009) or contact’s sedative effects (Cakal, Hewstone, Schwär, & Heath, 2011) in intergroup relations literature. According to Wright and Lubensky (2009), contact is likely to reduce collective action tendencies, since intergroup contact and social change theories have opposite motivations and functions, thereby requiring individuals to hold opposite emotions and perceptions. For example, while intergroup contact strategy aims to reduce ingroup identification and perceived discrimination and thereby bring group members closer, motivation for social change is often fueled by increased perceptions of inequality, as well as negative emotions and dislike toward outgroup members (see also Van Zomeren, Postmes, & Spears, 2008). While previous studies investigating the potential sedative effects of intergroup contact have often found contact to be associated with reduced motivation for social change and collective action tendencies, as well as reduced support for policies favoring ingroup’s rights among both minority (Cakal et al., 2011; Dixon et al., 2010; Saguy et al., 2009) and majority group members (Çakal et al., 2016; Saab, Harb, & Moughalian, 2017), other research has demonstrated that intergroup contact may not unconditionally deter group members from social change. For example, using a multi-level analysis, Kauff, Green, Schmid, Hewstone, and Christ (2016) demonstrated that contact was not associated with reduced collective action tendencies Ó 2018 Hogrefe Publishing

in neighborhoods where majority group members reported positive intergroup contact with minority group members. In line with this, Becker, Wright, Lubensky, and Zhou (2013) found contact to be unrelated to social change motivation when outgroup members explicitly affirmed their contestation for intergroup equality. Incorporating recent research on the positive–negative contact asymmetry and contact’s unintended effects, Reimer et al. (2017) provided initial evidence for how positive and negative contact may differentially influence collective action tendencies. The authors found that LGBT students’ collective action tendencies were dependent on the level of negative contact (but not of positive contact) with heterosexuals cross-sectionally and longitudinally, showing the importance of distinguishing positive and negative contact on the extent to which ingroup members were motivated for social change. In light of these initial findings, we suggested that only negative contact would be related to higher levels of collective action tendencies, whereas positive contact would be unrelated to the extent to which group members engage in collective action.

Intergroup Contact and Psychological Well-Being Extending previous studies which focused on the desired and undesired contact effects, the current study aimed to reveal whether and how contact was associated with psychological well-being among majority and minority participants in a conflict area. Previous research investigating the role of intergroup contact on psychological outcomes at the individual level has been restricted, but a few research papers provide initial evidence for this association. For example, Mendoza-Denton and Page-Gould (2008) demonstrated that cross-group friendships, which represent the ultimate form of positive intergroup contact, were related to increased institutional belonging and satisfaction among minority group university students. Voci, Hadziosmanovic, Cakal, Veneziani, and Hewstone (2017) focused on the association between contact and mental health among Bosnians recruited from a post-conflict area. More specifically, the authors found that postwar contact with the outgroup led Bosnians to display reduced levels of morbidity and posttraumatic stress disorder through increased forgiveness and decreased social distance, suggesting the potential importance of intergroup contact for the psychological well-being and mental health of group members, especially in conflict areas. Bagci, Turnuklu, and Bekmezci (2018), found that cross-group friendships were related to greater psychological well-being among physically disabled adults through increased perceptions of integration and empowerment. Other research from developmental Social Psychology (2019), 50(1), 7–23

psychology literature has demonstrated cross-group friendships, in particular their quality, to be positively related to psychological well-being and resilience (Bagci, Rutland, Kumashiro, Smith, & Blumberg, 2014), and relational inclusiveness (Kawabata & Crick, 2008) among children recruited from multi-ethnic classrooms. While positive contact forms are likely to foster psychological well-being among both majority and minority groups, negative intergroup experiences such as intergroup anxiety, race-based rejection fear, and negative expectations from the outgroup have been previously demonstrated to be associated with lower psychological well-being and mental health (e.g., Schmitt, Branscombe, Postmes, & Garcia, 2014), which implies the possible detrimental effects of negative contact experiences on psychological well-being. Moreover, negative contact experiences can be equally destructive for majority group members too, who also perceive race-based rejection fear and intergroup anxiety in competitive intergroup settings (e.g., Barlow, Sibley, & Hornsey, 2012; Tropp et al., 2017). Therefore, we suggested that, for both groups, positive and negative intergroup contact experiences should be both related to psychological well-being, albeit in opposite directions.

Potential Mediators We also aimed to test three different mediators that could potentially explain the associations between positive and negative contact and outgroup attitudes, collective action tendencies, and psychological well-being: ingroup identification, relative deprivation, and perceived discrimination. Ingroup Identification We proposed that ingroup identification would function as one of the critical mediators in these associations. Previous research has shown that intergroup contact does not only relate to the formation of outgroup attitudes, but it also influences how group members evaluate their ingroup membership. Accordingly, intergroup contact is likely to improve outgroup attitudes, while simultaneously distancing individuals from their ingroup, a phenomenon that is known as “deprovincialization” (Pettigrew, 1997, 2009). Consequently, through the deprovincialization process, group members show positive attitudes not only toward groups that are contacted, but also toward other groups that are not involved in contact, making contact generalizable to other groups and situations (Secondary Transfer Effect, Pettigrew, 1997). Past research provided empirical evidence for this hypothesis showing intergroup contact to be associated negatively with ingroup identification (Verkuyten, Thijs, & Bekhuis, 2010) and specifically, cross-group friendships, a Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

positive and strong intergroup contact form, to be related to lower levels of ingroup identification (e.g., Tausch, Saguy, & Bryson, 2015). On the contrary, negative contact was found to be even more strongly related to increased ingroup identification (Reimer et al., 2017). Therefore, we proposed that positive contact would be related to decreased ingroup identification, whereas negative contact which has relatively higher category salience (e.g., Graf et al., 2014) would be related to increased levels of ingroup identification. Ingroup identification, in turn, should relate to more negative outgroup attitudes, based on the Social Identity Theory (Tajfel & Turner, 1979), which suggests that ingroup members would naturally favor the ingroup over the outgroup in order to affirm their personal sense of self-esteem through a strong ingroup identification. Moreover, ingroup identification should relate positively to collective action tendencies suggested by Social Identity Model of Collective Action (SIMCA, Van Zomeren et al., 2008) demonstrating ingroup identification as a strong precursor of collective action tendencies (e.g., Cakal et al., 2011). Finally, numerous studies have shown ingroup identification to be positively related to psychological well-being and mental health among both advantaged and disadvantaged group members; recent social identity research studies have suggested that social identities constitute “social cures” for group members and make them more resilient against negative life experiences, in particular when they satisfy some basic human needs (e.g., Çelebi, Verkuyten, & Bagci, 2017; Haslam, Jetten, Postmes, & Haslam, 2009; Jetten, Haslam, & Haslam, 2012; Vignoles, 2011). Accordingly, social identities improve well-being by providing positive feelings such as belongingness, self-esteem, and self-efficacy, thereby increasing group members’ overall mental and physical health. In light of these findings, it is possible that ingroup identification would be related to lower levels of positive outgroup attitudes, but higher levels of support for collective action and psychological well-being.

Relative Deprivation We further suggested that positive and negative contact would be related to outgroup attitudes, collective action tendencies, and psychological well-being through the process of relative deprivation. Previous studies in intergroup relations literature have suggested that contact would provide opportunities for ingroup members to compare themselves to outgroup members and thereby influence their relative sense of deprivation (Pettigrew & Hewstone, 2017). While previous research has mainly focused on contact without taking into account its valence, it is possible that positive contact reduces relative deprivation by decreasing the perception of intergroup inequality (e.g., Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Saguy et al., 2009), whereas negative contact is likely to enhance the sense of relative deprivation by making intergroup differences more salient (Graf et al., 2014). In turn, compared to the ones with lower levels of relative deprivation, group members with higher levels of relative deprivation are more likely to engage in collective action tendencies (Cakal et al., 2011) and report less positive outgroup attitudes (e.g., Guimond & Dambrun, 2002; Pettigrew et al., 2008), as well as lower sense of psychological well-being (e.g., Mishra & Carleton, 2015). Perceived Discrimination Finally, we assumed that perceived discrimination would be another potential mediator linking the associations between positive and negative contact and outgroup attitudes, collective action tendencies, and psychological well-being. The relevant literature has indicated that contact may change perceived attitudes from outgroup members. For example, Dixon et al. (2010) and Ellison and Powers (1994) demonstrated that intergroup contact was associated with decreased levels of perceived discrimination. Further research has demonstrated that in particular commonality-focused intergroup contact led group members to perceive lower status hierarchy and to be less likely to attribute negative treatment to discrimination (Saguy & Chernyak-Hai, 2012). Therefore, it is expected that while positive contact decreases the perception of perceived discrimination, negative contact would make intergroup differences and inequality more evident and thereby enhance perceived discrimination (Reimer et al., 2017). In turn, perceived discrimination, in the form of negative expectancies from the outgroup and anticipation of race-based rejection, is likely to produce outgroup negativity among both minority and majority group members (Barlow, Sibley, & Hornsey, 2012), lead to increased motivation for social change (Tropp, Hawi, Van Laar, & Levin, 2012), and reduce psychological well-being at the individual level (e.g., Jasinskaja-Lahti, Liebkind, Jaakkola, & Reuter, 2006; Schmitt et al., 2014). We have also tested the possibility of a sequential mediation where the path from contact to outcome measures via perceived discrimination may be further mediated by ingroup identification. This is based on the rejectionidentification model (Branscombe, Schmitt, & Harvey, 1999) which suggested that feelings of discrimination may result in strong identification with the ingroup. Empirical evidence among various social groups such as women, ethnic minorities, and people with body piercing demonstrated that perceived discrimination relates to increased ingroup identification, which in turn results in greater psychological well-being (e.g., Branscombe et al., 1999; Jetten, Branscombe, Schmitt, & Spears, 2001). Therefore, it could be expected that some effects of perceived discrimination Ó 2018 Hogrefe Publishing

on the suggested outcome variables may be further mediated by increased ingroup identification.

The Current Study In summary, the current study aimed to provide a thorough understanding of intergroup contact’s consequences among majority and minority group participants in a conflictridden interethnic context, by simultaneously investigating the role of positive and negative contact on three distinct constructs. Furthermore, we intended to investigate whether three mechanisms – ingroup identification, relative deprivation, and perceived discrimination – significantly explained these associations. We aimed to contribute to the relevant literature in various ways. First, although previous research has focused on intended and unintended consequences of intergroup contact strategy mostly exclusively, only a few researchers have included positive outgroup attitudes and collective action tendencies as simultaneous outcomes of intergroup contact (e.g., Çakal et al., 2016; Reimer et al., 2017; Tausch et al., 2015). Other research has indicated that contact may not necessarily decrease collective action tendencies by improving positive outgroup attitudes, and attitudes and motivation for social change may be independent outcomes of intergroup contact (Bagci, Stathi, & Piyale, in press). Second, to our knowledge, no research has specifically focused on whether and how contact may relate to the overall psychological well-being of majority and minority group members and no research to date has distinguished between the differential role of positive and negative contact on this outcome. This is critical since, beyond how contact influenced outgroup attitudes and collective action tendencies at the societal or group level, the psychological outcomes of intergroup contact at the individual level should be also considered an essential consequence of the contact strategy. Third, by testing indirect associations on these relationships via ingroup identification, perceived discrimination, and relative deprivation, we aimed to provide a complete picture of the function of negative and positive contact in a conflict society. Fourth, we included both minority and majority group members in the context of positive–negative contact asymmetry, which has not been investigated until recently (Árnadóttir et al., 2018; Hayward, Tropp, Hornsey, & Barlow, 2017). Finally, we examined our research questions in the context of Turkish–Kurdish interethnic relationship which is characterized by high status differences, conflict, and hostility. Examining the role of positive and negative intergroup contact on attitudes, collective action tendencies, and psychological well-being in this setting may provide opportunities to evaluate how intergroup contact strategy may be beneficial and/or detrimental for both group members at the applied level. Social Psychology (2019), 50(1), 7–23

Method Participants and Procedure A total of 5271 university students completed the study (290 females and 237 males, Mage = 20.57, SD = 2.02). Participants self-reported their ethnic background with a single item (“We are all Turkish citizens, but may belong to different ethnic groups. Which ethnic group do you feel you belong to?”, Konda, 2011); 376 participants stated their ethnic background as Turkish, whereas 151 participants selfdescribed as Kurdish. The mean subjective socioeconomic status of participants was middle class with a mean of 4.02 (SD = 1.02), ranging from 1 (= very low) to 7 (= very high). Participants were recruited from two public universi_ ties in the West of Turkey (Izmir). Pen and pencil questionnaires were distributed during Fall 2017 by research assistants and main researchers within university campuses through convenience and snowball sampling. Participants were fully debriefed at the end of the study.

Measures Positive and Negative Contact Positive and negative contact were each assessed by a single item (Barlow, Paolini, et al., 2012). Participants were first given a specific instruction asking them to consider their Turkish peers if they were Kurdish or their Kurdish peers if they were Turkish (“If you are Turkish, please think about your Kurdish peers” and “If you are Kurdish, please think about your Turkish peers”). Next, they were asked to report the extent to which they had positive and negative contact with these peers (“How frequently do you have positive/negative contact with these peers?”). The response scale ranged from 1 (= never) to 7 (= all the time). Items were either not significantly correlated (r = .09, p = .27, for the minority), or weakly correlated (r = .12, p = .02, for the majority), suggesting the distinctive nature of both contact types. Ingroup Identification Ingroup identification was measured by four items originally proposed by Verkuyten (2007) and adapted to the Turkish context by Bagci and Çelebi (2018). Participants were asked to report the extent to which they identified with and belonged to their ethnic ingroup (e.g., “I identify with my ethnic group” and “I feel connected to my ethnic

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

group”). The response scale ranged from 1 (= strongly disagree) to 7 (= strongly agree). The reliability of the scale across the two groups was excellent (α = .92 for Turks and .88 for Kurds). Relative Deprivation Relative deprivation was measured by two items assessing how much participants felt that their ethnic group was relatively deprived compared to the outgroup. Participants were first given a specific instruction about which group they needed to compare themselves (“If you are Turkish, compared to Kurds. . .; If you are Kurdish, compared to Turks. . .”). The instruction was followed by two items: “To what extent your ethnic group is economically advantaged?” and “To what extent your ethnic group is socially advantaged?” (adapted from Cakal et al., 2011). The response scale ranged from 1 (= very disadvantaged) to 7 (= very advantaged). Both items were reverse-coded so that higher scores indicated higher levels of relative deprivation. Reliabilities assessed by Cronbach’s α coefficients demonstrated acceptable reliability scores across the two groups (α = .68 and .71 for Turks and Kurds, respectively). Perceived Discrimination We assessed the level of perceived discrimination by two items previously used by Tropp et al. (2012). One of the items measured perceived discrimination at the individual level (“To what extent do you feel you are personally discriminated because of your ethnic background?”), and the other item assessed perceived discrimination at the group level (“To what extent do you feel your ingroup members are discriminated because of their ethnic background?”). The response scale ranged from 1 (= never) to 7 (= all the time). The reliability of the scales was good for both groups (α = .87 for Turks and .79 for Kurds). Outgroup Attitudes This was measured by the feeling thermometer developed by Esses, Haddock, and Zanna (1993), using two single items which asked participants to rate their feelings toward Turkish and Kurdish group members, ranging from 0 (= extremely unfavorable) to 100 (= extremely favorable) degrees (reported in number in a single textbox), with higher scores indicating more positive attitudes toward the group. A new variable “outgroup attitudes” was then formed by selecting the appropriate outgroup based on participants’ selfreported ethnic group.

The total number of participants reported is the net number of participants who reported themselves as either Turkish or Kurdish, after the exclusion of participants from other backgrounds (such as Arab, Armenian, Georgian), who were instructed to skip sections that are only concerned with Turkish/Kurdish ethnic group membership. We also excluded 5 participants over the age of 29, which stood as outliers in initial analyses.

Social Psychology (2019), 50(1), 7–23

Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Collective Action Tendencies We assessed participants’ tendencies to engage in collective action by four items adapting the collective action tendencies scale previously used by Çakal et al. (2016) to the Turkish–Kurdish context (e.g., “I would be willing to sign a petition to improve the current situation of my ethnic group in Turkey” and “I would be willing to participate in a peaceful demonstration to improve the current conditions for my ethnic group.”). The response scale ranged from 1 (= strongly disagree) to 7 (= strongly agree), with higher scores indicating greater support for collective action on behalf of ingroup rights. The reliability of the scale was high among both Turks (α = .90) and Kurds (α = .90).

ing the following cut-off points: w2/df < 2, CFI .93, RMSEA .07, and SRMR .07 (Bagozzi & Yi, 2012; Marsh, Hau, & Wen, 2004). In order to test specific direct and indirect associations from the majority and minority group perspectives, we modeled the relationships separately for Turks and Kurds.3 The significance of indirect effects was assessed by the Model Indirect command on Mplus.

Results Preliminary Analyses

Psychological Well-Being We measured psychological well-being by the 8-item Flourishing Scale (Diener et al., 2010) which taps into different aspects of psychological well-being such as self-esteem, positive relationships, and feelings of competence (e.g., “I lead a purposeful and meaningful life” and “I am a good person and have a good life”). The response scale ranged from 1 (= strongly disagree) to 7 (= strongly agree), with higher scores indicating greater psychological well-being. The Turkish adaptation of the scale has been previously shown to have good psychometric properties (Telef, 2013). The reliability of the scale in the current sample was high (α = .87 for the Turks and .89 for the Kurds).2 Please see Electronic Supplementary Materials 1 (ESM 1) for full materials.

Data Strategy Data were analyzed with structural equation modeling (SEM) using Mplus Version 7 (Muthén & Muthén, 1998– 2018). SEM is an effective strategy to analyze the current data, as it is able to handle multiple independent, mediator, and dependent variables in a single model and allows the use of latent constructs. While positive contact, negative contact, and outgroup attitudes were represented by observed variables as they were measured by a single item, other variables were represented by latent constructs. For psychological well-being which includes eight items, we used the partial disaggregation method by combining items into three different parcels representing the latent variable. This is an efficient way of modeling, since the method provides more stable estimates, reduces the level of random error, and improves approximation of normality distributions for complex data (Bagozzi & Heatherton, 1994; Von der Heidt & Scott, 2007). Model fit was achieved by obtain2

We first conducted a 2 (valence: Positive Negative) by 2 (status: Majority Minority) mixed analysis of variance (ANOVA) test to reveal the main effect of contact valence and ethnic group status on contact. There was a significant main effect of valence, Wilks’ Λ = .52, F(1, 517) = 476.02, p < .001, ηp2 = .48; positive contact was reported more frequently than negative contact. There was also a significant main effect of ethnic group, F(1, 517) = 30.79, p < .001, ηp2 = .06, showing Kurds to indicate greater contact compared to Turks. There was no significant interaction between contact valence and group status, Wilks’ Λ = .99, F(1, 517) = 1.66, p = .20, ηp2 = .003. Next, a series of independent samples t-tests were conducted to reveal significant differences between the main variables across group status (see Table 1). Findings demonstrated that Kurds reported significantly higher levels of both positive and negative contact, higher levels of ingroup identification, relative deprivation, and perceived discrimination, as well as higher levels of collective action tendencies and positive outgroup attitudes compared to Turks. There were no significant differences across group status in terms of psychological well-being. Table 2 further reveals bivariate correlations among the main variables across the two groups.

Structural Equation Models Initially, measurement models with latent variables were assessed for both groups. Results demonstrated that for both majority and minority groups, measurement models fitted data well, w2(80) = 182.54, p < .001, w2/df = 2.28, CFI = .97, RMSEA = .058, SRMR = .034; w2(80) = 133.59, w2/df = 1.67, CFI = .96, RMSEA = .067, SRMR = .049,

The questionnaire also included scales that are beyond the scope of this study for exploratory purposes (extended contact and other contact quality items, Hopkins Checklist-25, collective self-esteem, and life satisfaction scale). Although our sample size for the Kurdish group was relatively low (N = 151) for conducting an SEM model, previous research has indicated that a minimum sample size of 100 is satisfactory for the use of SEM (e.g., Bagozzi & Yi, 2012; Kline, 2005).

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Table 1. Means and standard deviations for the main study variables across groups M (SD) Turkish

Kurdish

p < .001

Positive contact

4.30 (1.83)

5.08 (1.56)

4.93

323.22

Negative contact

2.06 (1.42)

2.56 (1.69)

3.18

238.90

.002

Ingroup identification

4.47 (1.95)

4.93 (1.83)

2.51

524.00

.01

Relative deprivation

3.27 (1.07)

4.82 (1.32)

12.66

226.10

< .001

Perceived discrimination

2.31 (1.49)

4.48 (1.63)

14.07

249.39

< .001

54.14 (23.38)

60.77 (26.24)

2.66

240.19

.01

Collective action tendencies

4.08 (1.86)

5.00 (1.76)

5.16

518.00

< .001

Psychological well-being

4.97 (1.12)

5.13 (1.16)

1.50

525.00

.14

Outgroup attitudes

Table 2. Bivariate correlations among main study variables for Turkish and Kurdish groups 1

1. Positive Contact

–

.09

2. Negative Contact

.12*

–

3. Ingroup identification

.16**

.03

4. Relative deprivation

.01

.14**

5. Perceived discrimination

.07

.05

6. Outgroup attitudes 7. Collective action tendencies

.39*** .05

8. Psychological well-being

.08

.07

.18*

.25**

.21**

.04

.07

.05

.08

.05

–

.06

.13

.07

.29***

.17*

.30***

.32***

.30***

.16y

.28**

.24**

.03

.09y .21*** .14**

– .03 .08

– .12*

.08

.56***

.06

.19***

.16**

.21***

.19***

.01

.45***

– .15** .05

.10

.20* – .27***

.28*** .01

.19* .21* –

Notes. The upper right of the diagonal displays results for the Kurds, and the lower left of the diagonal displays results for Turks. yp < .10; *p < .05; **p < .01; ***p < .001.

respectively, with all observed variables significantly loading on the relevant latent factors. The Model With the Majority Group We started with a partially mediated model where direct paths between positive and negative contact4 and outcome measures were included. We controlled for age and socioeconomic status and included a priori correlational paths between ingroup identification and relative deprivation and relative deprivation and perceived discrimination. As our preliminary analyses showed perceived discrimination to predict ingroup identification, we also included a direct path from perceived discrimination to ingroup identification (in line with rejection-identification model). The partially mediated model yielded a good fit, w2(129) = 170.74, p = .01, w2/df = 1.32, CFI = .99, RMSEA = .029, SRMR = .036, indicated by excellent fit indices. While only positive contact was negatively associated with ingroup identification (β = .18, p = .001), only negative contact was significantly related to higher levels of relative deprivation (β = .15, p = .02) and perceived discrimination (β = .12, p = .03). On the other hand, ingroup 4

identification was significantly associated with greater collective action tendencies and higher levels of psychological well-being (β = .57 and β = .25, p < .001), but did not predict outgroup attitudes. Relative deprivation was also significantly, but negatively associated with psychological wellbeing (β = .15, p = .01). Perceived discrimination, on the other hand, was not associated with any outcome measures, but significantly predicted ingroup identification (β = .22, p < .001). An examination of direct effects between contact measures and outcome variables indicated that while positive contact was positively and substantively associated with positive outgroup attitudes (β = .41, p < .001), negative contact was also significantly and negatively related to outgroup attitudes (β = .12, p = .02). While neither of the contact types directly related to collective action tendencies, both positive and negative contact were related to psychological well-being in opposite directions (β = .16, p = .002 and β = .15, p = .004, respectively). Figure 1 presents the final model for the majority group. Among the indirect effects, the effects of positive contact on collective action tendencies and psychological

As our initial analyses demonstrated that positive and negative contact were not normally distributed (skewness = 1.79 for negative contact and .31 for positive contact, normality tests p < .001), we used the log-transformed (base 10) versions of these variables in further analyses.

Social Psychology (2019), 50(1), 7–23

Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Figure 1. Final SEM model for the majority group. Standardized path coefficients are displayed. ID = Identification, RD = Relative Deprivation, PD = Perceived Discrimination, CA = Collective Action, PWB = Psychological well-being. Final Model Fit: w2(129) = 170.74, w2/df = 1.32, CFI = .99, RMSEA = .029, SRMR = .036. Correlational paths between mediators, error terms, and the effects of control variables are not shown for simplicity. *p < .05; **p < .01; ***p < .001.

well-being were significantly mediated by decreased ingroup identification (β = .10, SE = .03, p = .001, 90% CI [ .52, .16] and β = .05, SE = .02, p = .01, 90% CI [ .16, .04], respectively). The effects of negative contact on psychological well-being were not significantly mediated by relative deprivation, although approaching significance (β = .02, SE = .01, p = .11, 90% CI [ .11, .00]). Further analyses demonstrated that the effects of perceived discrimination on collective action tendencies and psychological well-being were also mediated by ingroup identification (β = .12, SE = .03, p < .001, 90% CI [.08, .27], and β = .06, SE = .02, p = .004, %90 CI [.02, .08], respectively). Moreover, the sequential mediation from negative contact to perceived discrimination and from perceived discrimination to ingroup identification also seemed to be effective on collective action tendencies (β = .02, SE = .01, p = .07, 90% CI [.01, .10]) and psychological well-being (β = .01, SE = .004, p = .09, 90% CI [.00, .03]).

The Model With the Minority Group We started with a partially mediated model where positive and negative contact5 predicted ingroup identification, relative deprivation, and perceived discrimination, which in turn predicted outgroup attitudes, collective action tendencies, and psychological well-being. We considered age and socioeconomic status as control variables and included a priori correlational paths between the mediators, but omitted the path from perceived discrimination to ingroup identification, since perceived discrimination did not significantly predict ingroup identification in our preliminary analyses. The partially mediated model yielded a good fit, w2(130) = 198.98, w2/df = 1.53, CFI = .95, RMSEA = .059, SRMR = .052. The baseline model accepted demonstrated that positive contact was significantly and negatively associated with ingroup identification, perceived discrimination, and relative deprivation (β = .23, β = .25, and β = .25,

As our initial analyses demonstrated that positive and negative contact were not normally distributed (skewness = 1.18 for negative contact and .72 for positive contact, normality tests p < .001), we used the log-transformed (base 10) versions of these variables in further analyses.

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Figure 2. Final SEM model for the minority group. Standardized path coefficients are displayed. ID = Identification, RD = Relative Deprivation, PD = Perceived Discrimination, CA = Collective Action, PWB = Psychological well-being. Final Model Fit: w2(130) = 198.98, w2/df = 1.53, CFI = .95, RMSEA = .059, SRMR = .052. Correlational paths between mediators, error terms, and the effects of control variables are not shown for simplicity. **p < .01; ***p < .001.

respectively, all p = .01). On the other hand, negative contact was not significantly associated with our mediator variables. Ingroup identification was positively associated with both collective action tendencies (β = .30, p = .001) and psychological well-being (β = .25, p = .01), but did not predict outgroup attitudes. Relative deprivation, on the other hand, was significantly associated with higher levels of collective action tendencies (β = .33, p = .001), and lower levels of positive outgroup attitudes (β = .32, p = .001). While the association between relative deprivation and psychological well-being was not significant, perceived discrimination also did not significantly predict the outcome variables. A further investigation of the direct effects between contact and outcome variables indicated that while positive contact was significantly and positively associated with positive outgroup attitudes (β = .40, p < .001), negative contact was negatively, but non-significantly related to this variable (β = .06, p = .38). Similarly, positive contact, but not negative contact, was significantly and positively associated with psychological well-being (β = .23, p = .01). Direct associations between positive and negative contact and collective action tendencies were non-significant. Figure 2 displays the overall model for the Kurdish group. Social Psychology (2019), 50(1), 7–23

As regards indirect effects, the indirect effect of positive contact on outgroup attitudes was significant via decreased relative deprivation (β = .08, SE = .04, p = .04, 90% CI [0.60, 11.56]). The effects of positive contact on collective action tendencies were mediated via decreased ingroup identification (β = .07, SE = .03, p = .03, 90% CI [ .52, .05]) and decreased relative deprivation (β = .08, SE = .04, p = .04, 90% CI [ .71, .01]). The effect of positive contact on psychological well-being via decreased ingroup identification was also significant (β = .06, SE = .03, p = .05, 90% [ .27, .03]). Please see ESM 2 for bootstrapped results and ESM 3 for the final dataset.

Discussion Recent research in contact literature has increasingly concentrated on exploring the simultaneous effects of positive and negative contact on a more diverse range of outcomes, exploring all potential consequences of intergroup contact among majority and minority group members (e.g., Árnadóttir et al., 2018; Barlow, Paolini, et al., 2012; Hayward Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

et al., 2017; Paolini et al., 2010; Reimer et al., 2017; Visintin et al., 2017). In line with this new research avenue, we aimed to examine the role of both forms of contact on the intended, unintended, and relatively less known consequences of contact, examining the direct associations between positive and negative contact and outgroup attitudes, collective action tendencies, and psychological well-being, as well as testing these associations via ingroup identification, relative deprivation, and perceived discrimination. Moreover, we investigated these relationships among both majority and minority group members in a context of intergroup tension. Partially confirming our initial proposition about positive– negative contact effects on outgroup attitudes, among the majority group, we found both contact forms to predict outgroup attitudes in opposite directions, but positive contact to be a stronger predictor of attitudes. Consistently, only positive contact was a significant determinant of outgroup attitudes among the minority group. Moreover, we found that positive contact effects on attitudes were partially mediated by decreased sense of relative deprivation among the minorities, overall showing positive contact to be a more tangible precursor of attitudes than negative contact in the current intergroup setting. Although research has recently concentrated on negative contact as a more important aspect of intergroup relationships (e.g., Barlow, Paolini, et al., 2012), some research studies have shown that positive contact is still a major antecedent of attitudes (e.g., Pettigrew & Hewstone, 2017; Pettigrew et al., 2011). The role of positive contact may be relatively more critical than negative contact in conflict settings (Al Ramiah & Hewstone, 2013), since it is possible that in conflictual settings where both groups (especially the minority group) generally hold negative expectancies from the outgroup, positive contact may become more salient than negative contact and group members may pay more attention to positive intergroup experiences that disconfirm their initial preassumptions. In turn, positive contact may generalize to outgroup attitudes more easily (e.g., Brown & Hewstone, 2005; Graf et al., 2014; Paolini et al., 2010) and thereby predict outgroup attitudes more strongly. Confirming this, Paolini et al. (2014) demonstrated that positive contact may buffer the valence-salience effects in conflict-ridden settings. Findings are also critical in terms of demonstrating the role of positive–negative contact on minority group members’ outgroup attitudes, which has been only studied in relatively less conflictual intergroup contexts and shown negative contact to be either as critical as or more critical than positive contact (Árnadóttir et al., 2018; Hayward et al., 2017). Hence, for the Kurds who represent a relatively oppressed minority group in a conflictual setting, positive intergroup experiences may compensate the existence of negative contact experiences and function as Ó 2018 Hogrefe Publishing

a more important determinant of positive outgroup attitudes. We further found that direct associations between positive and negative contact and collective action tendencies were non-significant among both groups, partially contradicting our initial proposition, but in line with some of the previous research findings showing the lack of direct associations between contact and support for collective action among Turks and Kurds (Çakal et al., 2016). However, we found positive contact to have indirect effects on collective action tendencies; for both groups, positive contact, but not negative contact, was associated with lower ingroup identification (consistent with the deprovincialization hypothesis, Pettigrew, 2009), and ingroup identification, in turn, was related to greater levels of collective action tendencies. The same association was also significant via relative deprivation among the minority group. Therefore, positive contact seemed to have an indirect sedative effect on social change motivation via decreased ingroup identification among the majorities and decreased ingroup identification and relative deprivation among the minorities. To our knowledge, only one study so far has focused on positive–negative contact asymmetry on social change motivation and demonstrated that negative contact, but not positive contact, was associated with collective action through increased perceived discrimination among LGBT individuals (Reimer et al., 2017). We found the opposite among the Kurds recruited from a conflict-ridden context where intergroup inequality and social change motivation is relatively more noticeable. Similar to our findings in terms of outgroup attitudes, it is possible that, for both groups who generally display high levels of distrust, outgroup attribution of responsibility, and stereotypes (e.g., Bagci & Çelebi, 2017), the demobilizing effect of positive contact may be more critical than the mobilizing effect of negative contact. In line with this, we found only positive contact to be associated with lower perceived discrimination among the minority group; hence, positive contact is likely to have more influence than negative contact among a stigmatized community who already expect negative attitudes from the majority group. For the majority group, on the other hand, we also detected a sequential mediation in which negative contact increases perceived discrimination which leads to higher ingroup identification, which in turn leads to increased collective action tendencies, demonstrating the indirect mobilizing effect of negative contact among the Turks. This is an interesting finding showing both positive and negative contact to have indirect, but opposite effects on collective action tendencies, via different routes; while positive contact seems to reduce collective action tendencies by reducing ingroup identification, negative contact which reinforces ingroup identification through increased Social Psychology (2019), 50(1), 7–23

perceived discrimination may indirectly increase collective action tendencies. As regards associations between positive–negative contact and psychological well-being, we expected that both forms of contact would be related to psychological wellbeing in opposite directions. Previous research examining contact effects on psychological well-being is rare, but demonstrates that positive contact, in the form of crossgroup friendships, is likely to bolster psychological outcomes at the individual level (Bagci et al., 2014; Mendoza-Denton & Page-Gould, 2008). In line with our initial proposal, we found both positive and negative contact to be equally likely to predict this outcome among the majority group. On the other hand, we found only positive contact to be strongly and positively associated with psychological well-being among the minority group. This is consistent with our previous findings, which demonstrate the relative effectiveness of positive contact compared to negative contact among a stigmatized minority group recruited from a conflict area. We also observed positive contact to have indirect effects on psychological well-being via ingroup identification among both groups. More specifically, positive contact with the outgroup was related to reduced ingroup identification and ingroup identification was consequently positively associated with psychological well-being. These indirect effects are critical, since although positive contact has direct positive influences on well-being, it can also indirectly reduce psychological well-being by decreasing ingroup identification which has been suggested to be an important source of psychological well-being and self-esteem, especially among minority group members (e.g., Cameron, 1999). A further paradoxical finding is the fact that while negative contact had an indirect negative influence on well-being through increased relative deprivation among the majority group, it also had an indirect positive influence by increasing perceived discrimination which is positively related to higher ingroup identification, which in turn, promotes psychological well-being; this confirms the rejection-identification model which suggested that perceived discrimination would lead group members to identify more strongly with their ingroup and this would consequently foster psychological well-being (e.g., Branscombe et al., 1999). Overall, these findings indicate that both positive and negative contact may have a diverse set of influences on the psychological well-being of group members through different mechanisms. As opposed to our initial expectation, although for both groups contact predicted perceived discrimination, we failed to find any significant direct influence of perceived discrimination on our outcome measures. It is especially intriguing to note this lack of association for the minority group, given that numerous research papers evidenced its Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

positive association with collective action tendencies (e.g., Tropp et al., 2012) and its negative association with wellbeing (e.g., Schmitt et al., 2014). A possible explanation may be the inclusion of many mediating mechanisms in the same model simultaneously which could have underestimated the role of perceived discrimination on our outcome measures (since perceived discrimination was correlated with some of the dependent variables in zeroorder correlations). Another alternative may be the high prevalence of perceived discrimination which was above average among the minority group. Such high levels of perceived discrimination may have led Kurdish group members to be more resilient in the face of pervasive discrimination and thereby rendered them less vulnerable to its negative influences. Overall, our findings offer three new insights into intergroup contact research literature. First, findings are far away from providing evidence for the recently discussed positive–negative contact asymmetry hypothesis which assumes the relative strong influence of negative contact over positive contact (Barlow, Paolini, et al., 2012; Paolini et al., 2010); instead, we found an asymmetry favoring positive contact which was generally more strongly related to our outcome measures, especially to outgroup attitudes, indicating the need to study positive–negative contact effects on various intergroup settings including conflict-ridden ones. Second, we found positive contact (and negative contact among the majority) to be directly and indirectly associated with psychological well-being, which suggests that beyond the conventionally studied contact effects such as intergroup attitudes, positive intergroup experiences in conflict-ridden societies are also likely to be associated with positive psychological outcomes among both majority and minority group members. Third, we found that the role of contact is multidimensional and the distinction between positive and negative contact may help us understand various direct and indirect influences of contact on distinct outcomes at the intergroup, societal, and individual levels. Previous research has indicated that it is critical to study contact at the macro-level and how contact may be used to improve societies through social cohesion or social change (Pettigrew & Hewstone, 2017), but how ingroup members personally benefit from contact experiences at the psychological level may be a further avenue to discover in intergroup contact research. Therefore, we extend past research which has shown the differential effects of contact on attitudes and collective action tendencies (e.g., Wright & Lubensky, 2009), by highlighting the need to incorporate psychological benefits/costs of contact in this context. Some limitations of the current study should be addressed. Methodologically, our Kurdish sample size was relatively small, due to various constraints in finding minority group members and their resistance to participate in a Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

study with such a sensitive topic. Moreover, participants were recruited from a relatively stable area where individuals were often not directly exposed to intergroup violence, whereas findings may differ for Kurdish groups recruited from the East of Turkey where military encounters occur more frequently (Bagci et al., in press). Although research conducted in more representative Kurdish samples in Turkey has shown Kurds to display high levels of perceived interethnic conflict regardless their specific location (Bagci & Çelebi, 2017), more research is needed to explore differences based on interethnic conflict perception. Additionally, the cross-sectional nature of the findings provides little evidence for the causal relationships among the variables, and we were unable to assess the direction of the effects with such a single time point study. Previous research has shown that especially the association between contact and outgroup attitudes is bidirectional (e.g., Binder et al., 2009). The same bidirectional association may also exist between contact and collective action tendencies, as well as psychological well-being such that people with greater collective action tendencies and psychological well-being may be more likely to engage in positive contact and less likely to form negative contact. Future longitudinal research is needed to ascertain that our proposed mediational pathways are causal. Our findings showed that although perceived discrimination was relatively high among the Kurds, positive contact experiences were perceived to be higher than negative contact experiences. In fact, our preliminary analyses demonstrated positive contact experiences to be more common than negative ones among both groups, which is consistent with previous research showing the quantity of positive contact to compensate the stronger influence of negative contact on outgroup attitudes (Aberson, 2015; Graf et al., 2014). Nevertheless, it is interesting to see that positive contact is more frequent than negative contact especially in a conflictual setting where intergroup attitudes are fairly negative. One reason may be our use of a subjective measure of contact valence rather than a more objective behavioral measure (e.g., Aberson, 2015), which suggests that we may not ascertain whether perceived positive and negative contact experiences reflect actual positive and negative contact experiences. Future research may examine whether estimated contact valence is dependent on various situational factors such as intergroup membership salience and whether different measures of positive–negative contact produce the same results (e.g., Hayward et al., 2017). We also focused only on the affective aspect of outgroup attitudes and positive attitudes; previous research has shown that negative contact may be more important for the cognitive dimension of prejudice (Aberson, 2015) and related more closely to negative outgroup attitudes (Barlow, Paolini, et al., 2012; Hayward et al., 2017), which suggests Ó 2018 Hogrefe Publishing

the need to incorporate various aspects of attitudes in future research. Theoretically, the inclusion of cognitive processes in the examination of positive and negative contact effects in the current model would advance our understanding of how contact valence functions in a conflict society. We assumed that positive contact may be a more important determinant of our dependent measures, given that positive contact would be a more salient intergroup process in the context of conflict, yet our data cannot show whether the contactsalience effect which suggests the superiority of negative contact over positive contact has been reversed in the current setting (e.g., Paolini et al., 2010). Furthermore, we only considered mediating variables that are based on grouprelated processes, however interpersonal mediating mechanisms that could be enhanced by positive contact such as self-disclosure, empathy, and inclusion of the other to the self (e.g., Pettigrew & Tropp, 2008; Turner, Hewstone, & Voci, 2007) may be equally important, in particular, mediating the effects of contact on psychological well-being. Other classically studied contact mediators such as intergroup anxiety and perceived outgroup threat toward the ingroup may also be studied in this framework to explain how positive and negative contact are associated with outgroup attitudes, collective action tendencies, and psychological well-being (e.g., Çakal et al., 2016; Hayward et al., 2017). In summary, our findings highlight the importance of examining both types of contact – negative as well as positive – in relation to a diverse range of outcomes to fully illustrate whether contact is beneficial or detrimental for minority and majority group members in the context of intergroup conflict. We demonstrated both positive and (to a lesser degree) negative contact to have direct and indirect influences on positive outgroup attitudes, collective action tendencies, and psychological well-being through different routes. Future research may explore other mechanisms involved in these associations in various intergroup contexts.

Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1864-9335/a000355 ESM 1. Data (.docx) Materials in English and Turkish. ESM 2. Tables (.docx) Bootstrapped confidence intervals. ESM 3. Data (.csv) Final Data. Social Psychology (2019), 50(1), 7–23

References Aberson, C. L. (2015). Positive intergroup contact, negative intergroup contact, and threat as predictors of cognitive and affective dimensions of prejudice. Group Processes & Intergroup Relations, 18, 743–760. https://doi.org/10.1177/ 1368430214556699 Aberson, C. L., & Gaffney, A. M. (2009). An integrated threat model of explicit and implicit attitudes. European Journal of Social Psychology, 39, 808–830. https://doi.org/10.1002/ejsp.582 Allport, G. W. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley. Al Ramiah, A., & Hewstone, M. (2013). Intergroup contact as a tool for reducing, resolving, and preventing intergroup conflict: Evidence, limitations, and potential. American Psychologist, 68, 527–542. https://doi.org/10.1037/a0032603 Árnadóttir, K., Lolliot, S., Brown, R., & Hewstone, M. (2018). Positive and negative intergroup contact: Interaction not asymmetry. European Journal of Social Psychology, 48, 784– 800. https://doi.org/10.1002/ejsp.2365 Bagci, S. C., & Çelebi, E. (2017). Cross-group friendships and outgroup attitudes among Turkish–Kurdish ethnic groups: Does perceived interethnic conflict moderate the friendshipattitude link? Journal of Applied Social Psychology, 47, 59–73. https://doi.org/10.1111/jasp.12413 Bagci, S. C., & Çelebi, E. (2018). Are your cross-ethnic friends ethnic and/or national group identifiers? The role of own and perceived cross-ethnic friend’s identities on outgroup attitudes and multiculturalism. European Journal of Social Psychology, 48, O36–O50. https://doi.org/10.1002/ejsp.2278 Bagci, S. C., Rutland, A., Kumashiro, M., Smith, P. K., & Blumberg, H. (2014). Are minority status children's cross-ethnic friendships beneficial in a multiethnic context? British Journal of Developmental Psychology, 32, 107–115. https://doi.org/ 10.1111/bjdp.12028 Bagci, S. C., Stathi, S., & Piyale, E. (in press). Imagined contact facilitates acculturation, sometimes: Contradicting evidence from two socio-cultural settings. In Cultural Diversity and Ethnic Minority Psychology. Bagci, S. C., Turnuklu, A., & Bekmezci, E. (2018). Cross-group friendships and psychological well-being: A dual pathway through social integration and empowerment. British Journal of Social Psychology, 57, 773–792. https://doi.org/10.1111/ bjso.12267 Bagozzi, R. P., & Heatherton, T. F. (1994). A general approach to representing multifaceted personality constructs: Application to state self-esteem. Structural Equation Modeling, 1, 35–67. https://doi.org/10.1080/10705519409539961 Bagozzi, R. P., & Yi, Y. (2012). Specification, evaluation, and interpretation of structural equation models. Journal of the Academy of Marketing Science, 40, 8–34. https://doi.org/ 10.1007/s11747-011-0278-x Barlow, F. K., Louis, W. R., & Hewstone, M. (2009). Rejected! Cognitions of rejection and intergroup anxiety as mediators of the impact of cross-group friendships on prejudice. The British Journal of Social Psychology, 48, 389–405. https://doi.org/ 10.1348/014466608X387089 Barlow, F. K., Paolini, S., Pedersen, A., Hornsey, M. J., Radke, H. R., Harwood, J., . . . Sibley, C. G. (2012). The contact caveat: Negative contact predicts increased prejudice more than positive contact predicts reduced prejudice. Personality and Social Psychology Bulletin, 38, 1629–1643. https://doi.org/ 10.1177/0146167212457953 Barlow, F. K., Sibley, C. G., & Hornsey, M. J. (2012). Rejection as a call to arms: Inter-racial hostility and support for political

Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

action as outcomes of race-based rejection in majority and minority groups. The British Journal of Social Psychology, 51, 167–177. https://doi.org/10.1111/j.2044-8309.2011.02040.x Becker, J. C., Wright, S. C., Lubensky, M. E., & Zhou, S. (2013). Friend or ally: Whether cross-group contact undermines collective action depends on what advantaged group members say (or don’t say). Personality and Social Psychology Bulletin, 39, 442–455. https://doi.org/10.1177/0146167213477155 Bekhuis, H., Ruiter, S., & Coenders, M. (2013). Xenophobia among youngsters: The effect of inter-ethnic contact. European Sociological Review, 29, 229–242. https://doi.org/10.1093/esr/ jcr057 Bilali, R., Çelik, A. B., & Ok, E. (2014). Psychological asymmetry in minority–majority relations at different stages of ethnic conflict. International Journal of Intercultural Relations, 43, 253– 264. https://doi.org/10.1016/j.ijintrel.2014.09.002 Binder, J., Zagefka, H., Brown, R., Funke, F., Kessler, T., Mummendey, A., . . . Leyens, J. P. (2009). Does contact reduce prejudice or does prejudice reduce contact? A longitudinal test of the contact hypothesis among majority and minority groups in three European countries. Journal of Personality and Social Psychology, 96, 843–856. https://doi.org/10.1037/a0013470 Branscombe, N. R., Schmitt, M. T., & Harvey, R. D. (1999). Perceiving pervasive discrimination among African Americans: Implications for group identification and well-being. Journal of Personality and Social Psychology, 77, 135–149. https://doi. org/10.1037/0022-3514.77.1.135 Brown, R., & Hewstone, M. (2005). An integrative theory of intergroup contact. Advances in Experimental Social Psychology, 37, 255–343. https://doi.org/10.1016/S0065-2601(05) 37005-5 Çakal, H., Hewstone, M., Güler, M., & Heath, A. (2016). Predicting support for collective action in the conflict between Turks and Kurds: Perceived threats as a mediator of intergroup contact and social identity. Group Processes & Intergroup Relations, 19, 732–752. https://doi.org/10.1177/1368430216641303 Cakal, H., Hewstone, M., Schwär, G., & Heath, A. (2011). An investigation of the social identity model of collective action and the ‘sedative’ effect of intergroup contact among Black and White students in South Africa. The British Journal of Social Psychology, 50, 606–627. https://doi.org/10.1111/j.20448309.2011.02075.x Cameron, J. E. (1999). Social identity and the pursuit of possible selves: Implications for the psychological well-being of university students. Group Dynamics: Theory, Research, and Practice, 3, 179–189. https://doi.org/10.1037/1089-2699.3.3. 179 Çelebi, E., Verkuyten, M., & Bagci, S. C. (2017). Ethnic identification, discrimination, and mental and physical health among Syrian refugees: The moderating role of identity needs. European Journal of Social Psychology, 47, 832–843. https://doi.org/ 10.1002/ejsp.2299 Çelebi, E., Verkuyten, M., Köse, T., & Maliepaard, M. (2014). Outgroup trust and conflict understandings: The perspective of Turks and Kurds in Turkey. International Journal of Intercultural Relations, 40, 64–75. https://doi.org/10.1016/j.ijintrel.2014. 02.002 Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D., Oishi, S., & Biswas-Diener, R. (2010). New well-being measures: Short scales to assess flourishing and positive and negative feelings. Social Indicators Research, 97, 143–156. https://doi.org/ 10.1007/s11205-009-9493-y Dixon, J., Durrheim, K., & Tredoux, C. (2005). Beyond the optimal contact strategy: A reality check for the contact hypothesis. American Psychologist, 60, 697–711. https://doi.org/10.1037/ 0003-066X.60.7.697

Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Dixon, J., Tropp, L. R., Durrheim, K., & Tredoux, C. (2010). “Let them eat harmony” prejudice-reduction strategies and attitudes of historically disadvantaged groups. Current Directions in Psychological Science, 19, 76–80. https://doi.org/10.1177/ 0963721410363366 Ellison, C. G., & Powers, D. A. (1994). The contact hypothesis and racial attitudes among Black Americans. Social Science Quarterly, 75, 385–400. Esses, V. M., Haddock, G., & Zanna, M. P. (1993). Values, stereotypes, and emotions as determinants of intergroup attitudes. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition and stereotyping: Interactive processes in group perception (pp. 137–166). San Diego, CA: Academic Press. Finlay, K. A., & Stephan, W. G. (2000). Improving intergroup relations: The effects of empathy on racial attitudes. Journal of Applied Social Psychology, 30, 1720–1737. https://doi.org/ 10.1111/j.1559-1816.2000.tb02464.x Göçek, F. M. (2011). The transformation of Turkey: Redefining state and society from the Ottoman Empire to the modern era (Vol. 103) . London, UK: IB Tauris. Graf, S., Paolini, S., & Rubin, M. (2014). Negative intergroup contact is more influential, but positive intergroup contact is more common: Assessing contact prominence and contact prevalence in five Central European countries. European Journal of Social Psychology, 44, 536–547. https://doi.org/10.1002/ ejsp.2052 Guimond, S., & Dambrun, M. (2002). When prosperity breeds intergroup hostility: The effects of relative deprivation and relative gratification on prejudice. Personality and Social Psychology Bulletin, 28, 900–912. https://doi.org/10.1177/ 014616720202800704 Haslam, S. A., Jetten, J., Postmes, T., & Haslam, C. (2009). Social identity, health and well-being: An emerging agenda for applied psychology. Applied Psychology: An International Review, 58, 1– 23. https://doi.org/10.1111/j.1464-0597.2008.00379.x Hayward, L. E., Tropp, L. R., Hornsey, M. J., & Barlow, F. K. (2017). Toward a comprehensive understanding of intergroup contact: Descriptions and mediators of positive and negative contact among majority and minority groups. Personality and Social Psychology Bulletin, 43, 347–364. https://doi.org/10.1177/ 0146167216685291 Hewstone, M., Cairns, E., Voci, A., Hamberger, J., & Niens, U. (2006). Intergroup contact, forgiveness, and experience of “The Troubles” in Northern Ireland. Journal of Social Issues, 62, 99– 120. https://doi.org/10.1111/j.1540-4560.2006.00441.x Jasinskaja-Lahti, I., Liebkind, K., Jaakkola, M., & Reuter, A. (2006). Perceived discrimination, social support networks, and psychological well-being among three immigrant groups. Journal of Cross-Cultural Psychology, 37, 293–311. https://doi.org/ 10.1177/0022022106286925 Jetten, J., Branscombe, N., Schmitt, M., & Spears, R. (2001). Rebels with a cause: Group identification as a response to perceived discrimination from the mainstream. Personality and Social Psychology Bulletin, 27, 1204–1213. https://doi.org/ 10.1177/0146167201279012 Jetten, J., Haslam, C., & Haslam, S. A. (2012). The social cure: Identity, health and well-being. New York, NY: Psychology Press. Kauff, M., Green, E. G., Schmid, K., Hewstone, M., & Christ, O. (2016). Effects of majority members’ positive intergroup contact on minority members’ support for ingroup rights: Mobilizing or demobilizing effects? European Journal of Social Psychology, 46, 833–839. https://doi.org/10.1002/ejsp.2194 Kawabata, Y., & Crick, N. R. (2008). The role of cross-racial/ethnic friendships in social adjustment. Developmental Psychology, 44, 1177–1183. https://doi.org/10.1037/0012-1649.44.4.1177

Ó 2018 Hogrefe Publishing

Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York, NY: Guilford Press. Konda, (2011). Kurt Meselesinde Algı ve Beklentiler [Perceptions and Expectations in the Kurdish Issue]. Istanbul, Turkey: Iletisim Yayinlari. Levin, S., Van Laar, C., & Sidanius, J. (2003). The effects of ingroup and outgroup friendships on ethnic attitudes in college: A longitudinal study. Group Processes & Intergroup Relations, 6, 76–92. https://doi.org/10.1177/1368430203006001013 Marsh, H. W., Hau, K., & Wen, J. (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11, 320–341. https://doi.org/10.1207/s15328007sem1103_2 Mazziotta, A., Rohmann, A., Wright, S. C., Tezanos-Pinto, D., & Lutterbach, S. (2015). (How) does positive and negative extended cross-group contact predict direct cross-group contact and intergroup attitudes? European Journal of Social Psychology, 45, 653–667. https://doi.org/10.1002/ejsp.2110 Mendoza-Denton, R., & Page-Gould, E. (2008). Can cross-group friendships influence minority students' well-being at historically white universities? Psychological Science, 19, 933–939. https://doi.org/10.1111/j.1467-9280.2008.02179.x Mishra, S., & Carleton, R. N. (2015). Subjective relative deprivation is associated with poorer physical and mental health. Social Science & Medicine, 147, 144–149. https://doi.org/10.1016/ j.socscimed.2015.10.030 Muthén, L. K., & Muthén, B. O. (1998–2018). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén. Page-Gould, E., Mendoza-Denton, R., & Tropp, L. R. (2008). With a little help from my cross-group friend: Reducing anxiety in intergroup contexts through cross-group friendship. Journal of Personality and Social Psychology, 95, 1080–1094. https://doi. org/10.1037/0022-3514.95.5.1080 Paolini, S., Harwood, J., & Rubin, M. (2010). Negative intergroup contact makes group memberships salient: Explaining why intergroup conflict endures. Personality and Social Psychology Bulletin, 36, 1723–1738. https://doi.org/10.1177/ 0146167210388667 Paolini, S., Harwood, J., Rubin, M., Husnu, S., Joyce, N., & Hewstone, M. (2014). Positive and extensive intergroup contact in the past buffers against the disproportionate impact of negative contact in the present. European Journal of Social Psychology, 44, 548–562. https://doi.org/10.1002/ejsp.2029 Paolini, S., Hewstone, M., Cairns, E., & Voci, A. (2004). Effects of direct and indirect cross-group friendships on judgments of Catholics and Protestants in Northern Ireland: The mediating role of an anxiety-reduction mechanism. Personality and Social Psychology Bulletin, 30, 770–786. https://doi.org/10.1177/ 0146167203262848 Pettigrew, T. F. (1997). Generalized intergroup contact effects on prejudice. Personality and Social Psychology Bulletin, 23, 173– 185. https://doi.org/10.1177/0146167297232006 Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65–85. https://doi.org/10.1146/annurev. psych.49.1.65 Pettigrew, T. F. (2009). Secondary transfer effect of contact: Do intergroup contact effects spread to noncontacted outgroups? Social Psychology, 40, 55–65. https://doi.org/10.1027/18649335.40.2.55 Pettigrew, T. F. (2016). In pursuit of three theories: Authoritarianism, relative deprivation, and intergroup contact. Annual Review of Psychology, 67, 1–21. https://doi.org/10.1146/annurevpsych-122414-033327 Pettigrew, T. F., Christ, O., Wagner, U., Meertens, R. W., Van Dick, R., & Zick, A. (2008). Relative deprivation and intergroup

Social Psychology (2019), 50(1), 7–23

prejudice. Journal of Social Issues, 64, 385–401. https://doi. org/10.1111/j.1540-4560.2008.00567.x Pettigrew, T. F., & Hewstone, M. (2017). The single factor fallacy: Implications of missing critical variables from an analysis of intergroup contact theory. Social Issues and Policy Review, 11, 8–37. https://doi.org/10.1111/sipr.12026 Pettigrew, T. F., & Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90, 751–783. https://doi.org/10.1037/00223514.90.5.751 Pettigrew, T. F., & Tropp, L. R. (2008). How does intergroup contact reduce prejudice? Meta-analytic tests of three mediators. European Journal of Social Psychology, 38, 922–934. https://doi.org/10.1002/ejsp.504 Pettigrew, T. F., Tropp, L. R., Wagner, U., & Christ, O. (2011). Recent advances in intergroup contact theory. International Journal of Intercultural Relations, 35, 271–280. https://doi.org/ 10.1016/j.ijintrel.2011.03.001 Reimer, N. K., Becker, J. C., Benz, A., Christ, O., Dhont, K., Klocke, U, . . . Hewstone, M. (2017). Intergroup contact and social change: Implications of negative and positive contact for collective action in advantaged and disadvantaged groups. Personality and Social Psychology Bulletin, 43, 121–136. https://doi.org/10.1177/0146167216676478 Saab, R., Harb, C., & Moughalian, C. (2017). Intergroup contact as a predictor of violent and nonviolent collective action: Evidence from Syrian refugees and Lebanese nationals. Peace and Conflict: Journal of Peace Psychology, 23, 297–306. https:// doi.org/10.1037/pac0000234 Saguy, T., & Chernyak-Hai, L. (2012). Intergroup contact can undermine disadvantaged group members’ attributions to discrimination. Journal of Experimental Social Psychology, 48, 714–720. https://doi.org/10.1016/j.jesp.2012.01.003 Saguy, T., Schori-Eyal, N., Hasan-Aslih, S., Sobol, D., & Dovidio, J. F. (2016). The irony of harmony: Past and new development. In L. Vezzali & S. Stathi (Eds.), Intergroup contact theory: Recent developments and future directions (pp. 53–71). Abingdon, UK: Routledge. Saguy, T., Tausch, N., Dovidio, J. F., & Pratto, F. (2009). The irony of harmony: Intergroup contact can produce false expectations for equality. Psychological Science, 20, 114–121. https://doi. org/10.1111/j.1467-9280.2008.02261.x Schmitt, M. T., Branscombe, N. R., Postmes, T., & Garcia, A. (2014). The consequences of perceived discrimination for psychological well-being: A meta-analytic review. Psychological Bulletin, 140, 921–948. https://doi.org/10.1037/a0035754 Stephan, W. G., Boniecki, K. A., Ybarra, O., Bettencourt, A., Ervin, K. S., Jackson, L. A., . . . Renfro, C. L. (2002). The role of threats in the racial attitudes of Blacks and Whites. Personality and Social Psychology Bulletin, 28, 1242–1254. https://doi.org/ 10.1177/01461672022812009 Swart, H., Hewstone, M., Christ, O., & Voci, A. (2011). Affective mediators of intergroup contact: A three-wave longitudinal study in South Africa. Journal of Personality and Social Psychology, 101, 1221–1238. https://doi.org/10.1037/ a0024450 Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks-Cole. Tausch, N., Saguy, T., & Bryson, J. (2015). How does intergroup contact affect social change? Its impact on collective action and individual mobility intentions among members of a disadvantaged group. Journal of Social Issues, 71, 536–553. https:// doi.org/10.1111/josi.12127

Social Psychology (2019), 50(1), 7–23

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Telef, B. B. (2013). Psikolojik _Iyi Olusß Ölçeği (P_IOÖ): Türkçe uyarlama, geçerlik vegüvenirlik çalısßması [Psychological wellbeing scale: Adaptation to Turkish and study of reliability and validity]. Hacettepe Üniversitesi Eğitim Fakültesi Dergisi, 28, 374–384. Tropp, L. R. (2006). Stigma and intergroup contact among members of minority and majority status groups. In S. Levin & C. van Laar (Eds.), Stigma and group inequality: Social psychological perspectives (pp. 171–191). Mahwah, NJ: Erlbaum. Tropp, L. R., Hawi, D. R., O’Brien, T. C., Gheorghiu, M., Zetes, A., & Butz, D. A. (2017). Intergroup contact and the potential for post-conflict reconciliation: Studies in Northern Ireland and South Africa. Peace and Conflict: Journal of Peace Psychology, 23, 239–249. https://doi.org/10.1037/pac0000236 Tropp, L. R., Hawi, D. R., Van Laar, C., & Levin, S. (2012). Crossethnic friendships, perceived discrimination, and their effects on ethnic activism over time: A longitudinal investigation of three ethnic minority groups. British Journal of Social Psychology, 51, 257–272. https://doi.org/10.1111/j.2044-8309.2011. 02050.x Tropp, L. R., Mazziotta, A., & Wright, S. C. (2016). Recent developments in intergroup contact research: Affective processes, group status, and contact valence. In F. Barlow & C. Sibley (Eds.), Cambridge Handbook of the Psychology of Prejudice. Tropp, L. R., & Pettigrew, T. F. (2005). Relationships between intergroup contact and prejudice among minority and majority status groups. Psychological Science, 16, 951–957. https://doi. org/10.1111/j.1467-9280.2005.01643.x Turner, R. N., Hewstone, M., & Voci, A. (2007). Reducing explicit and implicit outgroup prejudice via direct and extended contact: The mediating role of self-disclosure and intergroup anxiety. Journal of Personality and Social Psychology, 93, 369– 388. https://doi.org/10.1037/0022-3514.93.3.369 Uluğ, Ö. M., & Cohrs, J. C. (2017). “If we become friends, maybe I can change my perspective:” Intergroup contact, endorsement of conflict narratives, and peace-related attitudes in Turkey. Peace and Conflict: Journal of Peace Psychology, 23, 278–287. https://doi.org/10.1037/pac0000216 Van Zomeren, M., Postmes, T., & Spears, R. (2008). Toward an integrative social identity model of collective action: A quantitative research synthesis of three socio-psychological perspectives. Psychological Bulletin, 134, 504–535. https://doi.org/ 10.1037/0033-2909.134.4.504 Verkuyten, M. (2007). Religious group identification and interreligious relations: A study Among Turkish Dutch Muslims. Group Processes & Intergroup Relations, 10, 341–357. https:// doi.org/10.1177/1368430207078695 Verkuyten, M., Thijs, J., & Bekhuis, H. (2010). Intergroup contact and ingroup reappraisal: Examining the deprovincialization thesis. Social Psychology Quarterly, 73, 398–416. https://doi. org/10.1177/0190272510389015 Vignoles, V. L. (2011). Identity motives. In S. J. Schwartz, K. Luyckx, & V. L. Vignoles (Eds.), Handbook of identity theory and research (pp. 403–432). New York, NY: Springer. Visintin, E. P., Green, E. G., Pereira, A., & Miteva, P. (2017). How positive and negative contact relate to attitudes towards Roma: Comparing majority and high-status minority perspectives. Journal of Community & Applied Social Psychology, 27, 240– 252. https://doi.org/10.1002/casp.2309 Voci, A., Hadziosmanovic, E., Cakal, H., Veneziani, C. A., & Hewstone, M. (2017). Impact of pre-war and post-war intergroup contact on intergroup relations and mental health: Evidence from a Bosnian sample. Peace and Conflict: Journal of Peace Psychology, 23, 250–259. https://doi.org/ 10.1037/pac0000222

Ó 2018 Hogrefe Publishing

S. C. Bagci & A. Turnuklu, Positive and Negative Contact

Von der Heidt, T., & Scott, D. R. (2007, December). Partial aggregation for complex structural equation modeling (SEM) and small sample sizes: An illustration using a multi-stakeholder model of cooperative interorganisational relationships (IORs) in product innovation. Paper presented at the 21st ANZAM 2007 Conference, Sydney. Wright, S. C., & Lubensky, M. E. (2009). The struggle for equality: Collective action versus prejudice reduction. In S. Demoulin, J. P. Leyens, & J. F. Dovidio (Eds.), Intergroup misunderstandings: Impact of divergent social realities (pp. 291–310). New York, NY: Psychology Press.

Authorship All authors were involved in the design of the study; Abbas Turnuklu collected the data and Sabahat C. Bagci wrote the manuscript. ORCID Sabahat Bagci https://orcid.org/0000-0003-1642-2067

History Received February 26, 2018 Revision received July 3, 2018 Accepted July 3, 2018 Published online December 10, 2018 Sabahat C. Bagci Department of Psychology Isik University S ß ile/_Istanbul, 34980 Turkey cigdem.bagci@isikun.edu.tr

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 7–23

Original Article

Meta-Accuracy and Perceived Reciprocity From the PerceptionMeta-Perception Social Relations Model Ben Porter,1 Camilla S. Øverup,2 Julie A. Brunson,3 and Paras D. Mehta4,5 1

San Antonio, TX, USA

School of Psychology, Fairleigh Dickinson University, Teaneck, NJ, USA

Department of Psychology, Pennsylvania State University, University Park, PA, USA

Department of Psychology, University of Houston, Houston, TX, USA

Texas Institute for Measurement, Evaluation, and Statistics (TIMES), Houston, TX, USA

Abstract: Meta-accuracy and perceptions of reciprocity can be measured by covariances between latent variables in two social relations models examining perception and meta-perception. We propose a single unified model called the Perception-Meta-Perception Social Relations Model (PM-SRM). This model simultaneously estimates all possible parameters to provide a more complete understanding of the relationships between perception and meta-perception. We describe the components of the PM-SRM and present two pedagogical examples with code, openly available on https://osf.io/4ag5m. Using a new package in R (xxM), we estimated the model using multilevel structural equation modeling which provides an approachable and flexible framework for evaluating the PM-SRM. Further, we discuss possible expansions to the PM-SRM which can explore novel and exciting hypotheses. Keywords: social relations model, perception, meta-perception, perceived reciprocity, meta-accuracy

Perceptions of others within our social environment influence how we interact with them. Perceptions can concern attributes of others (e.g., she is nice) or the perceptions that others hold, called meta-perceptions (e.g., he thinks I am nice). Perceptions and meta-perceptions promote awareness of one’s social role, and both are necessary to successfully navigate one’s social environment. Perceptions and meta-perceptions can have complex interrelations as one can inform the other and both can be accurate or inaccurate. Furthermore, perceptions and meta-perceptions can be affected by the person doing the perceiving as well as the person being perceived. Studies of perception and meta-perception sometimes utilize the social relations model (SRM; Kenny, 1994) which decomposes multiple ratings into several interpretable components; covariances between these components can provide insight into various interpersonal concepts. Kenny (1994) discusses interpretations of these parameters in detail. Accuracy of meta-perceptions and assumptions of reciprocity can be examined with the SRM, but only when SRMs of a perceptual and meta-perceptual item with parallel wording are examined together. Although these paramSocial Psychology (2019), 50(1), 24–37 https://doi.org/10.1027/1864-9335/a000356

eters were described by Kenny (1994), researchers have infrequently used such models to examine accuracy of meta-perceptions and assumptions of reciprocity. Often when examining these constructs, researchers report only one of several important parameters (e.g., reporting dyadic perceived reciprocity but not generalized perceived reciprocity; e.g., Elfenbein, Eisenkraft, & Ding, 2009). Examining and presenting all parameters are important because they provide a rich description of the relationship between perception and meta-perception. Excluding parameters either intentionally or by accident can prevent readers from gaining a complete understanding of the interrelations between perceptions and meta-perceptions. A single complete model would prevent such exclusions and can be estimated using N-level structural equations modeling (NL-SEM). We call this unified model the Perception-Meta-Perception Social Relations Model (PM-SRM; Figure 1). Only two items must be asked of participants to estimate a PM-SRM: another-perception item and a metaperception item. For another-perception item, a general stem is “I think person X is _____” and for a meta-perception item, a general stem is “Person X thinks I am _____,” where Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

Figure 1. PM-SRM with paths labeled.

“person X” is the name of another participant and ______ is some attribute. For example, when rating Samantha on extraversion, participants would be presented with two items: “I think Samantha is extraverted” and “Samantha thinks I am extraverted.” This model is a specific example of the more general bivariate SRM (Mehta, 2018), but parallel item content creates theoretically important interpretations of covariance parameters. The current manuscript is intended to serve as a guide introducing the unified PM-SRM and the NL-SEM framework to researchers interested in meta-accuracy and perceptions of reciprocity. Although the parameters from the PM-SRM itself are fairly well known (Kenny, 1994), the NL-SEM modeling framework significantly expands Ó 2018 Hogrefe Publishing

the realm of testable hypotheses. As such, this guide may be useful even for experienced researchers. In this manuscript, we describe the interpretation of each of the parameters of the PM-SRM which may be redundant for experienced researchers because the basis for this unified model has been described decades before. Two practical examples are then provided in their entirety to demonstrate the use of the model and the utility of the PM-SRM in the interpretation of findings. Electronic Supplementary Material 1 (ESM 1) has annotated scripts for estimating the model in xxM, a freely available package for estimating NL-SEM in R (Mehta, 2013). Moreover, executable code and data are posted at https://osf.io/4ag5m, along with various other helpful information. Finally, future Social Psychology (2019), 50(1), 24–37

applications and adaptations of the PM-SRM are discussed to show the flexibility of the NL-SEM modeling framework.

N-Level Structural Equations Modeling NL-SEM is a form of Multilevel Structural Equations Modeling that does not restrict the number of levels that can be estimated (Mehta, 2018). As such, it is the intersection of structural equation modeling (SEM), which estimates associations of variables (i.e., columns), and multilevel modeling (MLM), which accounts for associations between observations (i.e., rows). Past work has demonstrated that multilevel models can be incorporated and specified in structural equation modeling terms (Mehta & Neale, 2005). MLM relies on manifest (observed) variables; however, by applying a structural equation modeling framework, researchers are able to examine interrelations between both latent and manifest variables both within and across levels (e.g., Mehta & Neale, 2005; Muthén, 1994). Unfortunately, common software packages often limit the user to two-level models (Hox, 2013; Mehta, 2018). In the case of the SRM, dependency often exists along both rows and columns, and for that reason, neither SEM nor MLM can adequately account for the data structure individually (Mehta, 2018). Thus, employing an NLSEM approach allows for more flexible estimation of the complex interdependency inherent to the data. An NLSEM framework can be used to estimate the PM-SRM because of the complex interrelations between items’ content and response dependency. Others have explained ML-SEM in detail (e.g., Goldstein, Bonnet, & Rocher, 2007; Hox, 2013; Rabe-Hesketh, Skrondal, & Zheng, 2014), and a full coverage is beyond the scope of the current manuscript. Mehta (2018) demonstrates the utility of conceptualizing the SRM using an NL-SEM framework. Using NL-SEM to estimate the PMSRM may be useful to researchers because (1) it is specified using language similar to SEM (which more researchers are familiar with), including latent variables and measurement models, (2) it provides methods for testing complex hypotheses and performing model comparisons, (3) it provides a complete model (i.e., all parameters are simultaneously estimated), and (4) missing data are accommodated through the application of full information maximum likelihood (FIML) estimation.

Social Relations Model – Univariate The social relations model (SRM; Kenny, 1994) decomposes ratings from a round-robin design into five separate Social Psychology (2019), 50(1), 24–37

B. Porter et al., PM-SRM

parameters. The meaning of these parameters depends on the format of the question. A brief description of the univariate components follows; ESM 2 (also posted on https://osf.io/4ag5m) explains the meaning of each component of other-perceptions and meta-perceptions in greater detail. Kenny (1994) provides a thorough explanation of these parameters and SRMs in general. For traditional other-perceptions (“I think Person X is ____”), the meaning of the components has a rather straightforward interpretation. These components examine whether people tend to perceive others in consistent ways (e.g., individuals may tend to see others universally as extraverted or introverted regardless of whom they are rating) or whether there is consensus in how people are perceived by others (e.g., individuals are universally seen as extroverted or introverted). These two pieces of information may be related, such that there is correspondence or reciprocity between the two constructs (e.g., participants who are seen as extraverted may see others as extraverted [positive covariance] or introverted [negative covariance]). Additionally, we can examine whether perceptions are unique to the specific dyad (i.e., whether there is variance in the ratings that is due to the specific dyad configuration) and whether members of a dyad reciprocate such perceptions (e.g., if Bob sees Samantha as extraverted, is she likely to see him as extraverted?) For meta-perceptions, the meanings are, in our opinion, more difficult to grasp. Again, we can decompose ratings into five pieces of information. We can examine whether people think that others see them consistently (e.g., individuals believe they are universally seen as introverted or extraverted) and whether people think that a person sees others consistently (e.g., some are universally thought to see others as either introverted or extraverted). These components may also covary, capturing potential similarities in how people think others see them and how people are generally thought to see others. And again, meta-perceptive ratings may also be due to the specific dyad configuration, and similarly, such meta-perceptions may be reciprocated (e.g., if Bob thinks Samantha sees him as extraverted then she is likely to think Bob sees her as extraverted).

A Quick Note on Terminology Traditionally, SRMs rely on the terms “actor” and “partner” to describe the components of a single univariate SRM (see Kenny, 1994). These terms are general enough that they do not need to be adapted for use in different contexts and have even been used in animal behavior studies (e.g., Kwan, Gosling, & John, 2008; Malloy, Barcelos, Arruda, DeRosa, & Fonseca, 2005). However, we have found that such terms can become confusing when considering the Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

interpretation of several of the SRM components (particularly with the meta-SRM). To ease understanding, we instead use the terms “rater” and “target” to describe the roles that individuals play for each response, because the vast majority, if not all PM-SRMs, will use self-reported data. For example, if Bob is rating Samantha’s intelligence, then Bob is the rater and Samantha is the target. Similarly for meta-perceptual questions, if Bob is rating what he believes Samantha thinks of his own intelligence, Bob will still be the rater because he is still the one making a rating and Samantha will still be the target because the rating is of an attribute of Samantha’s (in this case, her rating of Bob’s intelligence). The roles that participants play for each response will have corresponding rater effects and target effects which correspond to the previously used terms actor effects/perceiver variance and partner effects/target variance.

others’ ratings about the self. A strong positive correlation indicates high correspondence in how individuals think others rate them and how they rate others. For example, this association would be strong and positive if individuals who think they are seen as unkind by others also rate others as being unkind. Unlike the other covariances in the PM-SRM, this covariance is composed of two sets of ratings completed by the same person and is estimated at the person level. Because of this attribute, this covariance will be affected by individual differences in participants’ use of the rating scale (i.e., tendency to use the low or high end of a scale irrespective of their thoughts of the target). This will upwardly bias this parameter. Clear anchors (e.g., labeling the middle point “average”) may alleviate this issue.

Generalized Assumed Reciprocity (Target to Meta-Target Covariance)

The PM-SRM When estimating the PM-SRM, four additional covariances at the person level and two additional covariances at the dyadic level can be estimated. These parameters can be used to test hypotheses that cannot be evaluated within a univariate SRM. The covariances tap different aspects of meta-accuracy and perceived reciprocity. Table 1 presents an overview of all effects and associations estimated within the PM-SRM, and Figure 1 provides an illustration of the model. However, in the following sections, we describe and provide further interpretation of the covariances from the PM-SRM. Interpretations of some parameters can inform understanding of other parameters. For example, meta-accuracy of affection may be low because individuals hold false beliefs that others will reciprocate their feelings. Of note, if a variance component (e.g., a target or rater effect) is not significantly different from zero, the parameter estimates of the three covariances associated with that variance component are unreliable and should be interpreted with great caution or not at all.

Perceiver Assumed Reciprocity (Rater to Meta-Rater Covariance) Perceiver assumed reciprocity is measured by the covariance between the traditional rater effect and the meta-rater effect (Kenny, 1994). Conceptually, this covariance captures the relationship between how individuals view others and how individuals believe others view them. It reflects an internal process that concerns the similarities or differences between one’s own ratings and one’s beliefs about Ó 2018 Hogrefe Publishing

Generalized assumed reciprocity is the covariance between the traditional target effect and the meta-target effect. Conceptually, this covariance is between how people are viewed by others and how others believe those people view them. In contrast to perceiver assumed reciprocity, the reciprocity demonstrated in this case is not encompassed within the particular individual making the rating but rather is estimated using multiple participants’ ratings, so differences in the use of rating scales will not affect this covariance. This covariance will be positive and strong if individuals are viewed to rate the group in the same way that the group rates the individual. For example, a positive generalized assumed reciprocity effect would indicate a general agreement among people that individuals generally seen as kind tend to see more kindness in others and that those who are universally rated as unkind are thought to view others as unkind.

Generalized Meta-Accuracy (Target to Meta-Rater Covariance) Generalized meta-accuracy is captured by the covariance between the target effect from a traditional SRM and the rater effect from a meta-SRM. This covariance captures the association between how individuals believe they are generally viewed and how they are actually viewed. Of the three covariances that capture meta-accuracy, generalized meta-accuracy has been studied to a much greater extent than other PM-SRM components, typically under the non-specific name “meta-accuracy.” This may be because generalized meta-accuracy is more intuitive to understand: An individual’s accurate knowledge of how Social Psychology (2019), 50(1), 24–37

B. Porter et al., PM-SRM

Table 1. Overview of parameters estimated in the PM-SRM Main components of the SRM Conceptual name

Effect name

Conceptual meaning

Traditional SRM (Item format: “I see ____ as extraverted”) –

Dyadic

Assimilation

Rater

Represents the extent to which people hold idiosyncratic views of each other. The effect also contains error variance. The extent to which people tend to view others in a similar fashion.

Consensus

Target

The extent to which a person is viewed similarly by others.

Dyadic Reciprocity

Dyadic association

Generalized Reciprocity

The extent to which one person’s rating of another is reciprocated (i.e., the other person provides a similar rating of the person). Rater–Target association Association between peoples’ tendency to rate others in a certain way and whether they are seen by others in a similar (or dissimilar) way. Meta-SRM (Item format: “_____ sees me as extraverted”)

–

Dyadic

Projected Consensus

Rater

The extent to which people hold idiosyncratic views of each other’s views of the self. The effect also contains error variance. The extent to which people think others see them in a similar way.

Projected Assimilation

Target

The extent to which people are thought by others to see all others as alike.

Projected Reciprocity

Dyadic association

Projected Generalized Reciprocity

Rater–Target association

The extent to which there is reciprocity in ratings between people in terms of how they believe they see each other. Association between how an individual believes others rate him/her and how he/she is thought to rate others.

Conceptual name

Variable 1

Variable 2

Conceptual meaning.

Dyadic Meta-Accuracy

Dyadic A

Dyadic meta B

Assumed Dyadic Reciprocity

Dyadic A

Dyadic meta A

Perceiver Meta-Accuracy

Rater

Meta-target

Perceiver Assumed Reciprocity

Rater

Meta-rater

Generalized Meta-Accuracy

Target

Meta-rater

Generalized Assumed Reciprocity

Target

Meta-target

Association between how one is viewed idiosyncratically and how well one knows that idiosyncratic view. Association between how one rates a target and how one believes that same individual will rate one on that same trait. Association between how people are thought to view others and how people actually view others. Association between how people view others and how people believe others view them. Association between how people believe they are generally viewed and how they are actually viewed. Association between how people are viewed by others and how others believe those people view them.

Associations between components of the two SRMs

they are viewed by others is more in line with the core concept of meta-perception accuracy.

Perceiver Meta-Accuracy (Rater to Meta-Target Covariance) Perceiver meta-accuracy is captured by the covariance between the target effect from a meta-SRM and the rater effect from a traditional SRM. Conceptually, this is the association between how people are thought to view others and how those people actually view others. This type of metaaccuracy is driven by how well raters express or provide cues about the processes, feelings, and cognitions being queried. In other words, perceiver meta-accuracy represents the extent to which raters understand targets’ universal dispositions; thus, it is accuracy caused by the targets’ clear expression of their feelings and opinions. For example, Scrooge makes it well known how little he thinks of everyone’s work ethic. When asked to rate Scrooge’s perSocial Psychology (2019), 50(1), 24–37

ceptions of their work effort, his employees will assume Scrooge will rate their effort as poor. However, this is not due to an accurate perception of how they are seen, but rather an accurate perception of how Scrooge will rate others. This exemplifies the importance of expressiveness to perceiver meta-accuracy because the employees are correct in Scrooge’s assessment, but the accuracy comes from their knowledge of Scrooge and his well-shared general worldview. Of the types of meta-accuracy covered here, perceiver meta-accuracy has been studied the least in empirical research. One reason for this is that perceiver meta-accuracy is less intuitive to our understanding of meta-accuracy. Furthermore, meta-SRMs may not have a substantial amount of meta-target variance (e.g., Levesque, 1997). This means that often, there is not consensus in how targets will rate others; in other words, individuals do not distinguish between other’s differing worldviews and so these differing worldviews cannot inform meta-accuracy. Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

Assumed Dyadic Reciprocity (WithinPerson, Residual to Meta-Residual Covariance) In addition to person-level covariances, two novel dyadic covariances emerge with the use of the PM-SRM. The first covariance is assumed dyadic reciprocity. This is the idiosyncratic association between how one rates a target and how one believes that target will rate them. This covariance is called assumed dyadic reciprocity because if the effect is strong and positive, individuals generally believe that a partner holds the same unique view of them that they hold of the partner. High assumed dyadic reciprocity in extraversion would suggest that if Bob thinks Samantha is particularly extraverted, he also believes that she sees him as particularly extraverted too. It is important to point out that this association, like the others, is estimated controlling for all other effects. As such, assumed dyadic reciprocity is independent of rater and target effects and separate from assumed dyadic accuracy.

Dyadic Meta-Accuracy (Between-Person, Residual to Meta-Residual Covariance) The second covariance is dyadic meta-accuracy. This is captured by the association between the traditional SRM residual of one person and the meta-SRM residual from the other person. This is the association between how one is viewed idiosyncratically and how well one knows that idiosyncratic view; that is, if Bob thinks Samantha is extraverted, dyadic meta-accuracy exists if Samantha thinks Bob sees her as extraverted. This concept has also been called differential meta-accuracy to emphasize that after controlling for generalized meta-accuracy what remains is the unique association at the dyadic level (Carlson & Furr, 2009).

Examples Two examples of PM-SRMs are presented below using a sample of campus organizations whose members completed a questionnaire about other members of their group. Participants responded to both other-perception and metaperception formats. For the purpose of the present study, we selected two constructs: the first construct assessed extraversion, and the second construct assessed willingness to provide emotional support. We chose these items 1

because we theorized that the presence of significant components (e.g., target vs. meta-target components) may differ depending on whether the construct of interest concerns personality characteristics or more interpersonal, relational phenomena. That is, extraversion represents a universal personality characteristic for which some components may not be significant, while willingness to provide emotional support is a more interpersonal construct, for which we expected more components to be significant. Differences in the results between the two constructs provide a better understanding regarding the processes underlying meta-perception, particularly with respect to the context and the attribute in question. We first present the results of the PM-SRM for extraversion, followed by the results of the PM-SRM for emotional support.1

Method Participants were recruited from sororities, fraternities, and other social groups at a large university in the Southern United States. Research assistants attended group meetings during which surveys were administered to the participants. Participants rated up to five other members of their group on a battery of measures. Targets were randomly selected prior to data collection using the group roster. Participants were compensated with a $10 gift card. Seven campus organizations (mostly fraternities and sororities) participated in the study; some organizations were further divided based on years within their organization to enhance familiarity between raters and targets. Altogether, there were 114 raters, 164 targets, and 433 dyadic ratings. The median number of raters per target was 3, and the median number of targets per rater was 4. There are more targets than raters because individuals who did not attend the meeting could not serve as a rater but would have been rated by their peers due to the target pre-assignment based on the roster. Two constructs were selected for the current analyses. Extraversion was selected because it has previously been examined in SRM studies and, as a major personality construct, is of general interest to many researchers. The items measuring extraversion were adapted from the first item of the Ten-Item Personality Inventory (TIPI; Gosling, Rentfrow, & Swann, 2003): “I see this person as... extraverted, enthusiastic” and “This person sees me as... extraverted, enthusiastic.” We also examined emotional supportiveness because we believed it would it display more variability, particularly with regards to the meta-target effect. The items used to measure emotional support were developed for the current study: “I see this person as... willing to

The other-perception item for emotional support has been used in a manuscript that examined SRM components and interrelations and selfother-agreement among neuroticism, emotional support and life satisfaction. See Brunson et al. (2016).

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 24–37

provide emotional support in a time of need” and “This person sees me as... willing to provide emotional support in a time of need.” All extraversion and emotional support items were rated on 1- to 9-point scales where higher scores indicate more extraversion or emotional support. All analyses were performed with the R package xxM version 0.6.0 (Mehta, 2013, 2018) which can be downloaded from https://xxm.times.uh.edu/. Technical details and code for how to set up the data and conduct the analyses can be found in ESM 1. Furthermore, data and executable code for the univariate as well the PM-SRM can be found on https://osf.io/4ag5m. This program is built on a multilevel structural equation modeling framework and presents a flexible tool that allows for the estimation of cross-classified models including social relations models (Mehta, 2018).

Model Specification We strongly recommend that you download the supplemental materials at this time from https://osf.io/4ag5m or the ESM 1. The text in this section corresponds to the information provided in the supplemental guide that shows step-by-step how to set up the PM-SRM in xxM; thus, it may be helpful to review the guide at this point. In specifying the PM-SRM, a model containing two virtual levels (“person” and “dyad”) must be specified such that it connects different components with the same ID (e.g., connecting a person’s rater effect with the same person’s target effect). Within the person virtual level, levels for rater effects and target effects must be specified, while within the dyad virtual level, levels for the residual ratings must be specified. These levels allow xxM to store numerical information, including ID codes, observed data, and latent constructs. In the tutorial guide, setting up the initial model, including virtual levels and data containers, corresponds to Steps 1–5. Matrix notation similar to that of LISREL is used to estimate the means of observed variables (“nu” matrices), covariances/variances of observed variables (“theta” matrices), covariances/variances of latent variables (“psi” matrices), and latent variable loadings (“lambda” matrices). These matrices specify the relation between the variables that make up the PM-SRM (see Steps 6–12). Model output includes individual parameters as well as overall model statistics (e.g., 2 loglikelihood).

B. Porter et al., PM-SRM

Results The PM-SRM for Extraversion All parameters from the extraversion PM-SRM are presented in Table 2. The majority of other-perception variance was at the residual level (which included dyadic variance) (θ = 2.45 [2.04, 2.97]; 57% of total variance2). The target effect explained a substantial proportion of variance in the ratings (ψ = 1.27 [0.80, 1.89]; 29% of total variance), whereas the rater effect explained the smallest proportion of variance (ψ = 0.60 [0.28, 1.04]; 14% of total variance). This indicates that there is a degree of universal agreement about individuals’ levels of extraversion (target effect), but the majority of ratings are based on unique, idiosyncratic assessments (residual effect). Raters have a tendency to rate others similarly (rater effect), but this effect is relatively small. The results for the meta-perception item revealed that the rater effect accounted for the largest portion of variance in meta-perceptions (θ = 2.38 [1.72. 3.31]; 56% of total variance), followed by the residual effect (ψ = 1.76 [1.46, 2.02]; 42% of total variance). The nonsignificant target effect (ψ = 0.08 [0.001, 0.33]3; 2% of total variance) accounted for the least amount of variance in meta-perceptions. The large rater effect indicates that individuals may have a perceptual bias, thinking that everyone sees them similarly. As in other-perceptions, residual variance including dyadic variance remains important. However, the small amount of target variance in meta-perceptions indicates that people are not believed to perceive others as universally similar on extraversion. The covariances among the rater and target variances are also examined within the PM-SRM. The rater–target covariance for the traditional SRM (i.e., generalized reciprocity) was nonsignificant, which indicates that individuals are not rated similarly to their own ratings of others (ψ = 0.18 [ 0.20, 0.56], r = .21).4 Similarly, the nonsignificant covariance between the meta-rater and meta-target effects (i.e., the projected generalized reciprocity) shows that individuals who are thought to rate others as more extraverted are not more likely to think they will be rated as more extraverted (ψ = 0.13 [ 0.25, 0.51], r = .30). However, the nonsignificant projected generalized reciprocity is to be expected given the nonsignificant meta-target variance. Additionally, there is no evidence of dyadic reciprocity between the traditional residuals (θ = 0.46 [ 1.25, 0.38], r = .019) or projected dyadic reciprocity between

in component Percent of total variance is found via the following formula: percent of total variance ¼ totalvariance variance of all components. As an example, for the residual of 2:45 the traditional other-perceptions: 0:567 ¼ 2:45þ:60þ1:27. In order to remain defined, xxM restricts variances and residual variances to have a lower bound of .001. Thus, confidence intervals of variances and residual variances with a lower bound of .001 can be assumed to not be statistically different from zero. Raw variances and covariances can be used to calculate additional metrics by hand (e.g., correlation). For example, the generalized reciprocity covariance between rater and target 0:18 ffi ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi ¼ :21: correlation for other-perceptions is calculated by the following formula: r ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðvariance of rater variance of targetÞ

Social Psychology (2019), 50(1), 24–37

ð0:60 1:27Þ

Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

Table 2. The estimated parameters from an extraversion PM-SRM Parameter name

Estimate

Item: “I see ____ as extraverted,” M = 6.49 (other-perceptions) Rater

Assimilation

0.60

[0.28, 1.04]

Target

Consensus

1.27

[0.80, 1.89]

Residual Rater–target covariance

Generalized reciprocity

Dyadic covariance

Dyadic reciprocity

2.45

[2.04, 2.97]

0.18

[ 0.20, 0.56]

.21

0.46

[ 1.25, 0.38]

.02

Item: “_____ sees me as extraverted,” M = 6.32 (meta-perceptions) Rater

Projected consensus

2.38

[1.72. 3.31]

Target

Projected assimilation

0.08

[0.01, 0.33]

Residual Rater–target covariance

Projected generalized reciprocity

Dyadic covariance

Projected dyadic reciprocity

1.76

[1.46, 2.02]

0.13

[ 0.25, 0.51]

.30

0.07

[ 0.66, 0.52]

.04 .59

Across item-format covariances Rater

Meta-rater

Perceiver assumed reciprocity

0.70

[0.33, 1.18]

Target

Meta-target

Generalized assumed reciprocity

0.21

[ 0.03, 0.48]

.66

Rater

Meta-target

Perceiver meta-accuracy

0.01

[ 0.24, 0.24]

.05

Target

Meta-rater

Generalized meta-accuracy

0.92

[0.38, 1.52]

.53

Residual

Meta-residual

Assumed dyadic reciprocity

0.42

[0.15, 0.73]

.20

Residual

Meta-residual

Dyadic meta-accuracy

0.09

[ 0.43, 0.61]

.04

Note. Significance testing is done via 95% confidence intervals; bolded fonts reflect significance.

the meta-residuals (θ = 0.07 [ 0.66, 0.52], r = .04). Thus, within a given dyad, ratings do not relate and are not thought to relate. There was significant generalized meta-accuracy which indicates that individuals are aware of how extraverted they are generally seen to be (ψ = 0.92 [0.38, 1.52], r = .53). Even though there is not significant generalized reciprocity or projected generalized reciprocity, significant perceiver assumed reciprocity demonstrates that individuals who see others as extraverted believe they are rated as more extraverted as well (ψ = 0.70 [0.33, 1.18], r = .59). Interestingly, assumed reciprocity exists at the dyadic level (θ = 0.42 [0.15, 0.73], r = .20). There was no evidence of dyadic accuracy (θ = 0.09 [ 0.43, 0.61], r = .04); one reason may be because there was assumed dyadic reciprocity, but there was no evidence of actual dyadic reciprocity. Generalized assumed reciprocity (ψ = 0.21 [ 0.03, 0.48], r = .66) and perceiver meta-accuracy (ψ = 0.01 [ 0.24, 0.24], r = .05) should either not be interpreted or interpreted with caution due to unreliable estimates because the lack of significant variance in the meta-target effect. The Emotional Support PM-SRM All parameters from the emotional support PM-SRM are presented in Table 3. Unlike with extraversion, the rater, target and residual effect for both question formats were significant. In the other-perception format, most variance was captured by the residual effect (θ = 2.45 [2.04, 2.96]; 57%) with less variance attributable to the target (ψ = Ó 2018 Hogrefe Publishing

1.00 [0.48, 1.56]; 23%) and the rater (ψ = .84 [0.47, 1.36]; 20%). This supports the idea that special idiosyncratic relationships between individuals promote closeness and increased willingness to provide emotional support (Aron, Aron, Tudor, & Nelson, 1991). When examining the meta-perception format, the largest amount of variance was in the meta-residual effect (θ = 1.84 [1.52, 2.25]; 48%), which was closely followed by the variance accounted for by the meta-rater effect (ψ = 1.47 [0.99, 2.17]; 39%), with the least amount of variance attributed to the meta-target effect (ψ = 0.50 [0.23, 0.87]; 13%). The significant meta-rater variance indicates that individuals believe that others see them fairly consistent. Unlike with extraversion, a significant meta-target effect indicates that specific individuals are thought to see others consistently, yet again indicating the importance of the individual. However, the most important sources of variance are dyadic factors. Generalized reciprocity (i.e., rater–target covariance) was not significant (ψ = 0.32 [ 0.10, 0.76], r = .35) which indicates that individuals’ feelings of other’s supportiveness are not related to the level of support they themselves provide. Furthermore, the residual covariance (dyadic reciprocity) was not significant (θ = 0.21 [ 0.59, 1.01], r = .09). This indicates that within dyads, seeing a partner as especially supportive is not associated with being seen as especially supportive. However, there was significant projected generalized reciprocity (i.e., covariance between meta-rater and Social Psychology (2019), 50(1), 24–37

B. Porter et al., PM-SRM

Table 3. The estimated parameters from an emotional support PM-SRM Parameter name

Estimate

Item: “I see ____ as willing to provide emotional support in a time of need,” M = 6.94 (other-perceptions) Rater

Assimilation

0.84

[0.47, 1.36]

Target

Consensus

1.00

[0.48, 1.56]

Residual

2.45

[2.04, 2.96]

Rater–target covariance

Generalized reciprocity

0.32

[ 0.10, 0.76]

.35

Dyadic covariance

Dyadic reciprocity

0.21

[ 0.59, 1.01]

.09

Item format: “_____ sees me as willing to provide emotional support in a time of need,” M = 7.00 (meta-perceptions) Rater

Projected consensus

1.47

[0.99, 2.17]

Target

Projected assimilation

0.50

[0.23, 0.87]

Residual

1.84

[1.52, 2.25]

Rater–target covariance

Projected generalized reciprocity

0.54

[0.19, 0.93]

.63

Dyadic covariance

Projected dyadic reciprocity

0.32

[ 0.39, 0.98]

.17 .86

Across item-format covariances Rater

Meta-rater

Perceiver assumed reciprocity

0.96

[0.59, 1.46]

Target

Meta-target

Generalized assumed reciprocity

0.69

[0.41, 1.05]

.97

Rater

Meta-target

Perceiver meta-accuracy

0.21

[ 0.13, 0.55]

.32

Target

Meta-rater

Generalized meta-accuracy

0.88

[0.43, 1.40]

.72

Residual

Meta-residual

Assumed dyadic reciprocity

0.62

[0.34, 0.95]

.29

Residual

Meta-residual

Dyadic meta-accuracy

0.30

[ 0.29, 0.88]

.14

Note. Significance testing is done via 95% confidence intervals; bolded fonts reflect significance.

meta-target effect; ψ = 0.54 [0.19, 0.93], r = .63) which indicates that individuals who believe they are seen as supportive are thought to see others as supportive. However, there was no projected dyadic reciprocity (ψ = 0.32 [ 0.39, 0.98], r = .17), suggesting that individuals who thought the other person saw them as particularly supportive were not more likely to be considered supportive by the other person. There was evidence for significant perceiver assumed reciprocity (ψ = 0.96 [0.59, 1.46], r = .86), indicating that individuals who generally believe that others provide emotional support also believe others feel they provide emotional support. There was also support for generalized assumed reciprocity (ψ = 0.69 [0.41, 1.05], r = .97). This suggests that individuals who are considered willing to provide emotional support are also thought to believe they can get emotional support from others. Furthermore, there is significant assumed dyadic reciprocity (θ = 0.62 [0.34, 0.95], r = .29) which indicates that when individuals feel they can rely on a particular partner they believe the partner feels the same way. These results highlight how individuals assume there is a large degree of reciprocity, when the results indicated little actual reciprocity at either the generalized or dyadic level. Of the associations assessing accuracy, only generalized meta-accuracy was significant (ψ = 0.88 [0.43, 1.40], r = .72) indicating that individuals can accurately determine if they are seen as emotionally supportive. Perceiver metaaccuracy was not significant (ψ = 0.21 [ 0.13, 0.55], r = .32) indicating that people are inaccurate with respect to Social Psychology (2019), 50(1), 24–37

how they think other people see others. Moreover, dyadic meta-accuracy was also not significant (θ = 0.30 [ 0.29, 0.88], r = .14) suggesting that individuals were not able to determine whether a partner was particularly willing to provide emotional support to them, but assumed that there was mutual reciprocity (i.e., assumed dyadic reciprocity).

Interpretations and Comparisons of the Two Examples The two examples demonstrate the differential utility of the PM-SRM for different concepts. When examining extraversion, perceptions and meta-perceptions were fairly straightforward: There was support for the main components, such that some tended to see others as generally extraverted, and people generally agreed on who was extraverted versus introverted. Furthermore, people generally believed all others saw their extraversion similarly. Those who were seen by others as extraverted knew that they were seen as such. Additionally, individuals assumed they would be rated in a similar fashion as they rated others. This could either be due to perceptual biases or this could indicate that different individuals have different implicit scales. In this example, the use of the PM-SRM provided additional information about why dyadic meta-accuracy may be not have emerged. The results examining emotional support provide insight into the process of interpersonal support. Although individuals tended to have strong expectations of reciprocity with Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

regards to willingness to provide emotional support, reciprocity did not exist at either the person level or at the dyadic level. To some extent, as with extraversion, the high level of assumed reciprocity may be due to perceptual biases or differences in the use of scales. However, participants’ biased use of the scale could only explain perceiver assumed reciprocity, not generalized assumed reciprocity or dyadic assumed reciprocity, because both of these covariances contain others’ ratings. The high level of assumed reciprocity may indicate that individuals seeking to provide emotional support are seeking to gain social capital to be used in times of need. Yet, this tactic is unlikely to work successfully because those providing a high degree of social support both generally and in a particular dyad are not more likely to receive additional support in return. Although the majority of variance existed at the residual level, the discrepancy between assumed and actual reciprocity is likely a significant reason why dyadic metaaccuracy was not observed, an explanation which would have been absent without the PM-SRM. Contrary to our expectations, only generalized meta-accuracy was present; perceiver meta-accuracy was not significant. However, having either present is particularly interesting in the context of the lack of generalized reciprocity. It appears that individuals are particularly adept at determining which others are likely to be willing to provide support. This is likely a process that is not overtly displayed by individuals (e.g., announcing one’s willingness to assist), but rather is likely to be inferred (e.g., potentially by observing another’s interpersonal behavior). This process is important interpersonally, because regardless of what beliefs are held about reciprocity, it is important to know which sources of social support are viable. The emotional support results emphasize the utility of the additional information gained from using the PMSRM. For instance, if the univariate SRMs were analyzed separately, results would be fairly perplexing in that most of the variance exists at the residual level, but there is not any significant dyadic reciprocity or projected reciprocity. The residual variance may have been interpreted as random error, and more focus may have been placed upon the person-level findings. However, the significant assumed dyadic reciprocity correlation indicates that these results can inform our understanding of emotional support.

Discussion Perceptions of others is one of the most studied fields in social psychology, and it consists of many subfields, including prejudice and discrimination (attitudes), close relationships, and personality. Meta-perception is a related, but Ó 2018 Hogrefe Publishing

less studied area; a search on PsycINFO (conducted March 13, 2018) yielded a total of 192 published articles, book chapters, and dissertations that mention “meta-perception” or “metaperception” in their title, keywords or abstract. Just as with other-perceptions, the topics of meta-perception studies are varied (e.g., attractiveness and body image, goal achievement and motivation, personality). Many meta-perception studies claim to examine accuracy of perceptions; however, the vast majority of these manuscripts do not utilize SRM (see http://davidakenny.net/srm/srm.htm for a bibliography of studies that do use the SRM). The current paper describes the components of the PMSRM and details how such a model can be run and presented using NL-SEM, specifically using the R package xxM. Disattenuated estimates of the latent rater/target effects can be calculated in part by applying a measurement model approach, with the dyadic ratings serving as indicators of the latent rater/target traits. A measurement model approach can also be applied at the dyadic level, allowing for a separation of the true dyadic effect and measurement error. Additionally, this framework can accommodate the nesting of the levels; dyadic ratings are at a lower level than rater/target, which acts as the “parent” of the dyadic ratings. Additionally, by examining the complete model, all components are examined; many previous studies have focused on only specific associations, and analyses were performed with estimation of these associations only (e.g., only reporting generalized meta-accuracy). Examining all components in a unified model may have the potential to provide additional insight into the perceptual and meta-perceptual processes of the topic of interest. By only examining certain components of the model, these studies may have missed the opportunity to discover additional relevant findings which may have provided context to the findings (e.g., missing significant perceiver meta-accuracy due to only testing generalized meta-accuracy, or as occurred in the extraversion example, a lack of reciprocity and significant assumed reciprocity may have explained the low dyadic accuracy). Moreover, providing information about when components and associations are not significant may be of theoretical importance and may assist other researchers in developing hypotheses. Aspects of the constructs being studied may also affect the patterns of association between components in the model. Dyadic meta-accuracy may be more important for attributes that are interpersonal because there is likely to be more idiosyncratic differences between dyads. For example, Levesque (1997) found there was relatively little variance in ratings of personality traits, relative to relational constructs. Additionally, target variance from meta-perceptions is often not significant which precludes the estimation of perceiver meta-accuracy (e.g., Ho & Chau, 2009; Social Psychology (2019), 50(1), 24–37

Levesque, 1997). Target variance of meta-perception is likely higher in constructs that are interpersonally differentiated like friendships (because individuals generally know who likes/dislikes them) or ratings of interviewers (because interviewees likely devote more attention to the attitudes of interviewers). However, there is generally not enough published studies to do more than hypothesize about these relationships. Presenting these parameters, whether significant or not, may yield theoretical advances and understanding about how the construct studied affects the patterns demonstrated in PM-SRM. Moderating effects of these parameters may also be a fruitful research direction. In one study, meta-accuracy was higher after participants watched videotapes of interactions between themselves and others, presumably because they developed a better understanding the others’ perspective (Albright & Malloy, 1999). Additionally, generalized meta-accuracy has been found to increase with increasing levels of closeness (e.g., higher among family than among acquaintances; Malloy et al., 1997). Similarly, generalized and dyadic meta-accuracy was higher among parents than their children, suggesting that meta-accuracy skills may develop over time (Cook & Douglas, 1998). Finally, increased confidence in ratings was associated with increased meta-accuracy showing that participants, to an extent, are aware of what they do not know about others’ perceptions (Carlson, Furr, & Vazire, 2010). These components can also be important for theory building and practical applications. For example, in a workplace or romantic relationships, perceived reciprocity of contributions may be more important to individual or relational well-being than actual reciprocity (e.g., Lemay, Clark, & Feeney, 2007). Assessing perceived reciprocity may identify problems in these settings and potentially provide avenues to address perceived rather than actual reciprocity.

Methodological Consideration One reason we believe previous research may not have reported all possible model components is that SRMs of perception and meta-perception data have usually been done separately in univariate models. This may partly be due to somewhat limited analytical tools. SAS PROC MIXED code can be specified to fit an SRM model (Kenny & Livi, 2009), and software has been built (e.g., SOREMO), but these tools may be cumbersome and inflexible depending on the analyses to be performed. In this paper, we used the approach outlined by Mehta (2018), which relies on NLSEM. This framework is more flexible, and it allows for the simultaneous estimation of all components and associations. However, when examining the PM-SRM using Social Psychology (2019), 50(1), 24–37

B. Porter et al., PM-SRM

NL-SEM, issues relating to structural equation modeling (SEM) are still relevant. Model estimation may have convergence problems, particularly with a low number of ratings per rater or target. To prevent convergence issues, studies intending to use the PM-SRM should aim to have at least an average of three ratings for each rater and three ratings of each target. This is the minimum recommended, but more are ideal; Malloy and Kenny (1986) recommended at least four. In the SRM literature, there has been debate with regard to the optimal number of groups and people within groups. Some research has found that it may be more important to have a larger number of participants within-group than a large number of groups (Lüdtke, Robitzsch, Kenny, & Trautwein, 2013). However, the optimal decision for group constructions would depend on the sample, specific methodology, and resources available for the study. Additionally, more research into group size and construction are needed. A related issue is that of measurement reliability. Naturally, researchers worry about the reliability and validity of their measurements, often using multi-item scales, in which more items often yield increased reliability. However, multi-item scales may often suffer from monomethod bias and response set, which may artificially increase reliability. The SRM instead derives reliability from using multiple ratings and informants, allowing systematic variance to be more readily interpretable. This means that short scales or even single items form reliable measures (Mehta, 2018). As such, researchers using the PM-SRM should aim to use briefer measures as the traditionally longer measures would create a huge amount of burden for participants asked to rate multiple targets on several time-consuming scales. Moreover, because the SRM, as presented in the current manuscript, relies on a measurement model approach, estimates of rater and target effects are disattenuated. That is, while each individual only completes a single item (or a few items) about each other person, much more information goes into the estimation of rater/target effects. Thus, the SRM yields reliable measures with a reduced response burden on any given participant, as they may complete only a few items rather than long and repetitive scales. Given that SRMs typically require the use of round-robin designs, data collection may be difficult and the likelihood of missing data can be high. Observations with missing perception or meta-perception items can still be included into the xxM-specified model, as xxM relies on the use of FIML estimation. Similarly, unbalanced data where certain individuals either are not raters or not targets can be included in the model. In fact, this happened in the example data, as some potential participants were absent when data were collected but were included on the group roster. Additionally, the main unit of analysis for the PM-SRM is the Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

response rather than participant. Because each participant completes multiple ratings, a large number of responses are obtained even when collecting information from a relatively small number of participants. Collecting round-robin data can be difficult, but the effort of conducting such studies is well rewarded by rich data from any pair of items on the survey.

Construct Assessment To assess social psychological and personality constructs, researchers often rely on self-report multi-item inventories that may be time-consuming to complete. Additionally, selfreport can suffer from systematic biases (Back & Vazire, 2012), possibly stemming from people’s (perhaps erroneous) self-views, their views of others (with social comparison as a potential mechanism), as well as their meta-perceptions. These biases may cast doubts on the accuracy of people’s self-perceptions. Rather than have a single informant, multiple informants may be beneficial in establishing what might be considered a more objective measure of “true personality” (Back & Vazire, 2012). Often, studies rely on multiple sources of information that can be used to obtain a more objective estimate of the desired construct (e.g., a student’s performance rated by the student, teacher, and parent). Multi-trait-multi-method (MTMM) models exemplify such methods, as this class of designs separates systematic variance related to the construct of interest from measurement method or source, in order identify and control for measurement bias and systematic mono-method variance (Widaman, 1985). Similarly, the SRM method of assessment capitalizes on multiple ratings for each target and rater, distributed across multiple people. As such, rater and target components, and by extension, covariance estimates should be more nuanced than were they derived from a single source of self-reported data. The visibility of the construct of interest may also play an important role in PM-SRMs and may influence the accuracy of perceptions. Overall, research suggests that self-other agreement exists, particularly for extraversion (Kenny, 1994). Some have argued that self-other agreement is higher for more visible traits (e.g., extraversion, charm, being funny), while self-report may be more accurate for less visible traits, such as neuroticism (Back & Vazire, 2012; Funder & Dobroth, 1987). Interestingly, Brunson, Øverup, and Mehta (2016) found that people who saw themselves as more neurotic were also seen by others as more neurotic (Brunson et al., 2016), a trait often thought to be less visible (Funder & Dobroth, 1987); this study did, however, rely on a sample of acquainted individuals. The visibility of the trait may be less important to self-other agreement the longer people are acquainted.

Ó 2018 Hogrefe Publishing

Future Directions In using the NL-SEM framework, innumerable modeling possibilities emerge because models can be specified and tested using nested model comparisons. Researchers may be able to examine other variables (e.g., self-report, observational data) in conjunction with the PM-SRM. Back and Vazire (2012) suggest that the accuracy of perceived selfknowledge may be examined by how well self-reported personality converges with others’ perceptions of one’s personality (the target effect in the SRM). By adding self-report of the construct of interest (e.g., “I am . . . extraverted”) and examining associations with the other-perception target effect, researchers are able to examine whether self-perceptions correspond to others’ perceptions of the person. Indeed, the social relations model allows for an examination of self-other agreement, as past research has done (e.g., Brunson et al., 2016; Levesque & Kenny, 1993). Meta-insight is another concept that may be of interest; meta-insight is the ability of individuals to accurately predict others’ meta-perceptions controlling for one’s own self-perceptions (Carlson, Vazire, & Furr, 2011). By estimating covariances between the self-rating and all other latent variables at the person level, meta-insight would be captured by significant generalized meta-accuracy controlling for self-perception. Outside of computing time, requisite sample size, and a model which becomes increasingly complex, there is no practical limitation on the number of items that can be estimated simultaneously. For example, one might combine the PM-SRMs from the examples to examine the interplay between extraversion and emotional support. This single model would have 4 rater effects, 4 target effects, 4 residual effects, 6 residual covariances, and 28 person-level covariances (although such results can become unwieldy!). The model can also be expanded to examine generalized other-perception (i.e., “Others think person X is _____”) through simply adding a third variable, which would have its own rater, target, and residual component along with all accompanying covariances. A major benefit of using NL-SEM to estimate PM-SRM is that the model is flexible enough to test nuanced hypotheses by applying strategies used in SEM. Rather than examining single parameters, nested models can be used to make more general statements. For example, a universal test of metaaccuracy could be performed by comparing the PM-SRM to a restricted model with all three meta-accuracy paths restricted to zero. In a related vein, it is possible to construct latent variables at the dyadic level; this would allow for separating error variance from systematic dyadic variance, and it would provide a true estimate of dyadic reciprocity (Øverup et al., 2014). Structural paths can also be estimated between components and other person-level variables to estimate a number of different hypotheses (e.g., mediation).

Social Psychology (2019), 50(1), 24–37

In certain cases, researchers may only be interested in certain components of a model. For example, Cook and Douglas (1998) focused solely on children’s perceptions of their parents’ ratings and not parents’ perceptions of their children’s ratings. Participants within this model are distinguishable rather than indistinguishable (for a thorough discussion of distinguishability within a dyadic context, see Kenny, Kashy, & Cook, 2006). This can be accomplished by specifying separate latent rater and target variables for children and parents.

Conclusion The PM-SRM is a useful tool for researchers wishing to examine perceptual and meta-perceptual processes in groups because it provides rich and nuanced results that can guide researchers toward a better understanding of their construct. This understanding is potentially greater than that which could be obtained in a piecemeal fashion. Furthermore, the use of NL-SEM in estimating this model opens up variations in modeling and testing that can provide results specifically tailored to researchers’ objectives. With increased use of the PM-SRM, better theory may emerge about perceptions of reciprocity and meta-accuracy. The current manuscript is not meant to be an exhaustive guide for what is possible using this method but a starting point to make more researchers comfortable engaging with and customizing these models.

Open Data/Material All data used in this study are available at https://osf.io/ 4ag5m

Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1864-9335/a000356 ESM 1. Instructions (.docx) PM-SRM xxM Guide (v1.0). ESM 2. Text (.docx) Description of the traditional SRM.

References Albright, L., & Malloy, T. E. (1999). Self-observation of social behavior and metaperception. Journal of Personality and Social Psychology, 77, 726–734. https://doi.org/10.1037/00223514.77.4.726

Social Psychology (2019), 50(1), 24–37

B. Porter et al., PM-SRM

Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close relationships as including other in the self. Journal of Personality and Social Psychology, 60, 241–253. https://doi.org/ 10.1037/0022-3514.60.2.241 Back, M. D., & Vazire, S. (2012). Knowing our personality. In S. Vazire & T. D. Wilson (Eds.), Handbook of self knowledge (pp. 131–156). New York, NY: Guilford Press. Brunson, J. A., Øverup, C. S., & Mehta, P. D. (2016). A social relations examination of neuroticism and emotional support. Journal of Research in Personality, 63, 67–71. https://doi.org/ 10.1016/j.jrp.2016.05.012 Carlson, E. N., & Furr, R. M. (2009). Evidence of differential metaaccuracy: People understand the different impressions they make. Psychological Science, 20, 1033–1039. https://doi.org/ 10.1111/j.1467-9280.2009.02409.x Carlson, E. N., Furr, R. M., & Vazire, S. (2010). Do we know the first impressions we make? Evidence for idiographic meta-accuracy and calibration of first impressions. Social Psychological and Personality Science, 1, 94–98. https://doi.org/10.1177/ 1948550609356028 Carlson, E. N., Vazire, S., & Furr, R. M. (2011). Meta-insight: Do people really know how others see them? Journal of Personality and Social Psychology, 101, 831–846. https://doi.org/10.1037/ a0024297 Cook, W. L., & Douglas, E. M. (1998). The looking-glass self in family context: A social relations analysis. Journal of Family Psychology, 12, 299–309. https://doi.org/10.1037/08933200.12.3.299 Elfenbein, H. A., Eisenkraft, N., & Ding, W. W. (2009). Do we know who values us? Dyadic meta-accuracy in the perception of professional relationships. Psychological Science, 20, 1081– 1083. https://doi.org/10.1111/j.1467-9280.2009.02396.x Funder, D. C., & Dobroth, K. M. (1987). Differences between traits: Properties associated with interjudge agreement. Journal of Personality and Social Psychology, 52, 409–418. https://doi. org/10.1037/0022-3514.52.2.409 Goldstein, H., Bonnet, G., & Rocher, T. (2007). Multilevel structural equation models for the analysis of comparative data on educational performance. Journal of Educational and Behavioral Statistics, 32, 252–286. https://doi.org/10.3102/ 1076998606298042 Gosling, S. D., Rentfrow, P. J., & Swann, W. B. Jr. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37, 504–528. https://doi.org/10.1016/ S0092-6566(03)00046-1 Ho, D. Y. F., & Chau, A. W. L. (2009). Interpersonal perceptions and metaperceptions of relationship closeness, satisfaction and popularity: A relational and directional analysis. Asian Journal of Social Psychology, 12, 173–184. https://doi.org/10.1111/ j.1467-839X.2009.01283.x Hox, J. (2013). Multilevel regression and multilevel structural equation modeling. In T. Little (Ed.), Oxford handbook of quantitative methods in psychology: Vol. 2: Statistical analysis (pp. 281–294). Oxford, UK: Oxford University Press. Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York, NY: Guilford Press. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic Data Analysis. New York: Guilford Press. Kenny, D. A., & Livi, S. (2009). A componential analysis of leadership using the Social Relations Model. In F. J. Yammarino & F. Dansereau (Eds.), Multi-level issues in organizational behavior and leadership (Vol. 8 of Research in multi-level issues, pp. 147–191). Bingley, UK: Emerald. Kwan, V. S. Y., Gosling, S. D., & John, O. P. (2008). Anthropomorphism as a special case of social perception: A cross-species

Ó 2018 Hogrefe Publishing

B. Porter et al., PM-SRM

comparative approach and a new empirical paradigm. Social Cognition, 26, 129–142. Lemay, E. J., Clark, M. S., & Feeney, B. C. (2007). Projection of responsiveness to needs and the construction of satisfying communal relationships. Journal of Personality and Social Psychology, 92(5), 834–853. https://doi.org/10.1037/00223514.92.5.834 Levesque, M. J. (1997). Meta-accuracy among acquainted individuals: A social relations analysis of interpersonal perception and metaperception. Journal of Personality and Social Psychology, 72, 66–74. https://doi.org/10.1037/0022-3514.72.1.66 Levesque, M. J., & Kenny, D. A. (1993). Accuracy of behavioral predictions at zero acquaintance: A social relations analysis. Journal of Personality and Social Psychology, 65, 1178–1187. Lüdtke, O., Robitzsch, A., Kenny, D.A., & Trautwein, U. (2013). A general and flexible approach to estimating the social relations model using Bayesian methods. Psychological Methods, 18, 101–119. https://doi.org/10.1037/a0029252 Malloy, T. E., Albright, L., Kenny, D. A., Agatstein, F., & Winquist, L. (1997). Interpersonal perception and metaperception in nonoverlapping social groups. Journal of Personality and Social Psychology, 72, 390–398. https://doi.org/10.1037/0022-3514. 72.2.390 Malloy, T. E., Barcelos, S., Arruda, E., DeRosa, M., & Fonseca, C. (2005). Individual differences and cross-situational consistency of dyadic social behavior. Journal of Personality and Social Psychology, 89, 643–654. http://dx.doi.org/10.1037/00223514.89.4.643 Malloy, T.E., & Kenny, D.A. (1986). The Social Relations Model: An integrative method for personality research. Journal of Personality, 54, 199–225. Mehta, P. D. (2013). xxM User’s Guide. Retrieved from http:// xxm.times.uh.edu/documentation/xxm.pdf Mehta, P. D. (2018). Virtual level and role models: N-level structural equations model of reciprocal ratings data. Multivariate Behavioral Research, 53, 315–334. https://doi.org/ 10.1080/00273171.2018.1443787 Mehta, P. D., & Neale, M. C. (2005). People are variables too: Multilevel structural equations modeling. Psychological Methods, 10, 259–284. Muthén, B. (1994). Multilevel covariance structure analysis. Sociological Methods & Research, 22, 376–398. https://doi. org/10.1177/0049124194022003006 Øverup, C. S., Hadden, B. W., Brunson, J. A., Porter, B., Wickham, R. W., & Mehta, P. (September 2014). Selfless ‘Self-Determina-

Ó 2018 Hogrefe Publishing

tion Theory’: A five level latent variable social relations model of need supportiveness. In Paper presented at the SRCD Special Topic Meeting: Developmental Methodology, San Diego, CA. Rabe-Hesketh, S., Skrondal, A., & Zheng, X. (2014). Multilevel structural equation modeling. In S. Y. Lee (Ed.), Handbook on Structural Equation Models. Amsterdam, The Netherlands: Elsevier. Widaman, K. F. (1985). Hierarchically nested covariance structure models for multitrait-multimethod data. Applied Psychological Measurement, 9, 1–26. https://doi.org/10.1177/ 014662168500900101 History Received June 21, 2017 Revision received March 25, 2018 Accepted July 16, 2018 Published online December 10, 2018 Acknowledgments We would like to thank Dr. Robert Wickham for his valuable insight while developing the utility of the reciprocal SRM and his feedback on earlier drafts of the manuscript. Funding Development of xxM was made possible by a generous grant (R305D090024) from the Institute of Education Sciences, US Department of Education awarded to Paras Mehta. Authorship Drs. Porter, Øverup, and Brunson drafted the initial manuscript and electronic supplementary materials. Drs. Porter & Øverup conducted the analyses. Dr. Mehta collected the data. All authors contributed to the conceptualization of the study and critically revised all drafts of the manuscript. Ben Porter Department of Psychology University of Texas-San Antonio 10 Cocke Dr. San Antonio, TX 78249 USA bennett.w.porter@gmail.com

Social Psychology (2019), 50(1), 24–37

Original Article

You Can Leave Your Glasses on Glasses Can Increase Electoral Success Alexandra Fleischmann1 , Joris Lammers1, Janka I. Stoker2, and Harry Garretsen2 1

Social Cognition Center Cologne, University of Cologne, Köln, Germany

Faculty of Economics and Business, University of Groningen, Groningen, The Netherlands

Abstract: Does wearing glasses hurt or help politicians in elections? Although some research shows that glasses signal unattractiveness, glasses also increase perceptions of competence. In eight studies, participants voted for politicians wearing (photoshopped) glasses or not. Wearing glasses increased politicians’ electoral success in the US (Study 1), independent of their political orientation (Studies 2a and 2b). This positive effect was especially strong when intelligence was important (Study 3), and even occurred if glasses were used strategically (Study 4). However, it did not extend to India (Study 5) due to different cultural associations with glasses (Study 6). Furthermore, while intelligence mediated the effect, warmth did not (Study 7). In summary, wearing glasses can robustly boost electoral success, at least in Western cultures. Keywords: glasses, voting, stereotypes, politicians, election

Politicians appear to shun glasses. For example, even though presently around 61% of US citizens wear glasses, no US president since Harry Truman wore glasses in public (Pindell, 2015). After becoming the UK Prime Minister, David Cameron waited a full 3 months before wearing glasses in public and felt noticeably uncomfortable doing so (Swinford, 2013). Jeb Bush even attributed his failure to defeat Trump in the Republican primaries to his glasses (Jamieson, 2016). It seems politicians try to avoid wearing glasses in public because of possible associations with vision deficiency, old age, and weakness (Elman, 1977; Swinford, 2013; Terry & Krantz, 1993). But are politicians actually right in doing so? The current research aims to test this.

Effects of Glasses on Appearance Politicians’ appearance, in particular their facial appearance, certainly influences their electoral success (Antonakis & Dalgas, 2009; Jäckle & Metz, 2015; Little, Burriss, Jones, & Roberts, 2007; Todorov, Mandisodza, Goren, & Hall, 2005; Todorov, Olivola, Dotsch, & Mende-Siedlecki, 2015). For example, Todorov et al. (2005) found that in the US, 72% of the Senate races and 67% of the House races could be predicted by asking people who looked more competent. Given that glasses may signal declining health and weakness, politicians may be right to avoid wearing glasses in public. Indeed, some research appears to support politicians’ fears. Glasses are linked with a range of negative Social Psychology (2019), 50(1), 38–52 https://doi.org/10.1027/1864-9335/a000359

characteristics, such as being less likeable, less attractive, and most importantly less dominant (Edwards, 1987; Hasart & Hutchinson, 1993; Leder, Forster, & Gerger, 2011; Lundberg & Sheehan, 1994; Terry & Krantz, 1993; Terry & Kroger, 1976). For example, an early study by Terry and Krantz (1993) found that wearing glasses reduced social forcefulness, a characteristic including dominance, aggressiveness, and courage. Leder et al. (2011) showed that wearing glasses, especially full-rimmed ones, reduces attractiveness. Yet, in the current paper, we aim to show that wearing glasses actually improves electoral success. Although people have some negative associations, they also have many positive associations with glasses – in particular with learning, studying, and wisdom (Harris, 1991; Hellström & Tekle, 1994; Merry, 2012). This positive stereotype dates back to the Middle Ages, when monks used glasses to study despite declining vision. Glasses have since been commonly worn by people who perform intellectual or other highly skilled work (Ilardi, 2007). As a result, people associate glasses with a variety of competence-related characteristics, such as success, dependability, and industriousness, and most strongly intelligence (Harris, 1991; Hellström & Tekle, 1994; Jäckle & Metz, 2015; Leder et al., 2011; Manz & Lueck, 1968; Terry & Krantz, 1993; Thornton, 1943, 1944). Based on this notion, we propose that wearing glasses improves electoral success due to the robust association between perceptions of competence and electoral success (Ballew & Todorov, 2007; Chiao, Bowman, & Gill, 2008; Todorov et al., 2005). Competence often seems to be the best facial predictor of participants’ likelihood to vote for Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

a candidate (Jäckle & Metz, 2015; Rosenberg & McCafferty, 1987; Sussman, Petkova, & Todorov, 2013; Todorov et al., 2005; but see Praino, Stockemer, & Ratis, 2014; Verhulst, Lodge, & Lavine, 2010, for a discussion of when and why attractiveness might be more important). For example, competence seems to be the best predictor for electoral success in the previously mentioned US Senate and House races, in the German parliament, and for Bulgarian presidential candidates. In contrast, trust, likeability, attractiveness, and dominance do not predict electoral success as strongly as competence (Jäckle & Metz, 2015; Sussman et al., 2013; Todorov et al., 2005). In summary, although earlier research suggests that glasses may have negative effects, we propose that because glasses clearly boost perceived competence, wearing glasses increases politicians’ electoral success. We test this in eight studies using an experimental approach, in which we use photoshopped images of actual politicians (see Figure 1) to show participants the exact same politician – but with or without glasses. This experimental approach is particularly important because correlational findings from local elections in Denmark show that wearing glasses is related to reduced electoral success (Laustsen, 2014). However, the correlational nature of that study precludes causal inferences. Furthermore, Laustsen found this effect after controlling for differences in perceived competence between candidates, with glasses being related to increased competence. This is problematic because we predict that glasses increase electoral success exactly for this reason – by increasing perceived competence. Hypothesis 1 (H1): Glasses increase electoral success.

Possible Moderators and Mediators of the Glasses Effect Next, we predict that the positive effect of wearing glasses is moderated by a range of personal and situational factors. Political Orientation First, participants’ or politicians’ political orientation may play a role. In general, Republicans prefer conservativelooking politicians (Olivola, Sussman, Tsetsos, Kang, & Todorov, 2012; Olivola, Tingley, & Todorov, 2018), meaning politicians who look dominant (Laustsen & Petersen, 2015, 2016). As glasses tend to reduce dominance (Terry & Krantz, 1993), liberal participants may show a stronger and conservative participants a weaker glasses effect. Furthermore, attractiveness is generally important for electoral success (Berggren, Jordahl, & Poutvaara, 2010; Lutz, 2010; Praino et al., 2014; Rosar, Klein, & Beckers, 2008), but conservative politicians profit especially from it (Berggren, Ó 2018 Hogrefe Publishing

Figure 1. Example of actual stimuli pairs used in studies with American participants. Included with the kind agreement of Mrs. Agneta Gille (MP of Swedish Parliament for the Social Democrats). These are the only stimuli for which we got the permission to include them in print, but we are happy to share stimuli for research purposes.

Jordahl, & Poutvaara, 2017). As glasses also reduce attractiveness (Leder et al., 2011; Lundberg & Sheehan, 1994), the positive effect of glasses may be stronger for liberal politicians and weaker for conservative politicians. Then, it is possible that the glasses effect would be stronger for politicians of the same political orientation as the participants than for politicians from another party. Therefore, we expect the glasses effect to be stronger for liberal participants and for liberal politicians. Hypothesis 2a (H2a): The positive effect of glasses is stronger for liberal participants and liberal politicians than for conservative participants and politicians. Finally, because the effect of party affiliation already predicts electoral decisions strongly in cross-party elections (Campbell, Converse, Miller, & Stokes, 1960), it is possible that the party affiliation of a politician overrides the glasses effect for partisans. Hypothesis 2b (H2b): Do glasses increase electoral success in cross-party elections?

Political Situation Next, the effect of glasses may depend on the traits required by a specific political situation. Based on evolutionary psychology, human beings look for effective leadership in order to be able to survive specific threats to their group (Spisak, Nicholson, & van Vugt, 2011; van Vugt, Johnson, Kaiser, & O’Gorman, 2008; van Vugt, 2006, 2014). For example, people prefer male presidential candidates, associated with agentic skills, when there is a threat of international conflict, have no preference in the absence of conflict (Falk & Kenski, 2006), and even prefer female candidates when the political situation demands a communal candidate to overcome internal divisions (Lammers, Gordijn, & Social Psychology (2019), 50(1), 38–52

Otten, 2009). Similarly, people prefer masculine, strong, and dominant leaders in times of conflict, and feminine, intelligent, and cooperative leaders in times of peace (Laustsen & Petersen, 2015, 2017; Little et al., 2007; Spisak, Blaker, Lefevre, Moore, & Krebbers, 2014; Spisak, Dekker, Krüger, & van Vugt, 2012; Spisak, Homan, Grabo, & van Vugt, 2012; Spisak et al., 2011). Based on these findings, we propose that the positive effect of glasses differs with the characteristics needed in the current political situation. Although glasses are associated with intelligence, they are also associated with reduced perceptions of dominance (Elman, 1977; Leder et al., 2011; Terry & Krantz, 1993). Therefore, we expect that glasses have a positive effect in political situations that demand intelligence, but do not have this effect if the political situation makes dominance a more valuable trait. We test this hypothesis in Study 3. Hypothesis 3 (H3): The positive effect of glasses on electoral success is stronger when the political situation calls for intelligent leaders than when it calls for dominant leaders.

Strategic Use of Glasses Next, we test whether the positive effect of glasses disappears when voters are aware that politicians deliberately use glasses to win votes. Trustworthiness and honesty are important predictors of electoral success (Hetherington, 1998; Little, Roberts, Jones, & Debruine, 2012; McCurley & Mondak, 1995; Shephard & Johns, 2008; Wojciszke & Klusek, 1996). Using glasses merely to increase electoral success may be seen as deceptive and thus activate the common stereotype that politicians are dishonest (Blair, 2002; Koch, Imhoff, Dotsch, Unkelbach, & Alves, 2016). In addition, people tend to control for stereotypes if they are aware of their influence (Devine, 1989; Sczesny & Kühnen, 2004; Wegner & Bargh, 1998). If voters realize that a politician only wears glasses to look more intelligent, then they may avoid relying on that positive stereotype (Kunda & Sinclair, 1999). In sum, voters’ awareness of the strategic use of glasses may block the positive effect on electoral success. We test this idea in Study 4. Hypothesis 4 (H4): Glasses fail to increase electoral success if voters are aware of their strategic use.

Cross-Cultural Differences Next, we examine the potential moderating effect of crosscultural differences. Until now, the role of facial appearance in electoral success has been mostly studied with Western populations (Ballew & Todorov, 2007; Little et al., 2007; Poutvaara, Jordahl, & Berggren, 2009; Sussman et al., Social Psychology (2019), 50(1), 38–52

A. Fleischmann et al., Glasses Increase Electoral Success

2013; Todorov et al., 2005). No research has tested whether the effect of glasses (e.g., on voting) generalizes across cultures. One reason why the effect of glasses may differ more across cultures than the general effect of facial appearance is that the cultural stereotypes of glasses may differ much more between cultures. Although in rich Western countries a majority of adults wears glasses, this is much less common in less wealthy countries where people only wear glasses for very serious vision problems. For example, while over 60% of US Americans wear glasses, only 7% of Indians do so (Center for Disease Control and Prevention, 2013; Karnani, Garrette, Kassalow, & Lee, 2010). Consequently, in India and other less wealthy countries, wearing glasses may be less strongly associated with intelligence but more with vision deficiency and weakness (Elman, 1977; Swinford, 2013; Terry & Krantz, 1993). Therefore, if an Indian politician wears glasses, this may more easily be perceived as a sign of weakness, rather than of competence, and thus not produce the same positive effects as in the West. We test this in Study 5. Hypothesis 5 (H5): The positive effect of glasses on electoral success is limited to Western countries like the US and does not show in less wealthy countries like India.

Mediation by Intelligence and Warmth Finally, we aim to demonstrate the process underlying the relation between glasses and electoral success. We predict that the positive effect of glasses is driven by glasses making candidates look more intelligent (Leder et al., 2011; Terry & Krantz, 1993), and perceived intelligence increasing electoral success (e.g., Little et al., 2012; Shephard & Johns, 2008; Todorov et al., 2005). We test this prediction in Study 6 and Study 7, while also focusing on the hypothesized difference between Americans and Indians in the cultural association of glasses with intelligence in Study 6 (Hypothesis 5). Hypothesis 6 (H6): Americans associate glasses with intelligence, whereas Indians do not. Therefore, wearing glasses increases electoral success because it increases perceived intelligence in the US, but in India, it does not. Another possible mediator could be warmth. Some research finds that glasses increase facets of warmth such as honesty and helpfulness (Hellström & Tekle, 1994; Thornton, 1943, 1944), while other research finds that it reduces likeability (Jäckle & Metz, 2015; Leder et al., 2011). In turn, warmth is – next to intelligence – an important predictor for electoral success (Poutvaara et al., 2009; Rosenberg, Kahn, Tran, & Le, 1991; Shephard & Johns, 2008). If glasses lead to Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

increased warmth, warmth could also explain why glasses lead to higher electoral success. We test this prediction in Study 7. Hypothesis 7 (H7): Glasses lead to increased warmth, so warmth explains electoral success for politicians with glasses additionally to intelligence.

we report how we set sample size. In some of these studies, we accidentally collected more participants than planned, as occurs often in online research. All participants who finished a study were included in the analyses. We did not exclude any data, and we report all measures and manipulations. We report all studies conducted as part of this research project.1 Data and analyses can be found at https://osf.io/8867z.

Overview of Studies In summary, eight studies test how wearing glasses affects electoral success. Study 1 tests our main prediction that politicians who wear glasses have a higher electoral success. Studies 2a and 2b look at the influence of political orientation of participants and politicians. Study 3 examines the impact of political demands. Study 4 examines the effect of voters’ awareness of politicians’ strategic uses of glasses. Study 5 explores cross-cultural moderation by testing whether the effect generalizes to India. Study 6 explains the underlying process by testing whether stereotypical associations between glasses and intelligence mediate the effect, and whether this is an explanation for the cross-cultural differences tested in Study 5. Finally, Study 7 tests warmth as another potential mechanism for the glasses effect. Methodological Notes Across these studies, participants were recruited on Amazon’s Mechanical Turk (MTurk). Participants on MTurk are attentive to instructions and more representative of the US population than many convenience samples. Research using MTurk leads to similar results as nationally representative samples (Berinsky, Huber, & Lenz, 2012; Hauser & Schwarz, 2015; Mullinix, Leeper, Druckman, & Freese, 2016). To ensure high data quality, only MTurkers with an approval rate of over 90% participated, and only in one of these studies. Participants were compensated according to the length of each study. All studies were conducted using SoSci Survey (Leiner, 2015). First, we developed one stimulus set for our studies with American samples and one for our studies with Indian samples. Each stimulus set consisted of eight original pictures of politicians and eight pictures photoshopped with glasses (see Electronic Supplementary Material, ESM 1, for more information on stimulus development). Few participants overall were suspicious of the glasses, and suspicion did not change results (see ESM 1). Throughout these studies, 1

Study 1: Glasses Increase Electoral Success Study 1 tests whether wearing glasses increases politicians’ electoral success (Hypothesis 1).

Method Participants Two hundred three American MTurkers (75 women, 128 men, Mage = 36 years) participated. Sample size was set a priori to 200, which gave us a power of .80 (Cohen, 1992) to detect a small effect of d = 0.20 (Faul, Erdfelder, Lang, & Buchner, 2007). Procedure and Materials Participants were shown 16 pairs of same-sex politicians in a random order in a mock election paradigm (e.g., Lammers et al., 2009; Todorov et al., 2005), and indicated who they would vote for. Half of the pairs were critical trials, consisting of one politician with and one politician without glasses (see ESM 1 for stimulus development). To avoid that participants would guess the aim of the study, the remaining half of the pairs were filler trials, consisting only of original, unaltered pictures of politicians. These pictures were also politicians from the pretest, but those that were not chosen for our target stimuli. Participants always only saw each politician once, either with or without glasses. Politicians were identified simply with the letter A or B. Measures To measure electoral success, participants indicated which politician they would vote for, on a 7-point scale from 1 = definitely politician A to 7 = definitely politician B. We also measured demographics, including whether participants

Two additional studies are omitted from the paper given their conceptual and theoretical overlap with the reported studies. Study 1b (N = 200) replicated the positive effect of glasses of Study 1, t(199) = 5.16, p < .001, but was otherwise uninteresting because it failed to show a hypothesized moderation by dispositional differences. Study 4b (N = 202) only administered the wear glasses condition of Study 4. Consistent with the Study 4, participants indicated that wearing glasses would improve election success, t(201) = 8.03, p < .001. Given that Study 4 provides a more complete picture, we omit Study 4b. Both studies are in ESM 1.

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 38–52

wore glasses themselves, exploratory variables (see ESM 1), and participants’ suspicion of glasses (see ESM 1).

A. Fleischmann et al., Glasses Increase Electoral Success

party) within-subjects design. Based on a small effect size of d = 0.20 and 90% power, indicating a sample size of 265, and to account for some analyses only with partisans, we set sample size to 350.

Results Data and analyses for all studies can be found at https:// osf.io/8867z. We recoded all ratings so that higher values indicated a preference for politicians with glasses and lower values a preference for politicians without. In line with Hypothesis 1, a one-sample t-test indicated that participants preferred politicians with glasses over those without, t(202) = 4.16, p < .001, d = 0.29, M = 4.20, SD = 0.69, CI95 [0.15, 0.43]. For additional analyses with participants’ or politicians’ characteristics, see ESM 1.

Discussion Study 1 provided first evidence for Hypothesis 1 that wearing glasses improves electoral success. In a mock election paradigm, participants were more likely to vote for politicians when these politicians wore glasses than when they did not. Because we used exactly the same pictures of each politician, only changing whether they wore glasses or not (and counterbalanced between participants), we can rule out that this effect was due to any confounds with the stimuli.

Studies 2a and 2b: Political Orientation Study 2a tests whether the effect of wearing glasses depends on the political orientation of participants or of the politicians wearing glasses. The preregistration for this study can be found at https://aspredicted.org/5e44q.pdf. Study 2b tests whether glasses have a positive effect for partisans in cross-party elections. The preregistration can be found at https://aspredicted.org/xg4cz.pdf.

Method Participants and Design In Study 2a, 200 American MTurkers (80 women, 118 men, 2 other, Mage = 34) participated. The design was a 2 (political orientation of politician: Democrat vs. Republican) within-subjects design. As in Study 1, sample size was set to 200 to obtain 80% power for a small effect of d = 0.20. In Study 2b, 351 American MTurkers (159 women, 192 men, Mage = 38) participated. The design was a 2 (political orientation of politician with glasses: own vs. other Social Psychology (2019), 50(1), 38–52

Procedure and Materials We used the same procedure as in Study 1. In Study 2a, participants were told that the first eight pairs of politicians were Democrats and the second eight pairs were Republicans, or vice versa. Additionally, we asked participants two questions about glasses and fashion (see ESM 1). In Study 2b, politicians in the filler trials were always from the same party, and politicians in the critical trials were always a Republican running against a Democrat. Additionally, politicians had random names and appropriate ages (see ESM 1).

Results Study 2a As expected, participants were more likely to vote for the politician with glasses than for the politician without, t(199) = 4.15, p < .001, d = 0.30, M = 4.20, SD = 0.67, CI95 [0.15, 0.43]. In line with Hypothesis 2a, the more the participants identified as liberal, the more they showed the glasses effect, r(198) = .208, p = .003, CI95 [0.071, 0.337]. However, contrary to Hypothesis 2a, this was the case regardless of whether the politician was described as Democrat or Republican, t(199) = 0.09, p = .929, d = 0.01, Mdiff = 0.01, CI95 [ 0.13, 0.15]. Then, we tested whether participants showed a stronger glasses effect for their own party. In this analysis, we only included participants who were identified as either Democrats or Republican. Again, liberals showed a stronger glasses effect than conservatives, F(1, 151) = 7.52, p = .007, η2 = .05, CI90 [.01, .11], while the effect did not differ by politicians’ political orientation, F(1, 151) = 0.07, p = .790, η2 < .01, CI90 [0.00, 0.02], or the interaction between the two, F(1, 151) = 0.29, p = .594, η2 < .01, CI90 [0.00, 0.03]. For effects of fashion, see ESM 1. Study 2b With regard to Hypothesis 2b, participants were more likely to vote for the politician with glasses even in cross-party elections, t(350) = 3.58, p < .001, d = 0.19, M = 4.11, SD = 0.60, CI95 [0.85, 0.30]. Only including partisans, participants strongly voted for politicians from their own party, t(254) = 12.65, p < .001, d = 0.79, M = 4.98, SD = 1.24, CI95 [0.65, 0.93]. However, they were less likely to do so when the politician from the other party was wearing glasses (M = 4.88, SD = 1.49) than when the politician from Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

their own party was wearing glasses (M = 5.09, SD = 1.24), t (254) = 2.83, p = .005, d = 0.18, M = 0.21, SD = 1.19, CI95 [0.05, 0.31].

Discussion Studies 2a and 2b replicate the basic finding of glasses leading to better electoral success. As expected in Hypothesis 2a, this effect is stronger the more liberal participants are. Interestingly, the same is not true for the politicians’ political orientation: No matter whether a politician was presented as Republican or Democrat, participants preferred politicians with glasses. Additionally, glasses increased electoral success for politicians with the same or a different political orientation, and they also increased electoral success in cross-party elections. Glasses therefore seem to have a positive effect across the political spectrum of politicians.

stated what they searched for in a politician and filled out exploratory variables (see ESM 1).

Results Eighty-seven percent of participants correctly identified the problem in the country before the vote. Participants in the peace condition also searched more for a politician who could deliberate well, rather than act fast, compared to participants in the war condition (see ESM 1 for analyses). In support of Hypothesis 3, we found that participants’ electoral decisions differed by experimental condition, t(198) = 2.10, p = .037, d = 0.30, CI95 [0.02, 0.58]. In the peace condition (n = 96), participants preferred the politicians with glasses over politicians without glasses, t(95) = 2.49, p = .015, d = 0.25, M = 4.20, SD = 0.78, CI95 [0.05, 0.46]. In the war condition (n = 104), participants had an equal preference for leaders with and without glasses, t(103) = 0.43, p = .670, d = 0.04, M = 3.97, SD = 0.78, CI95 [ 0.23, 0.15].

Study 3: Political Demands Study 3 tests whether the effect of wearing glasses depends on situational political demands, specifically on whether intelligence or dominance is desirable.

Method Participants and Design Two hundred American MTurkers (85 women, 115 men, Mage = 32) participated. Participants were randomly assigned to one of two between-participant conditions (peace vs. war). Expecting the same effect as in Study 1, we calculated sample size a priori with G*Power (Faul et al., 2007) to obtain a power of .80. We rounded the recommended sample size (192) up to a target N = 200.

Discussion In the peace condition, Study 3 replicated the finding from Studies 1 and 2 that glasses increase electoral success. As predicted by Hypothesis 3, Study 3 also showed that this positive effect only occurs in the political situation that asks for intelligence in a leader, but not if the country is threatened by an armed conflict (stereotypically requiring dominance instead). Nonetheless, it is important to note that even if participants believed the country to be threatened by war, wearing glasses did not produce a negative effect.

Study 4: Strategic Use Procedure and Material We used the same procedure as in Study 1, except for the addition of the political demands manipulation. Participants first learned about the most important problem facing the country, either complex legislative problems (peace condition), or an attack from a neighboring country (war condition). In the former condition, the country needed a president who could deliberate on the problem, in the latter, a president who could act fast. Measures In addition to the measures from Study 1, participants indicated what the most important problem of the country was as an attention check (see ESM 1 for details). They also Ó 2018 Hogrefe Publishing

Study 4 investigated whether the effect of glasses is undermined if participants are aware that glasses are used deliberately to improve electoral results (Hypothesis 4).

Method Participants and Design Two hundred one American MTurkers (77 women, 124 men, Mage = 34) participated. Participants were randomly assigned to one of two conditions (wear glasses vs. remove glasses). Based on the calculations in Study 3, sample size was set to 200 a priori. Social Psychology (2019), 50(1), 38–52

Material and Procedure In the wear glasses condition, participants were shown one random unaltered picture of a politician (from the American stimulus set). They were told that the politician was running for office but was not doing so well. Therefore, one advisor proposed that the politician should wear glasses. Then, participants were also shown the altered picture of the same politician. In the remove glasses condition, participants instead saw one of the photoshopped pictures of the politician with glasses. They were told the same story, but the advisor instead proposed to remove the glasses. Then, they also saw the original picture of the politician without glasses.

Measures To measure overall success of the candidate, participants indicated how likely they and how likely others would vote for the candidate, both between 1 = less likely and 7 = more likely, and how wearing/removing glasses would change the politician’s election results, between 1 = definitely hurt and 7 = definitely help. These three questions were averaged (wear glasses condition: Cronbach’s α = .90, remove glasses condition: Cronbach’s α = .79). For exploratory purposes, participants also indicated on two separate items whether glasses would help or hurt if the politician was running for the Republican Party or if the politician was running for the Democratic Party, both between 1 = definitely hurt and 7 = definitely help. We did not ask for suspicion in this study because we openly showed both the pictures with and without glasses.

Results We recoded votes so that in both conditions, higher numbers indicated that participants preferred the politician to wear glasses (either keep wearing their glasses or start wearing them). Reactions to the politician differed by advice given, t(199) = 2.81, p = .005, d = 0.40, CI95 [0.12, 0.68]. Contrary to Hypothesis 4, for the wear glasses condition (n = 101), participants preferred the politicians to wear glasses, t(100) = 3.74, p < .001, d = 0.37, M = 4.34, SD = 0.91, CI95 [0.17, 0.57]. For the remove glasses condition (n = 100), participants did not have a preference for the politicians to either keep or remove the glasses, t(99) = 0.32, p = .753, d = 0.03, M = 3.97, SD = 0.95, CI95 [ 0.16, 0.23]. This finding suggests that participants believe that adopting glasses can help a politician, while removing them does not, because they thought both they themselves and others would be more likely to vote for a politician with glasses. For additional analyses with participants’ demographics and politicians’ political orientation, see ESM 1. Social Psychology (2019), 50(1), 38–52

A. Fleischmann et al., Glasses Increase Electoral Success

Discussion Study 4 failed to find support for Hypothesis 4. We again found a positive impact of glasses on electoral success, even when we directly told participants of the strategic use of glasses for the purpose of winning an election. Even though participants were aware of the impression management, they indicated that they and others would be more likely to vote for the politician with glasses. In contrast to that, removing glasses for strategic reasons did not lead to a positive effect, ruling out that participants were merely going along with any strategy an advisor proposed. All in all, these results show that voters’ awareness that a politician wears glasses for strategic impression management goals does not seem to lead to any negative effects.

Study 5: Glasses in India Our first four studies focused on American samples. To test for cultural differences, Study 5 tests whether the effect generalizes to a non-Western setting. As stated in Hypothesis 5, we expected the effect to be limited to the US and not shown in India.

Method Participants Two hundred three Indian MTurkers (74 women, 129 men, Mage = 33 years) participated. Sample size was set a priori at 200 to obtain a power of .80 (Cohen, 1992) to detect a small effect of d = 0.20 (Faul et al., 2007). Procedure and Material The procedure closely followed that of Study 1. Participants were shown eight pairs of same-sex Indian politicians of similar baseline popularity, in random order. The four critical trials featured a politician with and one without glasses. Then, participants stated for whom they would vote, on a 7point scale from 1 = definitely politician A to 7 = definitely politician B. After that, participants answered demographic questions, whether they wore glasses themselves and whether they found anything suspicious.

Results Participants’ responses again were recoded so that higher values indicated a preference for politicians with glasses. In line with Hypothesis 5 and consistent with the idea that wearing glasses may even be seen as a sign of weakness, Indians marginally preferred politicians without glasses, Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

t(202) = 1.94, p = .054, d = 0.14, M = 3.84, SD = 1.20, CI95 [ 0.27, 0.002]. See ESM 1 for analyses with participants’ demographics.

Discussion Consistent with our reasoning for Hypothesis 5, the positive effect of glasses did not replicate in India. In fact, if anything, glasses led to marginally lower electoral success. As we described, glasses are rare in India (only 7% wear them) and are typically worn out of dire medical needs. Therefore, Indians may not associate glasses with intelligence as much as Americans do. Another possibility may be that Indians value different characteristics in politicians. For example, Rule et al. (2010) showed that dominance predicts electoral success in the US and warmth does not, but the opposite is true for Japan. Therefore, we directly compared ratings of intelligence for politicians with glasses and their predictive power for electoral success in India and the US in Study 6.

Study 6: Intelligence as Mediator Study 6 replicated the finding of Study 5 and directly compared an Indian sample and an American sample in one study. It also tested whether the unequal glasses effect in India and America is due to different associations of glasses with intelligence (Hypothesis 6). As an alternative explanation, it examined whether people in the US and India value intelligence differently.

Method Participants Four hundred seven participants were recruited on MTurk. We targeted our sample so that we obtained an equal number of Indians (NIndia = 206; 76 women, 130 men; Mage = 32 years) and Americans (NAmerica = 201; 80 women, 120 men, 1 other; Mage = 33 years). We set sample size a priori at 200 per sample so that again we would be able to detect a small effect of d = .20 in each sample. Procedure and Material Participants were shown a pair of same-sex politicians of similar baseline popularity from the respective stimulus set. Participants indicated their associations of intelligence for the two candidates, judging who was more intelligent, more rational, and more intellectual (Terry & Krantz, 2

1993), between 1 = definitely politician A and 7 = definitely politician B, Cronbach’s α = .77, and rated exploratory variables (see ESM 1). As in previous studies, all ratings were recoded so that higher ratings indicated that the politician with glasses was considered to be more intelligent. Next, we used the same measure of electoral success as in previous studies. After that, participants provided demographics. Given that measures of suspicion did not moderate any of the previous results, we did not measure suspicion.

Results We first discuss the influence of glasses on electoral success and intelligence in both India and the US. Then, we test whether the difference in electoral success is mediated by differences in intelligence ratings. Election Results Supporting Hypothesis 5, Indians and Americans differed in their preference for politicians with glasses, t(358) = 2.43,2 p = .016, d = 0.24, CI95 [0.05, 0.44]. Replicating the results of Studies 1–4, Americans (n = 201) preferred politicians with glasses over those without glasses, t(200) = 3.69, p < .001, d = 0.26, M = 4.41, SD = 1.57, CI95 [0.12, 0.40]. Consistent with Hypothesis 5, Indians (n = 206) again did not share this preference, t(205) = 0.44, p = .657, d = 0.03, M = 3.93, SD = 2.35, CI95 [ 0.17, 0.11], although we did not find the marginally significant reversal we found in Study 5. Intelligence In support of Hypothesis 6, Indians and Americans differed in their associations between glasses and intelligence, t (388) = 4.12 (see Footnote 2), p < .001, d = .41, CI95 [0.21, 0.60]. While Americans rated politicians with glasses as more intelligent than politicians without, t(200) = 6.95, p < .001, d = 0.49, M = 4.59, SD = 1.20, CI95 [0.34, 0.64], Indians did not, t(205) = 0.28, p = .784, d = 0.02, M = 4.03, SD = 1.52, CI95 [ 0.12, 0.16]. Mediation To test whether intelligence ratings explained the difference in electoral success between countries (Hypothesis 6), we tested for mediation with the Process macro (Hayes, 2013, 10,000 bootstrapping resamples). The full mediation model can be found in Figure 2. Mirroring the previous analyses, participants’ country predicted electoral success for politicians with glasses compared to politicians without, b = 0.481, SE = 0.199, p = .016, CI95 [ 0.871, 0.090], showing that Americans were more likely to vote for politicians with glasses than Indians. Similarly, participants’

Corrected for variance heterogeneity.

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 38–52

A. Fleischmann et al., Glasses Increase Electoral Success

Participants Two hundred and one American MTurkers participated (84 women, 116 men, Mage = 36). Sample size was set to 200 to obtain 80% power for a small effect of d = .20.

Figure 2. Mediation model for Study 6. Country of participants predicts intelligence ratings of politicians with glasses, which predicts electing politicians with glasses. *p < .05; ***p < .001.

country predicted intelligence ratings of politicians with glasses compared to politicians without, b = 0.558, SE = 0.136, p < .001, CI95 [ 0.825, 0.291], as Americans found politicians with glasses compared to those without more intelligent than did Indians. When both country and intelligence ratings were used to predict electoral success, intelligence ratings predicted electoral success, b = 0.957, SE = 0.055, p < .001, CI95 [0.849, 1.065], but country did not, b = 0.053, SE = 0.154, p = .730, CI95 [ 0.249, 0.355]. Therefore, the data is in line with Hypothesis 6: The different electoral preferences of Indians and US Americans for politicians with glasses can be explained through their different intelligence ratings for politicians with glasses, indirect effect ab = 0.534, SE = 0.134, CI95 [ 0.804, 0.272]. For additional analyses on how Indians and Americans value intelligence in a candidate, see ESM 1.

Discussion The current results replicate and integrate the findings of Studies 1–5, by showing that although wearing glasses produces a positive effect on political candidates’ electoral success in the US, this is not the case for political candidates in India. Furthermore, these results show that this is due to a difference in the stereotype of glasses. Where Americans strongly associate glasses with intelligence, Indians do not.

Study 7 – Intelligence and Warmth as Mediators Study 7 was used to replicate the effect that glasses make politicians look more intelligent in the US and therefore lead to higher electoral success. More importantly, it tested whether glasses make politicians look warmer in the US as another explaining mechanism.

Procedure and Materials Procedure and materials were very similar to Study 6 for the American MTurkers. The only difference was that instead of dominance, we measured warmth. To measure warmth, participants answered which of the two politicians was more compassionate, friendly, and honest (Terry & Krantz, 1993), on a 7-point Likert scale between 1 = definitely politician A and 7 = definitely politician B, Cronbach’s α = .80. Again, we did not measure suspicion due to absent effects of suspicion in the previous studies.

Results Replicating previous studies, participants marginally preferred politicians with glasses, t(200) = 1.93, p = .055, d = 0.14, M = 4.22, SD = 1.61, CI95 [ 0.003, 0.27], and perceived them as more intelligent, t(200) = 6.25, p < .001, d = 0.44, M = 4.52, SD = 1.18, CI95 [0.30, 0.59]. However, they did not perceive politicians with glasses to be warmer than politicians without, t(200) = 0.06, p = .956, d < 0.01, M = 4.00, SD = 1.28, CI95 [ 0.14, 0.13]. Even though warmth did not differ for politicians with and without glasses, both warmth of a politician, b = 0.558, SE = 0.064, p < .001, CI95 [0.432, 0.684], and intelligence of a politician, b = 0.621, SE = 0.070, p < .001, CI95 [0.483, 0.759], predicted electoral success.

Discussion Study 7 replicates the effect of glasses on perceived intelligence. Contrary to Hypothesis 7, we find no effect of glasses on perceived warmth. This is in line with the mixed effects found by previous research, finding glasses to be related to both increased and decreased warmth (Hellström & Tekle, 1994; Jäckle & Metz, 2015; Leder et al., 2011; Thornton, 1943, 1944). While both intelligence and warmth predict electoral success quite well, the missing influence of glasses on perceived warmth indicates that intelligence is responsible for the electoral success of politicians with glasses, while warmth is not.

Meta-Analysis Method The preregistration for this study can be found at https:// aspredicted.org/t6d5v.pdf. Social Psychology (2019), 50(1), 38–52

To precisely estimate the glasses effect, we conducted a meta-analysis for our US and Indian samples (see ESM 1 for inclusion of studies). Using R (R Core Team, 2017), Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

While politicians often believe that glasses can hurt their electoral success, we show in eight experiments that glasses actually increase electoral success. In Study 1 (basic paradigm), participants in mock elections voted for politicians without glasses or with photoshopped glasses. We found that participants clearly preferred candidates with glasses. The positive effect of glasses was particularly strong for liberal participants (Study 2a) – but also occurred in crossparty elections (Study 2b) – and when intelligent politicians were needed (Study 3). Furthermore, against our prediction, the positive effect of glasses held even for the strategic use of glasses (Study 4). Cross-culturally, the positive effect of glasses observed in the US did not extend to India, because people in India do not share Americans’ stereotype to associate glasses with intelligence (Studies 5 and 6). Finally, while glasses seem to increase perceptions of intelligence, which lead to higher electoral success, perceptions of warmth remain unchanged by glasses (Study 7). The fact that people preferred candidates with glasses even when these glasses were worn strategically (Study 4) is surprising, given that people usually control for their stereotypes when they are aware of them. A strategic use of glasses should rather activate the stereotype of a dishonest politician (Devine, 1989; Kunda & Sinclair, 1999). Possibly, voters know that in the US, presidential candidates spend millions of dollars and employ hundreds of staff members to find the best way to improve their self-presentation (Gilens, 2012; Morton & Cameron, 1992; Stratmann, 2005). In that light, a small amount spent on glasses seems trivial and may not lead to such strong reactions. Future research may seek to test this further.

and did not have any reason to assume that their decision would influence their lives. In real elections, voters have more access to substantive information (political programs) and may be more motivated to process these. By only offering peripheral cues, our design may have favored peripheral over central processing (Chaiken, 1980; Petty & Cacioppo, 1986; Thompson, Roman, Moskowitz, Chaiken, & Bargh, 1994) and increased the effect. At the same time, considerable political scientific literature argues that many real-life political decisions, including elections, are characterized by strong peripheral processing (Lupia, 1994; Popkin, 1991). Furthermore, peripheral cues may have an even stronger impact when people have other information to attribute their decision to (Yzerbyt, Schadron, Leyens, & Rocher, 1994). Finally, while cues like party affiliation or policy beliefs likely play a stronger role than glasses, glasses are – in contrast to these other factors – easily changeable. Another limitation concerns our participants and stimuli. While MTurkers are more representative of the US population than convenience samples, they are still somewhat more liberal than a representative sample (Berinsky et al., 2012). Previous research has shown that conservatives value facial dominance more than liberals (Laustsen & Petersen, 2015, 2016). We also found that liberal participants showed a stronger effect of glasses in Study 1 (see ESM 1) and tested this in Study 2a. Therefore, it is possible that very conservative voters would not show a glasses effect, as they value dominance more in their decision. Similar arguments can be made for the used glasses and stimuli. We chose contemporary glasses, which may be better in conveying intelligence than old-fashioned glasses, which perhaps only convey reduced dominance and decreased health. Additionally, some glasses might convey warmth while others do not, which would explain inconsistent findings in previous research on the relationship of glasses and warmth. In this case, warmth might additionally mediate the effect of glasses on electoral success, leading to a stronger effect. Pending future research, we assume that most glasses that are relatively neutral in color and shape are associated with intelligence, so that the effect of glasses should be positive. Similarly, we chose politicians that already looked likeable and trustworthy. For sinister-looking politicians or politicians with known negative characteristics, glasses might hurt, as competence is only seen as positive if it is coupled with morality (Landy, Piazza, & Goodwin, 2016). It is likely that people are more afraid of competent evil than incompetent evil.

Limitations and Strengths

Theoretical Implications

One limitation of our work is that participants were not offered any substantive information about the candidates

Our findings can only partly be explained by an evolutionary psychological view on leadership selection. From a

we computed effect sizes and standard errors with BootES (Kirby & Gerlanc, 2013) and then computed a randomeffects meta-analysis with metafor (Viechtbauer, 2010). Consistent with the studies reported above, for the US sample, we found a clear effect of glasses, d = 0.25, SE = 0.03, p < .001, CI95 [0.19, 0.31]. In contrast, for the Indian sample, there was no positive effect of glasses, d = 0.08, SE = 0.05, p = .115, CI95 [ 0.19, 0.02]. Most importantly, country moderated the glasses effect, b = 0.336, SE = 0.064, p < .001, CI95 [ 0.462, 0.211]. Therefore, glasses increased electoral success in the US, but not in India.

General Discussion

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 38–52

strong evolutionary standpoint, health and attractiveness should be preferred in leaders in all situations (Spisak et al., 2014), and they should be signaled by unchangeable biological factors such as a person’s height or their face’s width-to-height ratio (Poutvaara, 2014; Spisak et al., 2011; van Vugt et al., 2008). If glasses signal anything, it should be weakness and bad health, leading to less leadership potential regardless of culture. In contrast to that, our results indicate that glasses positively influence the perception of leaders in some countries, depending on the associated stereotypes. This is in line with a moderate perspective that assumes that people use facial information to detect desirable traits depending on the situation (Oosterhof & Todorov, 2008; Spisak et al., 2014), as leadership emerged as an adaptive strategy to solve important problems faced by the group (Spisak et al., 2011; van Vugt, 2006; van Vugt & Ronay, 2013). Accordingly, we find that wearing glasses increases electoral success when situational demands lead people to look for intelligent leaders, and only in countries where glasses are considered to be a reliable signal for intelligence. Another important theoretical implication concerns the cross-cultural differences we tested, as previous cross-cultural research on facial appearance shows mixed results. On the one hand, Americans’ and Indians’ ratings of politicians’ competence predicted Mexican and Brazilian election outcomes (Lawson, Lenz, Baker, & Myers, 2010). On the other hand, Americans’ and Japanese’s ratings of dominance only predicted election results in the US, while their ratings of warmth only predicted election results in Japan. Additionally, explicit ratings of electoral success were only predictive within country (Rule et al., 2010). Even though both links in our mediation model are well established, we show that our findings only apply to the US (and, tentatively, to Western cultures where glasses are common) and find that the effects of glasses on elections do not generalize to India (or, tentatively, to Eastern cultures, if glasses are less common). Furthermore, this difference seems to be due to factors that are quite susceptible to cultural influences (stereotypes of glasses, see, e.g., Cuddy et al., 2009, for cultural differences in stereotype content), while the basic mechanism of electing competent-looking politicians does not seem to differ.

Practical Implications We believe that our findings also have important practical implications. Politicians with competent faces seem to enjoy better electoral success than their less competentlooking colleagues. Although that finding has important theoretical implications, the applied, practical implications for politicians are limited. First, it is unclear what facial features actually contribute to this effect, as most studies Social Psychology (2019), 50(1), 38–52

A. Fleischmann et al., Glasses Increase Electoral Success

looked at how general impressions predict votes (Atkinson, Enos, & Hill, 2009; Ballew & Todorov, 2007; Chiao et al., 2008; Poutvaara et al., 2009; Shephard & Johns, 2008; Todorov et al., 2005). More importantly, in those few cases when specific features were examined, these features were mostly unchangeable, such as the shape of politicians’ faces or eyes (Rosenberg et al., 1991; Zebrowitz & Montepare, 2005). In contrast, we focused on an easily changeable feature – glasses – that politicians can change to their advantage by simply making a purchase! Our results suggest that wearing glasses does not harm politicians as many appear to think, but instead seems to offer politicians an advantage over their competitors (at least in Western cultures) without any real drawbacks. Even though we found that the positive effect of glasses was moderated by a number of factors, it did not reverse in any of these cases. For example, Study 2b showed that glasses had the same positive effect in cross-party elections. Although Study 3 showed that glasses do not have a positive effect when a country is threatened by conflict (suggesting the need for dominance over intelligence), glasses also did not have a negative effect even then. Additionally, Study 4 showed that voters’ awareness of self-presentation intentions behind wearing glasses does not undermine the positive effects of wearing glasses. Although the glasses manipulation produced small effects across studies, such small effects could nonetheless make a difference, in particular in a plurality or first-pastthe-post system, such as in the US, Canada, or Great Britain. In these elections, political candidates gain little beyond getting the smallest majority possible (50.1% of the vote), and thus, such elections are often decided by relatively narrow margins (Downs, 1957). A small effect could make a difference between winning and losing a battleground state and thus provide a candidate with an edge over their opponent. However, we note that election success depends not only on the decisions of the voters, but also on the decision to vote at all. As we have not tested the effect of glasses on voter turnout, we cannot say whether glasses would also help to mobilize voters to vote.

Conclusion Facial features are consistently found to be important predictors of electoral success (Ballew & Todorov, 2007; Todorov et al., 2005). Yet existing research has limited itself to mostly testing fixed, unchangeable features of faces. We focus on changeable features and show that wearing glasses reliably boosts electoral success. Glasses offer an easy, effective, and robust way for politicians to change their facial features and increase the probability of electoral success in the West (as long as competence is important). In fact, politicians may even use glasses Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

strategically, as voters are still more likely to vote for them when this strategic use is known. The positive effect of glasses may not generalize beyond Western settings such as India though. Overall, our research suggests that not only unchangeable, but also alterable facial features play an important role in impression formation and electoral success.

Open Data/Materials Data and analyses can be found at https://osf.io/8867z.

Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1864-9335/a000359 ESM 1. Analyses (.pdf) Stimulus development, Studies 1b and 4b, and additional analyses.

References Antonakis, J., & Dalgas, O. (2009). Predicting elections: Child’s play! Science, 323, 1183–1183. https://doi.org/10.1126/ science.1167748 Atkinson, M. D., Enos, R. D., & Hill, S. J. (2009). Candidate faces and election outcomes: Is the face-vote correlation caused by candidate selection? Quarterly Journal of Political Science, 4, 229–249. https://doi.org/10.1561/100.0000806 Ballew, C. C., & Todorov, A. (2007). Predicting political elections from rapid and unreflective face judgments. Proceedings of the National Academy of Sciences, 104, 17948–17953. https://doi. org/10.1073/pnas.0705435104 Berggren, N., Jordahl, H., & Poutvaara, P. (2010). The looks of a winner: beauty and electoral success. Journal of Public Economics, 94, 8–15. https://doi.org/10.1016/j.jpubeco.2009. 11.002 Berggren, N., Jordahl, H., & Poutvaara, P. (2017). The right look: Conservative politicians look better and voters reward it. Journal of Public Economics, 146, 79–86. https://doi.org/ 10.1016/j.jpubeco.2016.12.008 Berinsky, A. J., Huber, G. A., & Lenz, G. S. (2012). Evaluating online labor markets for experimental research: Amazon.com’s Mechanical Turk. Political Analysis, 20, 351–368. https://doi. org/10.1093/pan/mpr057 Blair, I. V. (2002). The malleability of automatic stereotypes and prejudice. Personality and Social Psychology Review, 6, 242– 261. https://doi.org/10.1207/s15327957pspr0603_8 Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York, NY: Wiley. Center for Disease Control and Prevention. (2013). 2013 national health interview survey functioning and disability file [Data file and SPSS code]. Retrieved from http://www.cdc.gov/nchs/ nhis/funcdisb2013.htm

Ó 2018 Hogrefe Publishing

Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39, 752–766. https://doi.org/10.1037/0022-3514.39.5.752 Chiao, J. Y., Bowman, N. E., & Gill, H. (2008). The political gender gap: Gender bias in facial inferences that predict voting behavior. PLoS One, 3, e3666. https://doi.org/10.1371/journal. pone.0003666 Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155– 159. https://doi.org/10.1037/0033-2909.112.1.155 Cuddy, A. J. C., Fiske, S. T., Kwan, V. S. Y., Glick, P., Demoulin, S., Leyens, J.-P., . . . Ziegler, R. (2009). Stereotype content model across cultures: Towards universal similarities and some differences. British Journal of Social Psychology, 48, 1–33. https://doi.org/10.1348/014466608X314935 Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. https://doi.org/10.1037/0022-3514. 56.1.5 Downs, A. (1957). An economic theory of democracy. New York, NY: Harper. Edwards, K. (1987). Effects of sex and glasses on attitudes toward intelligence and attractiveness. Psychological Reports, 60, 590– 590. https://doi.org/10.1177/003329418706000201 Elman, D. (1977). Physical characteristics and the perception of masculine traits. The Journal of Social Psychology, 103, 157– 158. https://doi.org/10.1080/00224545.1977.9713308 Falk, E., & Kenski, K. (2006). Issue saliency and gender stereotypes: Support for women as presidents in times of war and terrorism. Social Science Quarterly, 87, 1–18. https://doi.org/ 10.1111/j.0038-4941.2006.00365.x Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. https://doi.org/10.3758/BF03193146 Gilens, M. (2012). Affluence and influence: Economic inequality and political power in America. Princeton, NJ: Princeton University Press. Harris, M. B. (1991). Sex differences in stereotypes of spectacles. Journal of Applied Social Psychology, 21, 1659–1680. https:// doi.org/10.1111/j.1559-1816.1991.tb00497.x Hasart, J. K., & Hutchinson, K. L. (1993). The effects of eyeglasses on perceptions of interpersonal attraction. Journal of Social Behavior & Personality, 8, 521–528. Hauser, D. J., & Schwarz, N. (2015). Attentive Turkers: MTurk participants perform better on online attention checks than do subject pool participants. Behavior Research Methods, 48, 400–407. https://doi.org/10.3758/s13428-015-0578-z Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press. Hellström, Å., & Tekle, J. (1994). Person perception through facial photographs: Effects of glasses, hair, and beard on judgments of occupation and personal qualities. European Journal of Social Psychology, 24, 693–705. https://doi.org/10.1002/ejsp. 2420240606 Hetherington, M. J. (1998). The political relevance of political trust. The American Political Science Review, 92, 791–808. https://doi.org/10.2307/2586304 Ilardi, V. (2007). Renaissance vision from spectacles to telescopes. Philadelphia, PA: American Philosophical Society. Jäckle, S., & Metz, T. (2015). Attractiveness, competence or likability? Appearance effects in the 2013 German election. Paper presented at the ECPR General Conference, Montreal, Canada.

Social Psychology (2019), 50(1), 38–52

Jamieson, A. (2016). Jeb Bush’s glasses conundrum: Some unsolicited fashion advice. The Guardian. Retrieved from http://www.theguardian.com/us-news/2016/feb/20/jeb-bushno-glasses-fashion-advice Karnani, A. G., Garrette, B., Kassalow, J. S., & Lee, M. (2010). Better vision for the poor. Stanford Social Innovation Review, 9, 66–71. Kirby, K. N., & Gerlanc, D. (2013). BootES: An R package for bootstrap confidence intervals on effect sizes. Behavior Research Methods, 45, 905–927. https://doi.org/10.3758/ s13428-013-0330-5 Koch, A., Imhoff, R., Dotsch, R., Unkelbach, C., & Alves, H. (2016). The ABC of stereotypes about groups: Agency/socioeconomic success, conservative–progressive beliefs, and communion. Journal of Personality and Social Psychology, 110, 675–709. https://doi.org/10.1037/pspa0000046 Kunda, Z., & Sinclair, L. (1999). Motivated reasoning with stereotypes: Activation, application, and inhibition. Psychological Inquiry, 10, 12–22. https://doi.org/10.1207/ s15327965pli1001_2 Lammers, J., Gordijn, E. H., & Otten, S. (2009). Iron ladies, men of steel: The effects of gender stereotyping on the perception of male and female candidates are moderated by prototypicality. European Journal of Social Psychology, 39, 186–195. https:// doi.org/10.1002/ejsp.505 Landy, J., Piazza, J. R., & Goodwin, G. (2016). When it’s bad to be friendly and smart: The desirability of sociability and competence depends on morality. Personality and Social Psychology Bulletin, 42, 1272–1290. https://doi.org/10.1177/ 0146167216655984 Laustsen, L. (2014). Decomposing the relationship between candidates’ facial appearance and electoral success. Political Behavior, 36, 777–791. https://doi.org/10.1007/s11109-0139253-1 Laustsen, L., & Petersen, M. B. (2015). Does a competent leader make a good friend? Conflict, ideology and the psychologies of friendship and followership. Evolution and Human Behavior, 36, 286–293. https://doi.org/10.1016/j.evolhumbehav.2015.01.001 Laustsen, L., & Petersen, M. B. (2016). Winning faces vary by ideology: How nonverbal source cues influence election and communication success in politics. Political Communication, 33, 188–211. https://doi.org/10.1080/10584609.2015.1050565 Laustsen, L., & Petersen, M. B. (2017). Perceived conflict and leader dominance: Individual and contextual factors behind preferences for dominant leaders. Political Psychology, 38, 1083–1101. https://doi.org/10.1111/pops.12403 Lawson, C., Lenz, G. S., Baker, A., & Myers, M. (2010). Looking like a winner: Candidate appearance and electoral success in New Democracies. World Politics, 62, 561–593. https://doi.org/ 10.1017/S0043887110000195 Leder, H., Forster, M., & Gerger, G. (2011). The glasses stereotype revisited. Swiss Journal of Psychology, 70, 211–222. https://doi. org/10.1024/1421-0185/a000059 Leiner, D. J. (2015). SoSci Survey (Version 2.6.00-i) [Computer Software]. Retrieved from https://www.soscisurvey.com Little, A. C., Burriss, R. P., Jones, B. C., & Roberts, S. C. (2007). Facial appearance affects voting decisions. Evolution and Human Behavior, 28, 18–27. https://doi.org/10.1016/j. evolhumbehav.2006.09.002 Little, A. C., Roberts, S. C., Jones, B. C., & Debruine, L. M. (2012). The perception of attractiveness and trustworthiness in male faces affects hypothetical voting decisions differently in wartime and peacetime scenarios. The Quarterly Journal of Experimental Psychology, 65, 2018–2032. https://doi.org/ 10.1080/17470218.2012.677048

Social Psychology (2019), 50(1), 38–52

A. Fleischmann et al., Glasses Increase Electoral Success

Lundberg, J. K., & Sheehan, E. P. (1994). The effects of glasses and weight on perceptions of attractiveness and intelligence. Journal of Social Behavior and Personality, 9, 753–760. Lupia, A. (1994). Shortcuts versus encyclopedias: Information and voting behavior in California insurance reform elections. The American Political Science Review, 88, 63–76. https://doi.org/ 10.2307/2944882 Lutz, G. (2010). The electoral success of beauties and beasts. Swiss Political Science Review, 16, 457–480. https://doi.org/ 10.1002/j.1662-6370.2010.tb00437.x Manz, W., & Lueck, H. E. (1968). Influence of wearing glasses on personality ratings: Crosscultural validation of an old experiment. Perceptual and Motor Skills, 27, 704–704. https://doi. org/10.2466/pms.1968.27.3.704 McCurley, C., & Mondak, J. J. (1995). Inspected by #1184063113: The influence of incumbents’ competence and integrity in US House elections. American Journal of Political Science, 39, 864– 885. Merry, S. (2012). Eye see you: How criminal defendants have utilized the nerd defense to influence jurors’ perceptions. Journal of Law & Policy, 21, 725–771. https://doi.org/10.1111/ j.1468-0343.1992.tb00056.x Morton, R., & Cameron, C. (1992). Elections and the theory of campaign contributions: A survey and critical analysis. Economics & Politics, 4, 79–108. https://doi.org/10.1111/j.14680343.1992.tb00056.x Mullinix, K. J., Leeper, T. J., Druckman, J. N., & Freese, J. (2016). The generalizability of survey experiments. Journal of Experimental Political Science, 2, 109–138. https://doi.org/10.1017/ XPS.2015.19 Olivola, C. Y., Tingley, D., & Todorov, A. (2018). Republican voters prefer candidates who have conservative-looking faces: New evidence from exit polls. Political Psychology, 39, 1157–1171. https://doi.org/10.1111/pops.12489 Olivola, C. Y., Tingley, D., & Todorov, A. (2018). Republican voters prefer candidates who have conservative-looking faces: New evidence from exit polls. Political Psychology, 39, 1157–1171. https://doi.org/10.1111/pops.12489 Oosterhof, N. N., & Todorov, A. (2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences, 105, 11087–11092. https://doi.org/10.1073/pnas.0805664105 Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In R. E. Petty & J. T. Cacioppo (Eds.), Communication and persuasion: Central and peripheral routes to attitude change (pp. 1–24). New York, NY: Springer. Pindell, J. (2015). Four-eyes for president. The Boston Globe. Retrieved from https://www.bostonglobe.com/metro/2015/02/ 13/four-eyes-for-president/S8q2IX7pr1MzZLS981qV5J/story. html Popkin, S. L. (1991). The reasoning voter: Communication and persuasion in presidential campaigns. Chicago, IL: University of Chicago Press. Poutvaara, P. (2014). Facial appearance and leadership: An overview and challenges for new research. The Leadership Quarterly, 25, 801–804. https://doi.org/10.1016/j.leaqua.2014. 08.003 Poutvaara, P., Jordahl, H., & Berggren, N. (2009). Faces of politicians: Babyfacedness predicts inferred competence but not electoral success. Journal of Experimental Social Psychology, 45, 1132–1135. https://doi.org/10.1016/j.jesp.2009.06.007 Praino, R., Stockemer, D., & Ratis, J. (2014). Looking good or looking competent? Physical appearance and electoral success in the 2008 congressional elections. American Politics Research, 42, 1096–1117. https://doi.org/10.1177/ 1532673x14532825

Ó 2018 Hogrefe Publishing

A. Fleischmann et al., Glasses Increase Electoral Success

R Core Team. (2017). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org/ Rosar, U., Klein, M., & Beckers, T. (2008). The frog pond beauty contest: Physical attractiveness and electoral success of the constituency candidates at the North Rhine-Westphalia state election of 2005. European Journal of Political Research, 47, 64–79. https://doi.org/10.1111/j.1475-6765.2007.00720.x Rosenberg, S. W., Kahn, S., Tran, T., & Le, M.-T. (1991). Creating a political image: Shaping appearance and manipulating the vote. Political Behavior, 13, 345–367. https://doi.org/10.2307/ 586121 Rosenberg, S. W., & McCafferty, P. (1987). The image and the vote: Manipulating voters’ preferences. Public Opinion Quarterly, 51, 31–47. Rule, N. O., Ambady, N., Adams, R. B. Jr., Ozono, H., Nakashima, S., Yoshikawa, S., & Watabe, M. (2010). Polling the face: Prediction and consensus across cultures. Journal of Personality and Social Psychology, 98, 1–15. https://doi.org/10.1037/ a0017673 Sczesny, S., & Kühnen, U. (2004). Meta-cognition about biological sex and gender-stereotypic physical appearance: Consequences for the assessment of leadership competence. Personality and Social Psychology Bulletin, 30, 13–21. https://doi. org/10.1177/0146167203258831 Shephard, M., & Johns, R. (2008). Candidate image and electoral preference in Britain. British Politics, 3, 324–349. https://doi. org/10.1057/bp.2008.8 Spisak, B. R., Blaker, N. M., Lefevre, C. E., Moore, F. R., & Krebbers, K. F. B. (2014). A face for all seasons: Searching for context-specific leadership traits and discovering a general preference for perceived health. Frontiers in Human Neuroscience, 8, 792. https://doi.org/10.3389/fnhum.2014. 00792 Spisak, B. R., Dekker, P. H., Krüger, M., & van Vugt, M. (2012). Warriors and peacekeepers: Testing a biosocial implicit leadership hypothesis of intergroup relations using masculine and feminine faces. PLoS One, 7, e30399. https://doi.org/10.1371/ journal.pone.0030399 Spisak, B. R., Homan, A. C., Grabo, A., & van Vugt, M. (2012). Facing the situation: Testing a biosocial contingency model of leadership in intergroup relations using masculine and feminine faces. The Leadership Quarterly, 23, 273–280. https://doi. org/10.1016/j.leaqua.2011.08.006 Spisak, B. R., Nicholson, N., & van Vugt, M. (2011). Leadership in organizations: An evolutionary perspective. In G. Saad (Ed.), Evolutionary psychology in the business sciences (pp. 165–190). Berlin, Heidelberg, Germany: Springer. Stratmann, T. (2005). Some talk: Money in politics. A (partial) review of the literature. In W. F. Shughart II & R. D. Tollison (Eds.), Policy challenges and political responses (pp. 135–156). Dordrecht, The Netherlands: Springer. Sussman, A. B., Petkova, K., & Todorov, A. (2013). Competence ratings in US predict presidential election outcomes in Bulgaria. Journal of Experimental Social Psychology, 49, 771– 775. https://doi.org/10.1016/j.jesp.2013.02.003 Swinford, S. (2013). David Cameron wears glasses in public for first time. The Telegraph. Retrieved from http://www.telegraph. co.uk/news/politics/david-cameron/10300348/DavidCameron-wears-glasses-in-public-for-first-time.html Terry, R. L., & Krantz, J. H. (1993). Dimensions of trait attributions associated with eyeglasses, men’s facial hair, and women’s hair length. Journal of Applied Social Psychology, 23, 1757– 1769. https://doi.org/10.1111/j.1559-1816.1993.tb01064.x

Ó 2018 Hogrefe Publishing

Terry, R. L., & Kroger, D. L. (1976). Effects of eye correctives on ratings of attractiveness. Perceptual and Motor Skills, 42, 562– 562. https://doi.org/10.2466/pms.1976.42.2.562 Thompson, E. P., Roman, R. J., Moskowitz, G. B., Chaiken, S., & Bargh, J. A. (1994). Accuracy motivation attenuates covert priming: The systematic reprocessing of social information. Journal of Personality and Social Psychology, 66, 474–489. https://doi.org/10.1037/0022-3514.66.3.474 Thornton, G. R. (1943). The effect upon judgments of personality traits of varying a single factor in a photograph. The Journal of Social Psychology, 18, 127–148. https://doi.org/10.1080/ 00224545.1943.9921704 Thornton, G. R. (1944). The effect of wearing glasses upon judgments of personality traits of persons seen briefly. Journal of Applied Psychology, 28, 203–207. https://doi.org/10.1037/ h0055862 Todorov, A., Mandisodza, A. N., Goren, A., & Hall, C. C. (2005). Inferences of competence from faces predict election outcomes. Science, 308, 1623–1626. https://doi.org/10.1126/ science.1110589 Todorov, A., Olivola, C. Y., Dotsch, R., & Mende-Siedlecki, P. (2015). Social attributions from faces: Determinants, consequences, accuracy, and functional significance. Annual Review of Psychology, 66, 519–545. https://doi.org/10.1146/annurev-psych113011-143831 van Vugt, M. (2006). Evolutionary origins of leadership and followership. Personality and Social Psychology Review, 10, 354–371. https://doi.org/10.1207/s15327957pspr1004_5 van Vugt, M. (2014). On faces, gazes, votes, and followers: Evolutionary psychological and social neuroscience approaches to leadership. In J. Decety & Y. Christen (Eds.), New frontiers in social neuroscience (pp. 93–110). Cham Switzerland: Springer. van Vugt, M., Johnson, D. D. P., Kaiser, R., & O’Gorman, R. (2008). Evolution and the social psychology of leadership: The mismatch hypothesis. In C. L. Hoyt, G. R. Goethals, & D. R. Forsythe (Eds.), Leadership at the crossroads (Vol. 1, pp. 267–282). Westport, CT: Praeger. van Vugt, M., & Ronay, R. (2013). The evolutionary psychology of leadership: Theory, review, and roadmap. Organizational Psychology Review, 4, 74–95. https://doi.org/10.1177/ 2041386613493635 Verhulst, B., Lodge, M., & Lavine, H. (2010). The attractiveness halo: Why some candidates are perceived more favorably than others. Journal of Nonverbal Behavior, 34, 111–117. https://doi. org/10.1007/s10919-009-0084-z Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36, 1–48. https://doi.org/10.18637/jss.v036.i03 Wegner, D. M., & Bargh, J. A. (1998). Control and automaticity in social life. In D. T. Gilbert, S. T. Fiske, G. Lindzey, D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vols. 1 and 2, pp. 446–496). New York, NY: McGraw-Hill. Wojciszke, B., & Klusek, B. (1996). Moral and competence-related traits in political perception. Polish Psychological Bulletin, 27, 319–324. Yzerbyt, V. Y., Schadron, G., Leyens, J.-P., & Rocher, S. (1994). Social judgeability: The impact of meta-informational cues on the use of stereotypes. Journal of Personality and Social Psychology, 66, 48–55. https://doi.org/10.1037/0022-3514. 66.1.48 Zebrowitz, L. A., & Montepare, J. M. (2005). Appearance DOES matter. Science, 308, 1565–1566. https://doi.org/ 10.1126/science.1114170

Social Psychology (2019), 50(1), 38–52

History Received November 30, 2017 Revision received July 23, 2018 Accepted July 23, 2018 Published online December 10, 2018 Acknowledgments We would like to thank Niek de Boer and Andrea Przegendza for developing the stimuli, and Agneta Gille for letting us include her pictures in this paper. Funding This research was partly funded by a Leo-Spitzer-Prize for Junior Researchers awarded to Joris Lammers by the University of Cologne.

A. Fleischmann et al., Glasses Increase Electoral Success

ORCID Alexandra Fleischmann https://orcid.org/0000-0001-8290-4561 Alexandra Fleischmann Sozialpsychologie Social Cognition Center Cologne University of Cologne Richard-Strauss-Str. 2 50931 Köln Germany alexandra.fleischmann@uni-koeln.de

Authorship Alexandra Fleischmann, Joris Lammers, Janka I. Stoker, and Harry Garretsen designed the study concept. A. Fleischmann and J. Lammers designed the studies, and A. Fleischmann collected and analyzed the data. A. Fleischmann drafted the manuscript; all authors edited the manuscript and approved of the final version.

Social Psychology (2019), 50(1), 38–52

Ó 2018 Hogrefe Publishing

Original Article

Attribution of Egoistic Versus Altruistic Motives to Acts of Helping The Role of the Helper’s Status and the Act’s Intended Publicity Birte Siem

and Stefan Stürmer

Department of Psychology, FernUniversität in Hagen, Germany

Abstract: This research investigated the effects of helpers’ status and the intended publicity of the helping act on observers’ attributions of altruistic versus egoistic motives to helpers. Results from two studies (overall N = 670), in which we varied helpers’ status (low vs. high) and intended publicity of the helping act (private vs. public) suggest that observers attribute more egoistic (and less altruistic) motives to the helping act when helpers intended to make their helping public (vs. keeping it private) and were of high (vs. low) status. Moreover, status and publicity had a negative indirect effect on observers’ willingness to assist helpers via increased attribution of predominantly egoistic motives. Implications of these findings for theory and real-world helping will be discussed. Keywords: status, publicity, altruistic motives, egoistic motives, helping, volunteering

Acts of prosocial behavior such as helping or volunteering are important for society’s collective well-being and are often referred to as the “social glue” that binds people together and enables the emergence of harmonious and productive relationships (e.g., Lay & Hoppmann, 2015). Acts of prosocial behavior do not necessarily produce unanimously positive responses, though. Reactions to Band Aid 30, the 2014 incarnation of the charity pop group Band Aid founded in 1984, are a case in point. Band Aid 30 was founded by Bob Geldof and Midge Ure with the goal of releasing a version of the commercial pop song “Do They Know It’s Christmas?” to raise money to battle the Ebola epidemic. Even though sales of the Band Aid 30 single raised millions of dollars for Ebola victims, the public response to the campaign was partly negative. Specifically, some commentators contended that the celebrities’ underlying motive was not concern for other people’s welfare but boosting their own public images as responsible, caring, and humanitarian citizens (see Stürmer, Rohmann, & van der Noll, 2016). Similar motives have been attributed to other celebrities who volunteered for or initiated charity campaigns. For instance, Angelina Jolie, who regularly visits sites of deprivation around the globe and has established

Ó 2018 Hogrefe Publishing

a number of charity foundations, is frequently accused of using her “celebrity do-gooding” to improve her public image as a caring cosmopolitan (e.g., Littler, 2008). The nature of a helper’s motivation in terms of egoistic relative to altruistic concerns is a leading theme in research on human prosocial behavior (e.g., Batson, 1991; Penner, Dovidio, Piliavin, & Schroeder, 2005). Importantly, this research shows that people do not only act upon these motives (for an overview, see Batson, 2011), but – as illustrated in the above examples – also ascribe them to the good deeds of others. Specifically, a number of studies within social, organizational, and developmental psychology show that people attribute helping behavior, including volunteering, to egoistic and altruistic motives of the actor and that such attributions affect their evaluations of the helper and their willingness to assist the helper or to join the helper’s cause (i.e., attribution to altruistic motives leads to a more positive evaluation and enhances willingness to help; e.g., Plewa, Conduit, Quester, & Johnson, 2015; Rodell & Lynch, 2016; Schlenker, Hallam, & McCown, 1983). Research on do-gooder derogation demonstrates, for instance, that prosocial actors who benefit from their good deeds are evaluated less positively (e.g., less charitable) than those who do

Social Psychology (2019), 50(1), 53–66 https://doi.org/10.1027/1864-9335/a000360

not benefit (e.g., Lin-Healy & Small, 2013; see also the “tainted-altruism effect,” Newman & Cain, 2014). In sum, people seem to be particularly susceptible to cues of the helpers’ egoistic versus altruistic motives. In the present research, we focus on the role of two potential attributional cues that have only recently received empirical attention in the research literature, the perceived intended publicity of the helping act and the helper’s status.

Publicity and the Attribution of Prosocial Behavior to Altruistic Versus Egoistic Motives As helping is a behavior that is highly regarded in most social contexts, when helping in public, helpers usually receive social benefits in return (social recognition, praise, gratitude, and the like). Ironically, however, when observes know (or believe to know) that the helper intentionally decided to make the good deed public, they should be less willing to recognize the helper’s good intentions (presumably because they assume that the helper’s primary motivation was to obtain social benefits). In line with these assumptions, research by Heyman and colleagues shows that children who were told about another child who helped a peer judged the helper and the prosocial act more favorably when told that the actor purposefully provided help privately rather than publicly and explained their judgments by referring to egoistic impression management motives (Heyman, Barner, Heumann, & Schenck, 2014; Heyman et al., 2016). Likewise, research by Berman, Levine, Barasch, and Small (2015, Studies 1a and 1b) demonstrated that, when others are already aware of actors’ good deeds, actors who brag about their deeds (e.g., post about them on social media) are perceived as less altruistic than actors who did not brag. Complementary evidence comes from organizational and consumer psychology studies on corporate social responsibility (CSR) activities, that is, “voluntary corporate actions designed to improve social conditions” (Mackey, Mackey, & Barney, 2007, p. 818) such as reducing waste or engaging in humanitarian aid. Yoon, Zeynep, and Schwarz (2006) found, for instance, that people ascribed less altruistic motives to a company’s CSR activities when they learned about them through company advertising rather than an unbiased source and that this, in turn, led to more negative company evaluations. Taken together, these pieces of evidence suggest that observers ascribe more egoistic motives (and less altruistic motives) to helpers who actively make their good deeds public (e.g., by talking about them) than to helpers who keep their good deeds private or whose good deeds just happened to be witnessed by others. Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Status and the Attribution of Prosocial Behavior to Altruistic Versus Egoistic Motives Another potentially relevant cue for observers’ attributions of motives is the helper’s status. Status can result from different hierarchical bases, with socioeconomic status (SES) being among the most widely researched ones. SES is based on objective criteria, including a target’s income, education, and occupation (e.g., Adler & Stewart, 2007). SES has been acknowledged as an important factor in previous research on prosocial behavior. Yet, these studies have mainly focused on how a person’s SES affects his or her willingness to help (e.g., Guinote, Cotzia, Sandhu, & Siwa, 2015; Kraus & Callaghan, 2016; Piff, Kraus, Côté, Cheng, & Keltner, 2010), while the question of how this variable affects observers’ attributions of motives to the helper is an underresearched issue. In the present research, we assume that observers tend to ascribe more egoistic motives (and less altruistic motives) to higher-SES helpers. This assumption is informed by the stereotype content model (e.g., Fiske, Cuddy, Glick, & Xu, 2002) which hypothesizes two primary dimensions of stereotype content: warmth and competence. Importantly, a number of studies consistently show that members of high-SES groups (e.g., the rich) are seen as more competent (e.g., more ambitious, confident, and dominant) than lowSES individuals, but also as less warm (e.g., less trustworthy, sincere, and good-natured) (e.g., Fiske et al., 2002; see also Conway, Pizzamiglio, & Mount, 1996; Koch, Imhoff, Dotsch, Unkelbach, & Alves, 2016). The stereotypically perceived lack of warmth might fuel the perception that members of high-SES groups tend to be motivated more by egoistic factors than members of low-SES groups. A similar conclusion can be drawn from sociological approaches suggesting that the status attainment process itself might contribute to the ascription of more egoistic motives to high-SES individuals, because its incentive structure encourages high-SES actors to feign capability and commitment or to assert their superiority and another’s inferiority (Hahl & Zuckermann, 2014). While there is theoretical evidence to assume that higher-SES helpers are perceived as egoistically motivated, so far, we know of only one study which tests the effects of variables related to a helper’s SES on observers’ motive attributions. Specifically, Berman et al. (2015, Study 2) manipulated a helper’s occupation (indicating higher vs. lower SES) and found that observers regarded the higherSES helper as less altruistic than the lower-SES helper. The potential effects of a helper’s status on observers’ attributions of altruistic versus egoistic motives thus deserve further investigation. Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

The Present Research The main aim of the present research was to subject the role of the helper’s status and the intended publicity of the helping act in observers’ attributions of underlying motives (altruistic vs. egoistic) to systematic empirical testing. For this purpose, we conducted two experiments in which we simultaneously manipulated the helper’s SES and whether he intended to make the helping act public or keep it private. We then investigated how these differences affected the extent to which observers attributed the helping behavior to egoistic versus altruistic motives. Further, we examined to what extent SES and intended publicity had indirect effects on participants’ willingness to assist the volunteer via their attributions of egoistic and altruistic motives. Building on our theoretical review, our main assumptions were as follows. First, we expected observers to attribute more altruistic and less egoistic motives to a volunteer who signaled that he or she intended to keep the helping act private vs. make it public. Second, we expected observers to ascribe more egoistic and less altruistic motives to the behavior of a higher-SES versus a lower-SES helper. We did not make any specific assumptions about a potential Status Publicity interaction. Given that different interaction patterns seem plausible (e.g., an additive effect of status and publicity; a moderating effect of status on the publicity-attribution link), we felt justified in keeping this part of our analyses exploratory in nature. The studies were conducted in the context of volunteering to help refugees and are thus also of high practical relevance. A number of studies suggest that people’s motivational attributions of helping behavior also affect their willingness to assist the helper, with altruistic attributions leading to a greater willingness to join the helper’s cause (e.g., Plewa et al., 2015; Rodell & Lynch, 2016). Knowledge about the determinants of these attribution processes can thus be used to help humanitarian campaigns attract “fellow helpers.”

Study 1 The main aim of Study 1 was to test how manipulations of the helper’s SES and the intended publicity of the helping act affected observers’ attributions of altruistic versus egoistic motives to the helper. Participants were presented with a (fictitious) media report about a man volunteering to help refugees in Germany. Inspired by the real-world examples 1

of celebrity charity mentioned in the introduction, we varied the helper’s SES by depicting him either as a celebrity (in the higher-SES condition) or as an average citizen (in the lower-SES condition). Celebrities typically have a high income (“the rich and famous”) what makes them representative for the higher-SES category (e.g., Adler & Stewart, 2007). As a second factor, we varied the intended publicity by telling participants that the helper had either intended to make his voluntary work public or keep it private. We measured participants’ attributions of altruistic versus egoistic motives as well as their willingness to assist the volunteer in order to investigate the relationships between status, publicity, observers’ attributions, and observers’ intentions to join the volunteer’s cause.

Method We report how we determined our sample size, all data exclusions, all manipulations, and all measures for both Study 1 and Study 2 (Simmons, Nelson, & Simonsohn, 2012). Auxiliary measures and additional analyses for both studies are reported in Electronic Supplemental Material, ESM 1. All materials and data files are available on the Open Science Framework (https://osf.io/r3db6/). Participants Following the recommendations of Simmons, Nelson, and Simonsohn (2013), we aimed to have > 50 participants in each of the four conditions. Data collection was thus terminated after 262 participants had completed the study. After excluding nine participants who did not agree to allow their data to be used for analyses, the final sample consisted of 253 participants (71.15% female, Mage = 32.99 years, SD = 11.17). Post hoc power analyses indicated that the study had moderate power (.75) to detect small effects (f = .10) (conducted with G*Power; Faul, Erdfelder, Lang, & Buchner, 2007). Design and Procedure Participants were randomly assigned to the conditions of a 2 (Status: low vs. high) 2 (Publicity: private vs. public) between-subjects design. The study was conducted online and was introduced to participants as concerning the evaluation of media reports about current social issues. In line with this cover story, participants were told that they would be randomly assigned to one of several media reports. In truth, all participants were assigned a short text about a man who volunteers handing out food to refugees at German refugee accommodations.1 All participants received

To rule out the possibility that our results were influenced by different gender relationships between the target (male) and participants (male and female), we reran our main analyses in Studies 1 and 2 while controlling for participants’ gender. Including gender as a covariate in our analyses left our main findings virtually unchanged. Moreover, models including gender as a potential moderator in our main analyses found that gender did not moderate any of the observed effects.

Ó 2018 Hogrefe Publishing

Social Psychology (2019), 50(1), 53–66

the same text except for the parts used to manipulate status and publicity (see ESM 2 for the full text). Status Manipulation To manipulate the volunteer’s SES, we varied whether the volunteer was described in the text as a famous actor (high-status condition) or an average citizen (low-status condition). Publicity Manipulation To manipulate the intended publicity of the helping act, the text either stated that the man posted information about his project on social media (public condition) or that he did not intend to make his volunteering public, but the tabloid press found out about it and decided to publish it (private condition). After some filler items about participants’ media usage and attitudes toward journalism, which were included to strengthen the credibility of the cover story, participants filled in a questionnaire containing the central variables of our study. At the end of the study, participants answered some sociodemographic questions; earned course credits; and were fully debriefed. Measures Unless otherwise noted, we used 7-point Likert scales for all measures (from 1 = do not agree at all to 7 = fully agree). We created composite scores for each participant and each measure by averaging across all corresponding items.

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

We ran a principal axis analysis with oblimin rotation on the eight items to test whether altruistic motives and egoistic motives loaded on two different factors. This analysis yielded two factors with eigenvalues greater than 1 (1.47, 4.13) that explained 69.93% of the variance. The items measuring egoistic motives exhibited high loadings on the first factor ( .55) and weak loadings on the second factor ( |.17|), while the items measuring altruistic motives exhibited high loadings on the second factor ( .68) and weak loadings on the first factor ( |.18|). We thus felt justified in building separate scales for altruistic and egoistic motives. Perceived Status In order to check the effectiveness of the status manipulation, we asked participants to indicate what status they would assign to the helping person on a 5-point scale (1 = very low, 2 = low, 3 = medium, 4 = high, 5 = very high). Willingness to Assist the Volunteer Two items measured participants’ willingness to assist the volunteer depicted in the media report: “I would help the depicted person with words and deeds” and “I would provide a financial contribution to the depicted person to assist him with his helping” (r[251] = .55, p < .001).2

Results Table 1 presents intercorrelations, means, and standard deviations for all theoretically relevant variables.

Egoistic Motives After participants read the vignette and worked through the filler items, they were asked to rate the volunteer’s motives by indicating their agreement with eight different statements. Four statements referred to egoistic motives for volunteerism (adapted from Ellen, Webb, & Mohr, 2006): “The helper wants to use the helping act for his own publicity”, “. . .expects benefits for his own career and reputation from his helping behaviour”, “. . .wants to ease his conscience”, and “. . .wants to impress other people” (Cronbach’s α = .86).

Manipulation Check A 2 (Status: low vs. high) 2 (Publicity: private vs. public) analysis of variance (ANOVA) on perceived status yielded a significant main effect of status, F(1, 249) = 57.49, p < .001, η2p = .19, 95% CI [0.11, 0.27], indicating that, as intended, participants in the high-status condition assigned a higher status to the helper (M = 4.12, SD = 0.79) than participants in the low-status condition (M = 3.36, SD = 0.78). The main effect of publicity and the interaction effect were not significant, both ps .788.

Altruistic Motives These statements were presented intermixed with four statements referring to altruistic motives (adapted from different scales; Ellen et al., 2006; Sun, 2014; Walker, Heere, Parent, & Drane, 2010): “The helper has a long-term interest in society,” “. . .likes to work for the welfare of others,” “. . .is trying to give something back to society,” and “. . .feels a moral obligation to help” (Cronbach’s α = .82).

Main Analyses Effects on Attributions To test the effects of our manipulations on participants’ attributions, we conducted a 2 (Status: low vs. high) 2 (Publicity: private vs. public) 2 (Attribution: egoistic vs. altruistic) mixed ANOVA with attributed egoistic vs. altruistic motives as a within-subject factor. Four significant effects were found. Specifically, the analysis revealed a

For exploratory reasons, our studies contained an additional dependent variable, participants’ evaluation of the help provided by the volunteer. For the sake of simplicity, we do not report this variable and related findings in the article. These can be found in ESM 1.

Social Psychology (2019), 50(1), 53–66

Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Table 1. Intercorrelations, means (M), and standard deviations (SD) for all theoretically relevant variables (Study 1) Variables

–

1. Status manipulation 2. Publicity manipulation

.44**

.26**

.41**

.44**

.07

.23**

.16**

.23**

.03

.13*

3. Egoistic motives

.49**

4. Altruistic motives

.88**

.26**

.34*

.84**

.04

.56**

5. Attribution difference score

.18**

6. Perceived status

.51** .03

7. Willingness to assist the volunteer M

–

3.31

5.05

1.74

3.76

4.16

–

1.46

1.26

2.35

0.87

1.66

Notes. The status manipulation variable was coded 1 (low-status condition) and 1 (high-status condition), and the publicity manipulation variable was coded 1 (private condition) and 1 (public condition). The attribution difference score was computed by subtracting ratings of altruistic motives from ratings of egoistic motives and can thus vary between 6 and 6. *p < .05; **p < .01 (two-tailed).

significant main effect of status, F(1, 249) = 11.81, p < .001, η2p = .05, 95% CI [0.01, 0.11], indicating that, independent of the type of attribution, participants in the high-status conditions made stronger attributions (M = 4.33, SD = 0.74) than participants in the low-status conditions (M = 4.01, SD = 0.61). Moreover, the main effect of attribution was significant, F(1, 249) = 169.28, p < .001, η2p = .41, 95% CI [0.32, 0.48], indicating that participants made more attributions to altruistic (M = 5.05, SD = 1.26) than egoistic motives (M = 3.31, SD = 1.46) in general. More importantly, these main effects were qualified by a significant Status Attribution interaction, F(1, 249) = 49.85, p < .001, η2p = .17, 95% CI [0.09, 0.25]: As expected, participants attributed more egoistic motives to the helping act when the helper was of high status (M = 3.91, SD = 1.43) than when he was of low status (M = 2.63, SD = 1.18), F(1, 249) = 57.22, p < .001, η2p = .19, 95% CI [0.11, 0.27], while the reverse was true for altruistic motives. Here, participants attributed more altruistic motives to the helping behavior when the helper was of low status (M = 5.39, SD = 1.04) than when he was of high status (M = 4.74, SD = 1.35), F(1, 249) = 17.47, p < .001, η2p = .07, 95% CI [0.02, 0.13]. In addition, the ANOVA revealed a significant Publicity Attribution interaction, F(1, 249) = 15.28, p < .001, η2p = .06, 95% CI [0.01, 0.12], indicating that, as expected, participants attributed more egoistic motives to the helping act when they learned that the helper himself made it public (M = 3.69, SD = 1.52) than when they learned that the helper wanted to keep it private (M = 3.01, SD = 1.35), F(1, 249) = 15.94, p < .001, η2p = .06, 95% CI [0.02, 0.13], while the reverse was true for altruistic motives. Here, participants attributed more altruistic motives to the helping behavior when the helper’s intention was to keep the helping act private (M = 5.23, SD = 1.20) than when he made it public (M = 4.81, SD = 1.29), F(1, 249) = 6.36, p = .012, η2p = .03, 95% CI [0.00, 0.07]. The main effect of publicity, the Status Publicity interaction effect, and Ó 2018 Hogrefe Publishing

the Status Publicity Attribution interaction effect were not significant, all ps .127. To further explore the effects of our manipulations, we conducted simple effects analyses. The results suggested an additive effect: Participants’ attributions to egoistic motives were significantly highest when the helper was of high status and made his helping act public, for all three comparisons, t(249)s |2.16|, ps .032, η2ps .02, 95% CIs [ 0.00, 0.06], and significantly lowest when he was of low status and intended to keep his help private, for all three comparisons, t(249)s 3.45, ps .001, η2ps .05, 95% CIs [ 0.01, 0.11]. The reverse pattern emerged for attributions of altruistic motives. Specifically, participants in the high status/public condition reported (marginally) significantly lowest levels of attributed altruistic motives, for all three comparisons, t(249)s 1.92, ps .056, η2ps .02, 95% CIs [ 0.00, 0.06], while participants in the low status/private condition reported significantly highest levels of attributed altruistic motives, for all three comparisons, t(249)s |1.67|, ps .098, η2ps .05, 95% CIs [ 0.00, 0.05]. Effects on Willingness to Assist A 2 (Status: low vs. high) 2 (Publicity: private vs. public) ANOVA on participants’ willingness to assist the volunteer revealed a significant main effect of publicity, F(1, 249) = 4.29, p = .039, η2p = .02, 95% CI [0.00, 0.06], indicating that participants in the private condition were more willing to assist (M = 4.34, SD = 1.64) than participants in the public condition (M = 3.92, SD = 1.67). The main effect of status and the interaction effect were not significant, both ps .178. Indirect Effects In line with previous research on the mediating role of observers’ attributions (e.g., Plewa et al., 2015; Rodell & Lynch, 2016), we tested whether our manipulations and Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Figure 1. Path model of the indirect effects of status, publicity, and Status Publicity on willingness to assist via attribution difference score (higher values indicate greater predominance of egoistic motives) (Study 1). Coefficients are unstandardized regression weights. Total effects are reported in parentheses *p < .05; **p < .01; ***p < .001.

their interaction had indirect effects on participants’ willingness to assist the volunteer via their attributions of egoistic and altruistic motives.3 For this purpose, we first computed a difference score by subtracting participants’ ratings of the volunteer’s altruistic motives from their ratings of his egoistic motives (i.e., higher positive values indicate greater dominance of ascribed egoistic motives over altruistic motives). We then conducted a paths analysis with the status manipulation (coded 1 = low status, 1 = high status), the publicity manipulation (coded 1 = private, 1 = public) and the Status Publicity interaction term as the predictors, the attribution difference score as the mediator, and willingness to assist as the criterion (see Figure 1).4 The results revealed that the status manipulation, b = 0.94, t(249) = 7.06, p < .001, pr2 = .17, 95% CI [0.09, 0.25], and the publicity manipulation, b = 0.52, t(249) = 3.91, p < .001, pr2 = .06, 95% CI [0.01, 0.12], were significantly and positively related to the attribution difference score, indicating that high-status volunteers (vs. low-status volunteers) and intended publicity (vs. intended privacy) both led to the stronger ascription of predominantly egoistic motives. The attribution difference score, in turn, was significantly negatively related to willingness to assist the volunteer, b = 0.41, t(248) = 9.54, p < .001, pr2 = .27, 95% CI [0.18, 0.35], indicating that the more participants ascribed egoistic (compared to altruistic) motives to the helper, the less their willingness to assist was. We tested 3

the indirect effects of status and publicity using a bootstrap estimation approach with 5000 samples and 95% bias-corrected confidence intervals (using PROCESS; Hayes, 2017). Importantly, both indirect effects were significant, status: IE = .38, 95% CI [ 0.53, 0.26]; publicity: IE = .21, 95% CI [ 0.34, 0.11]. The Status Publicity interaction, in contrast, was not significantly related to the mediator, b = 0.08, t(249) = 0.59, p = .558, nor was its indirect effect significant, IE = 0.03, 95% CI [ 0.08, 0.14].

Discussion Study 1 provided solid evidence for our assumptions on the role of the helper’s SES and the intended publicity of the helping act on observers’ attributions of altruistic vs. egoistic motives to the helper. Specifically, participants made higher levels of egoistic attributions (and lower levels of altruistic attributions) when the helper was depicted as a celebrity (compared to an average citizen) and when he intended to make his helping public (compared to keeping it private). Moreover, simple effects analyses provided some evidence for an additive effect of status and publicity, with participants in the high status/public condition exhibiting highest levels of egoistic attributions and lowest levels of altruistic attributions, while the reverse was true for participants in the low status/private condition.

We felt justified in investigating potential indirect effects of these variables even though the total effects of status and of Status Publicity on willingness to help were not significant. There is relatively wide consensus among statisticians that the total effect should not be a prerequisite for tests of indirect effects (e.g., Rucker, Preacher, Tormala, & Petty, 2011). Please note that, in both studies, we rerun the indirect effects analyses with perceived egoistic and altruistic motives as parallel mediators instead of the attribution difference score as a single mediator. Results of these additional analyses parallel results from analyses using the attribution difference score as a mediator and are reported in ESM 1.

Social Psychology (2019), 50(1), 53–66

Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

In addition, indirect effects analyses confirmed that SES and publicity have independent indirect effects on participants’ willingness to assist the volunteer via specific attributions (see also Plewa et al., 2015; Rodell & Lynch, 2016): Both SES and publicity had a negative indirect effect on willingness to assist the volunteer via increased attribution of predominantly egoistic motives.

Study 2 The main aim of Study 2 was to replicate the findings of Study 1 and enhance their generalizability. For this purpose, we implemented the following changes. First, we used a different SES manipulation than in Study 1 based not on the helper’s celebrity status but on his occupation. Thus, we wanted to rule out the possibility that our findings in Study 1 were mainly due to factors specific to the group of celebrities (e.g., frequent media reports about celebrities volunteering to help refugees at the time our studies were conducted; e.g., Apin, 2016). Occupation is one of the fundamental components of SES (e.g., Kraus & Stephens, 2012; Lareau & Conley, 2008) and, as an individual’s occupation usually is highly correlated with his or her income and level of education, often includes the two other components. Second, we used a more comprehensive measure of participants’ attributions of altruistic vs. egoistic motives. Specifically, while the measure of egoistic motives used in Study 1 focused mainly on impression management and reputation concerns, the measure employed in Study 2 also encompassed other egoistic concerns (e.g., enhancing selfesteem, distracting from one’s own troubles; see also Clary et al., 1998).

Method Participants Given that Study 1 was only moderately powered, we aimed to have > 80 participants in each of the four conditions in Study 2. Data collection was thus terminated after 428 participants had completed the study. After excluding 11 participants who did not agree to allow their data to be used for analyses, the final sample consisted of 417 people (71.46% female, Mage = 31.56 years, SD = 10.09). This sample size was sufficient to detect even small effects (f = .10) with high power (.87) (Faul et al., 2007). Design and Procedure As in Study 1, participants were randomly assigned to the conditions of a 2 (Status: low vs. high) 2 (Publicity: private vs. public) between-subjects design. The study was conducted online. The procedure was virtually identical to the procedure of Study 1, and the text of the fictitious Ó 2018 Hogrefe Publishing

newspaper article closely resembled the text used in Study 1 (see ESM 2 for the full text). Status Manipulation To manipulate the volunteer’s SES, we varied the occupation of the volunteer depicted in the newspaper article. He was a stockbroker in the higher-SES condition and a carpenter in the lower-SES condition. Publicity Manipulation To manipulate the intended publicity of the helping act, the text either stated that the man himself had informed the tabloid press about his voluntary project (public condition) or that he did not intend to make his volunteering public, but the tabloid press found out about it and decided to publish it (private condition). Measures Unless otherwise noted, we used 5-point Likert scales for all measures (from 1 = do not agree at all to 5 = fully agree) and created composite scores by averaging across all corresponding items. Altruistic Motives After participants had read the vignette, they were asked to indicate their agreement with 16 different statements. Eight statements referred to altruistic motives for volunteerism (adapted from different scales; e.g., Oostlander, Güntert, van Schie, & Wehner, 2014; Simon, Stürmer, & Steffens, 2000; Van de Vliert, Huang, & Levine, 2004): “The helper feels compassion for needy people,” “. . .is expressing his concern for other people,” “. . .wants to identify with suffering people,” “. . .shows sympathy for others’ fate,” “. . .considers it important to help other people,” “. . .feels a responsibility to help people who are worse off than himself,” “. . .wants to contribute to disadvantaged people’s hope and dignity,” and “. . .is concerned about other people’s welfare.” Egoistic Motives These statements were presented intermixed with eight statements referring to egoistic motives (adapted from different scales; e.g., Carlo & Randall, 2002): “The helper expects benefits to himself from his helping behaviour,” “. . .wants to improve his self-esteem,” “. . .wants to win approval,” “. . .enjoys the attention,” “. . .is praising himself for being a responsible citizen,” “. . .thinks that he can assuage his feelings of guilt about being better off than others,” “. . .helps in order to distract himself from his own troubles,” and “. . .helps in order to avoid looking bad in front of others.” A principal axis analysis with oblimin rotation on the 16 items yielded a three-factor solution, with two items showing loadings .43 and cross-loadings on two factors (“The Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Table 2. Intercorrelations, means (M), and standard deviations (SD) for all theoretically relevant variables (Study 2) Variables

–

1. Status manipulation 2. Publicity manipulation

.12*

.06

.11**

.50**

.12*

.35**

.10*

.29**

.004

.07

3. Egoistic motives

.37**

4. Altruistic motives

.87**

.01

.25*

.77**

.03

.36**

5. Attribution difference score

.02

6. Perceived status

.36** .03

7. Willingness to assist the volunteer M

–

2.52

4.19

1.67

3.62

3.25

–

0.90

0.69

1.32

0.75

1.07

helper wants to identify with suffering people” and “The helper helps in order to avoid looking bad in front of others”). We thus reran the analysis without these two items. This analysis yielded two factors with eigenvalues greater than 1 (2.72, 5.97) that explained 62.07% of the variance. The items measuring altruistic motives exhibited high loadings on the first factor ( .64) and weak loadings on the second factor ( .14), while the items measuring egoistic motives exhibited satisfactory loadings on the second factor ( .48) and weak loadings on the first factor ( |.13|). We thus felt justified in building separate scales for altruistic motives (Cronbach’s α = .90) and egoistic motives (Cronbach’s α = .88) consisting of seven items each. However, we conducted all analyses reported in the Results section with the eight-item scales as well. The results were virtually identical to those observed with the seven-item scales. Perceived Status and Willingness to Assist the Volunteer Perceived status and willingness to assist the volunteer (r[415] = .59, p < .001) were assessed with items identical to those used in Study 1.

Results Table 2 presents intercorrelations, means, and standard deviations for all theoretically relevant variables. Manipulation Check A 2 (Status: low vs. high) 2 (Publicity: private vs. public) ANOVA on perceived status yielded a significant main effect of status, F(1, 413) = 135.25, p < .001, η2p = .25, 95% CI [0.18, 0.31], indicating that, as intended, participants in the high-status condition assigned a higher status to the target (M = 4.00, SD = 0.71) than participants in the low-status condition (M = 3.26, SD = 0.60). The main effect of publicity and the interaction effect were not significant, both ps .263. Social Psychology (2019), 50(1), 53–66

Main Analyses Effects on Attributions A 2 (Status: low vs. high) 2 (Publicity: private vs. public) 2 (Attribution: egoistic vs. altruistic) mixed ANOVA with attributed egoistic versus altruistic motives as a within-subject factor yielded five (marginally) significant effect. Specifically, in accordance with the findings of Study 1, the analysis revealed a marginally significant main effect of status, F(1, 413) = 2.76, p = .098, η2p = .01, 95% CI [0.00, 0.03], indicating that participants in the high-status conditions generally tended to make slightly stronger attributions (M = 3.39, SD = 0.45) than participants in the lowstatus conditions (M = 3.32, SD = 0.45), and a significant effect of attribution, F(1, 413) = 726.16, p < .001, η2p = .64, 95% CI [0.59, 0.68], indicating that participants made more attributions to altruistic (M = 4.19, SD = 0.69) than egoistic motives (M = 2.52, SD = 0.90) in general. Moreover, the main effect of publicity was significant, F(1, 413) = 31.02, p < .001, η2p = .07, 95% CI [0.03, 0.12], indicating that participants in the public condition made stronger attributions in general (M = 3.48, SD = 0.44) than participants in the private condition (M = 3.24, SD = 0.45). More importantly, these main effects were qualified by a Status Attribution and a Publicity Attribution interaction. Specifically, as indicated by a significant Status Attribution interaction, F(1, 413) = 5.64, p = .018, η2p = .01, 95% CI [0.00, 0.04], participants attributed more egoistic motives to the helping act when the helper was of high status (M = 2.63, SD = 0.90) than when he was of low status (M = 2.41, SD = 0.89), F(1, 413) = 6.70, p = .008, η2p = .02, 95% CI [0.00, 0.05], while they did not differ on the attribution of altruistic motives F(1, 413) = 1.24, p = .267. Thus, contrary to expectations and in contrast to Study 1, participants in the low-status (M = 4.23, SD = 0.66) and the highstatus conditions (M = 4.15, SD = 0.72) made equally strong attributions to altruistic motives. In addition, the ANOVA revealed a significant Publicity Attribution interaction, F(1, 413) = 38.67, p < .001, η2p = .09, 95% CI [0.04, 0.14], indicating that, as in Study 1, Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

participants attributed more egoistic motives to the helping act when they learned that the helper himself made it public (M = 2.84, SD = 0.83) than when they learned that the helper wanted to keep it private (M = 2.21, SD = 0.86), F(1, 413) = 57.27, p < .001, η2p = .12, 95% CI [0.07, 0.18], while the reverse was true for altruistic motives. Here, participants attributed more altruistic motives to the helping behavior when the helper’s intention was to keep the helping act private (M = 4.27, SD = 0.69) than when he made it public (M = 4.12, SD = 0.69), F(1, 413) = 4.54, p = .034, η2p = .01, 95% CI [0.00, 0.04]. As in Study 1, the Status Publicity and the Status Publicity Attribution interaction effects were not significant, both ps .228. As in Study 1, results from additional simple effects analyses were in line with an additive effect. Participants’ attributions of egoistic motives were (marginally) significantly highest in the high status/public condition (for all three comparisons, t(413)s |1.89|, ps .060, η2ps .01, 95% CIs [ 0.00, 0.04]), and (marginally) significantly lowest in the low-status/private condition (for all three comparisons, t(413)s 1.85, ps .065, η2ps .01, 95% CIs [ 0.00, 0.03]). A different pattern emerged for participants’ attributions of altruistic motives. Participants in the high status/public condition reported lowest levels of attributed altruistic motives, but, in contrast to Study 1, they were only significantly different from levels in the low status/private condition, t(413) = 2.29, p = .022, η2p = .01, 95% CI [0.00, 0.04] (for the remaining two comparisons, t(413)s |0.45|, ps .654). As in Study 1, participants in the low-status/private condition reported (marginally) significantly highest levels of attributed altruistic motives (for all three comparisons, t(413)s |1.85|, ps .065, η2ps .01, 95% CIs [ 0.00, 0.03]).

Effects on Willingness to Assist A 2 (Status: low vs. high) 2 (Publicity: private vs. public) ANOVA on participants’ willingness to assist the volunteer revealed a significant main effect of status, F(1, 413) = 6.02, p = .015, η2p = .01, 95% CI [0.00, 0.05], indicating that participants in the low-status condition were more willing to assist the volunteer (M = 3.38, SD = 1.08) than participants in the high-status condition (M = 3.12, SD = 1.04). This main effect was qualified by a significant interaction, F(1, 413) = 5.08, p = .025, η2p = .01, 95% CI [0.00, 0.04], indicating that participants presented with a low-status volunteer were more willing to assist him when he intended to keep his volunteering private (M = 3.56, SD = 1.07) than when he made it public (M = 3.19, SD = 1.07), F(1, 413) = 6.38, p = .012, η2p = .02, 95% CI [0.00, 0.05], while publicity had no effect on participants presented with a high-status volunteer, F(1, 413) = 0.47, p = .492 (private condition: M = 3.07, SD = 1.07; public condition: Ó 2018 Hogrefe Publishing

M = 3.17, SD = 1.01). The main effect of publicity was not significant, p = .205. Indirect Effects The procedure of the indirect effects analyses was identical to the procedure used in Study 1 (see Figure 2 for the results). In line with the findings from Study 1, the results revealed that the status manipulation, b = 0.15, t(413) = 2.37, p = .018, pr2 = .01, 95% CI [0.00, 0.04], and the publicity manipulation, b = 0.38, t(413) = 6.22, p < .001, pr2 = .09, 95% CI [0.04, 0.14], were significantly and positively related to the attribution difference score, indicating that high-status volunteers (vs. low-status volunteers) and intended publicity (vs. intended privacy) both led to a stronger ascription of predominantly egoistic motives. The attribution difference score, in turn, was significantly negatively related to willingness to assist the volunteer, b = 0.29, t(412) = 7.47, p < .001, pr2 = .12, 95% CI [0.07, 0.18], indicating that the more participants’ ascribed egoistic (compared to altruistic) motives to the helper, the less their willingness to assist was. Importantly, the indirect effects of status and publicity were both significant, status: IE = .04, 95% CI [ 0.08, 0.01]; publicity: IE = .11, 95% CI [ 0.16, 0.07]. The Status Publicity interaction, in contrast, was not significantly related to the mediator, b = 0.05, t(413) = 0.78, p = .436, nor was its indirect effect significant, IE = 0.01, 95% CI [ 0.02, 0.05]. In sum, then, these findings clearly replicated the Study 1 results.

Discussion Study 2, using a different operationalization of the helper’s SES and a more comprehensive measure of participants’ attributions of altruistic versus egoistic motives, generally replicated the main findings from Study 1. Again, participants attributed the volunteer’s behavior more to egoistic motives (and less to altruistic motives) when he intended to make his help public (vs. keeping it private). Also, they attributed more egoistic motives to the helping act when the helper was of higher (vs. lower) SES. Moreover, similar to Study 1, we found evidence for additive rather than interactive effects of status and publicity on attributions. Mediation analyses also replicated the indirect effects of the status and publicity manipulations on participants’ willingness to assist the volunteer via increased attributions of predominantly egoistic motives. The only effect on ascribed motives that did not replicate across studies was the effect of the status manipulation on altruistic motives: Unlike Study 1, participants in the lowand high-status conditions made equally strong attributions to altruistic motives. We can only speculate about the reasons for this inconsistency. As stockbrokers’ working hours Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Figure 2. Path model of the indirect effects of status, publicity, and Status Publicity on willingness to assist via attribution difference score (higher values indicate greater predominance of egoistic motives) (Study 2). Coefficients are unstandardized regression weights. Total effects are reported in parentheses + p < .10; *p < .05; ***p < .001.

are known to be extremely long, it may well be that participants gave a stockbroker who devotes his very rare leisure time to volunteering extra credit for altruistic motives. However, with this exception, the two experiments created a clear and theoretically consistent pattern of the intended publicity and SES effects on ascribed motives.

General Discussion A large number of studies have examined how a person’s SES (e.g., Guinote et al., 2015; Piff et al., 2010) and the public versus private nature of a helping opportunity (e.g., Ariely, Bracha, & Meier, 2009; Kraus & Callaghan, 2016) affect his or her willingness to help. Comparably less is known about how these variables function as attributional cues guiding observers to infer helpers’ egoistic or altruistic motivation. Two experiments, in which we simultaneously manipulated the helper’s SES and the intended publicity of the helping act, produced a converging and robust pattern of effects. Observers attributed more egoistic (and less altruistic) motives to the helping act when helpers intended to make their helping public (vs. keeping it private). A similar pattern emerged when the helpers were of higher (vs. lower) SES. Of particular importance, the present research also provided solid evidence for the expected mediational role of observers’ attributions. Specifically, status and publicity indirectly affected observers’ willingness to assist the helper via motive attributions. A high-status (vs. low-status) helper and a helping act that was intended to be public (vs. private) Social Psychology (2019), 50(1), 53–66

both led to increased attributions of predominantly egoistic motives, which in turn were negatively related to observers’ willingness to join the helper’s cause. In conjunction with other research suggesting that people’s motivational attributions of helping behavior affect their willingness to assist the helper or join the helper’s cause (e.g., Plewa et al., 2015; Rodell & Lynch, 2016), these findings have immediate practical implications. High-status persons are important role models in the context of attracting new volunteers or donors (e.g., Wheeler, 2009). Also, going public with one’s voluntary work is often a necessary part of drawing attention to a charitable project. Still, as high SES and publicity both function as cues for “selfishness,” campaign initiators might take appropriate care to reduce the attribution of egoistic motives among their audiences. This can be done, for instance, by focusing on the project itself and its beneficiaries rather than on the high-status helper, by explicitly stating the helper’s altruistic reasons for going public, or by emphasizing that the media reports were not initiated by the high-status person him- or herself. Although the vignettes in our studies model real-world contexts, and the manipulations produced the theoretically expected effects, the design of our studies does not allow us to delineate precisely the “active ingredients” of our status manipulation. We acknowledge this as a major limitation of the present work. Our theoretical analyses suggest that observers ascribe egoistic motives to higher-SES helpers because they hold negative stereotypes (in terms of perceived lack of warmth) about them (e.g., Fiske et al., 2002; see also Conway et al., 1996; Koch et al., 2016). Still, as SES is highly correlated with wealth, the helper’s perceived wealth might provide an alternative explanation. Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Research suggests, for instance, that compared to less wealthy individuals, wealthier individuals have to donate larger amounts to charity or another person in order to be perceived as altruistic, probably simply because for those who have much, giving resources away makes less of a difference than for those who have only little (e.g., McCrink, Bloom, & Santos, 2010; Ng, Heyman, & Barner, 2011). Future research is needed to delineate and disentangle stereotype-based and wealth-based explanations. We wish to emphasize, however, that documenting the effects of status on the observers’ attributions (as we have achieved in the present studies) is an important contribution in and of itself as it helps to explain when people are punished rather than praised for their good deeds. A more precise understanding of the active ingredients of our manipulations might also help to interpret their effects on observers’ willingness to assist the volunteer. While the effects of our manipulations on ascribed motives, and also their indirect effects on willingness to assist via ascribed motives, were highly consistent across studies, the total effects pattern was less clear. Specifically, while we observed a publicity main effect on willingness to assist in Study 1, Study 2 revealed a status main effect qualified by an interaction effect. A possible explanation for these inconsistent findings could be that the helper’s status and the intended publicity of the helping act activate a variety of psychological processes, apart from the ascription of egoistic versus altruistic motives, that affect people’s willingness to assist in different ways. To illustrate, a helper’s status might have a positive effect on observers’ willingness join the helper’s cause (because a high-status helper is perceived as more efficacious) as well as a negative effect (because a high-status helper probably already has sufficient resources at hand and does not depend on additional assistance). This could result in rather inconsistent effects of status and publicity on willingness to assist the volunteer, while, at the same time, the indirect effects via the ascription of specific motives are quite stable. Future research might consider additional mediators over and above the motives observers’ ascribe to the helper, in order to more fully understand the effects of a helper’s status and the intended publicity of the helping act on observers’ responses. Before concluding, we wish to outline one other direction for future work. This concerns the role of group processes. Research on in-group bias shows that people tend to evaluate the behavior of in-group members in more positive terms than the behavior of out-group members (e.g., Brewer, 2007). Thus, one explanation for our observation that people tend to ascribe less altruistic (and more egoistic) motives to a higher-SES target could be that they perceived the higher-SES target as an out-group member and the lower-SES target as an in-group member. Both experiments Ó 2018 Hogrefe Publishing

included a measure of participants’ perceptions of their own subjective status (MacArthur Scale of Subjective Social Status, e.g., Operario, Adler, & Williams, 2004). In both experiments, this measure was unrelated to participants’ motive attributions (rs |.14|, ps .180, in each of the experimental conditions), thus rendering SES-based in-group/outgroup categorizations rather unlikely as an alternative explanation for our findings. Nevertheless, it seems an interesting task for future research to integrate the present research with existing work on the role of helpers’ and recipients’ group membership in helping relations (e.g., Hopkins et al., 2007; LotzSchmitt, Siem, & Stürmer, 2017; Nadler & Halabi, 2006; Siem & Stürmer, 2012; Stürmer, Snyder, & Omoto, 2005; van Leeuwen & Täuber, 2010). A particularly well-documented finding in this domain is that recipients of help respond less positively to help from members of high-status groups, especially when the assistance seeks to maintain the privileged social status of the helper’s group (e.g., Becker, Ksenofontov, Siem, & Love, 2018; Nadler & Halabi, 2006). The dominant explanation for these findings is that members of low-status groups perceive such kind of help provided by members of high-status groups as a threat to their social identity (e.g., Nadler, 2002). The current work suggests that recipients’ attribution of egoistic motives to the high-status helper’s assistance can meaningfully complement this explanation. This reasoning also relates to work on intergroup helping as a strategic image enhancement tool (e.g., Hopkins et al., 2007; van Leeuwen & Täuber, 2010). This line of research suggests that one reason for why people help out-group members is that they gain collective benefits related to the in-group’s image, such as preventing damages or threats to the in-group’s positive image or counteracting negative stereotypes about the ingroup. While the present research focused on cues guiding observers to infer a helper’s egoistic motives in the form of individual benefits, an interesting task for future research could be to investigate whether observers are equally susceptible to cues signaling collective benefits of helping.

Concluding Comment The present research suggests that, in real-world helping contexts, a helper’s high status and the publicity of the helping act can be a double-edged sword. On the one hand, a helper’s high status can be advantageous, because high-status individuals often serve as role models for others and simply have more resources than lower-status people. On the other hand, the present research indicates that their help is more likely to be attributed to egoistic motives, which, in turn, might keep others from joining the cause. Similarly, a certain amount of publicity is necessary to attract other potential volunteers to a charity campaign, Social Psychology (2019), 50(1), 53–66

but might also undermine the attribution of altruistic motives. We believe that awareness of these potential negative effects can help increase the effectiveness of public humanitarian campaigns initiated by high-status individuals, thus increasing the likelihood that more people join such causes and volunteer to help other people or communities.

Open Data/Materials All materials and data files are available on the Open Science Framework (https://osf.io/r3db6/).

Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1864-9335/a000360 ESM 1. Text (.docx) Auxiliary measures and additional analyses. ESM 2. Text (.docx) Vignettes.

References Adler, N. E., & Stewart, J. (2007). The MacArthur Scale of subjective social status. Retrieved from https://macses.ucsf. edu/research/psychosocial/subjective.php Apin, N. (2016, June 9). Kommentar Promis und Flüchtlinge: Das geteilte Haus am See [Comment on celebs and refugees: The shared house by the lake]. Taz. Retrieved from http://www.taz. de/!5308699/ Ariely, D., Bracha, A., & Meier, S. (2009). Doing good or doing well? Image motivation and monetary incentives in behaving prosocially. American Economic Review, 99, 544–555. https://doi.org/ 10.1257/aer.99.1.544 Batson, C. D. (1991). The altruism question: Toward a social psychological answer. Hillsdale, NJ: Erlbaum. Batson, C. D. (2011). Altruism in humans. New York, NY: Oxford University Press. Becker, J. C., Ksenofontov, I., Siem, B., & Love, A. (2018). Antecedents and consequences of autonomy- and dependency-oriented help toward refugees. European Journal of Social Psychology,. Advance online publication. https://doi.org/ 10.1002/ejsp.2554 Berman, J. Z., Levine, E. E., Barasch, A., & Small, D. A. (2015). The braggart’s dilemma: On the social rewards and penalties of advertising prosocial behavior. Journal of Marketing Research, 52, 90–104. https://doi.org/10.1509/jmr.14.0002 Brewer, M. B. (2007). The social psychology of intergroup relations: Social categorization, ingroup bias, and outgroup prejudice. In A. W. Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles (pp. 695–715). New York, NY: Guilford Press.

Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Carlo, G., & Randall, B. A. (2002). The development of a measure of prosocial behaviors for late adolescents. Journal of Youth and Adolescence, 31, 31–44. https://doi.org/10.1023/ A:1014033032440 Clary, E. G., Snyder, M., Ridge, R. D., Copeland, J., Stukas, A. A., Haugen, J., & Miene, P. (1998). Understanding and assessing the motivations of volunteers: A functional approach. Journal of Personality and Social Psychology, 74, 1516–1530. https://doi. org/10.1037/0022-3514.74.6.1516 Conway, M., Pizzamiglio, M. T., & Mount, L. (1996). Status, communality, and agency: Implications for stereotypes of gender and other groups. Journal of Personality and Social Psychology, 71, 25–38. https://doi.org/10.1037/00223514.71.1.25 Ellen, P. S., Webb, E. J., & Mohr, L. A. (2006). Building corporate associations: Consumer attributions for corporate socially responsible programs. Journal of the Academy of Marketing Science, 34, 147–157. https://doi.org/10.1177/ 0092070305284976 Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. https://doi.org/10.3758/BF03193146 Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878–902. https://doi. org/10.1037/0022-3514.82.6.878 Guinote, A., Cotzia, I., Sandhu, S., & Siwa, P. (2015). Social status modulates prosocial behavior and egalitarianism in preschool children and adults. Proceedings of the National Academy of Sciences of the United States of America, 112, 731–736. https://doi.org/10.1073/pnas.1414550112 Hahl, O., & Zuckermann, E. W. (2014). The denigration of heroes? How the status attainment process shapes attributions of considerateness and authenticity. American Journal of Sociology, 120, 504–554. https://doi.org/10.1086/678304 Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). New York, NY: Guilford Press. Heyman, G. D., Barner, D., Heumann, J., & Schenck, L. (2014). Children’s sensitivity to ulterior motives when evaluating prosocial behavior. Cognitive Science, 38, 683–700. https:// doi.org/10.1111/cogs.12089 Heyman, G. D., Fu, G., Barner, D., Zhishan, H., Zhou, L., & Lee, K. (2016). Children’s evaluation of public and private generosity and its relation to behavior: Evidence from China. Journal of Experimental Child Psychology, 150, 16–30. https://doi.org/ 10.1016/j.jecp.2016.05.001 Hopkins, N., Reicher, S., Harrison, K., Cassidy, C., Bull, R., & Levine, M. (2007). Helping to improve the group stereotype: On the strategic dimension of prosocial behavior. Personality and Social Psychology Bulletin, 33, 776–788. https://doi.org/ 10.1177/0146167207301023 Koch, A., Imhoff, R., Dotsch, R., Unkelbach, C., & Alves, H. (2016). The ABC of stereotypes about groups: Agency/socioeconomic success, conservative-progressive beliefs, and communion. Journal of Personality and Social Psychology, 110, 675–709. https://doi.org/10.1037/pspa0000046 Kraus, M. W., & Callaghan, B. (2016). Social class and prosocial behavior: The moderating role of public versus private contexts. Social Psychological and Personality Science, 7, 769–777. https://doi.org/10.1177/1948550616659120 Kraus, M. W., & Stephens, N. M. (2012). A road map for an emerging psychology of social class. Social and Personality

Ó 2018 Hogrefe Publishing

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

Psychology Compass, 6, 642–656. https://doi.org/10.1111/ j.1751-9004.2012.00453.x Lareau, A., & Conley, D. (2008). Social class: How does it work? New York, NY: Russell Sage Foundation. Lay, J. C., & Hoppmann, C. A. (2015). Altruism and prosocial behavior. In N. A. Pachana (Ed.), Encyclopedia of geropsychology (pp. 1–9). New York, NY: Springer. Lin-Healy, F., & Small, D. A. (2013). Nice guys finish last and guys in last are nice: The clash between doing well and doing good. Social Psychological and Personality Science, 4, 692–698. https://doi.org/10.1177/1948550613476308 Littler, J. (2008). “I feel your pain”: Cosmopolitan charity and the public fashioning of the celebrity soul. Social Semiotics, 18, 237–251. https://doi.org/10.1080/10350330802002416 Lotz-Schmitt, K., Siem, B., & Stürmer, S. (2017). Empathy as a motivator of dyadic helping across group boundaries: The disinhibiting effect of the recipient’s perceived benevolence. Group Processes & Intergroup Relations, 20, 233–259. https://doi.org/ 10.1177/1368430215612218 Mackey, A., Mackey, T. B., & Barney, J. B. (2007). Corporate social responsibility and firm performance: Investor preferences and corporate strategies. Academy of Management Review, 32, 817–835. https://doi.org/10.5465/AMR.2007. 25275676 McCrink, K., Bloom, P., & Santos, L. R. (2010). Children’s and adults’ judgments of equitable resource distributions. Developmental Science, 13, 37–45. https://doi.org/10.1111/j.14677687.2009.00859.x Nadler, A. (2002). Inter-group helping relations as power relations: Maintaining or challenging social dominance between groups through helping. Journal of Social Issues, 58, 487–502. https:// doi.org/10.1111/1540-4560.00272 Nadler, A., & Halabi, S. (2006). Intergroup helping as status relations: Effects of status stability, identification, and type of help on receptivity to high-status group’s help. Journal of Personality and Social Psychology, 91, 97–110. https://doi.org/ 10.1037/0022-3514.91.1.97 Newman, G. E., & Cain, D. M. (2014). Tainted altruism: When doing some good is evaluated worse than doing no good at all. Psychological Science, 25, 648–655. https://doi.org/10.1111/ j.1467-7687.2009.00859.x Ng, R., Heyman, G. D., & Barner, D. (2011). Collaboration promotes proportional reasoning about resource distribution in young children. Developmental Psychology, 47, 1230–1238. https:// doi.org/10.1037/a0024923 Oostlander, J., Güntert, S. T., van Schie, S., & Wehner, T. (2014). Volunteer Functions Inventory (VFI): Konstruktvalidität und psychometrische Eigenschaften der deutschen Adaption [Volunteer Functions Inventory (VFI): Construct validity and psychometric characteristics of the German adaptation]. Diagnostica, 60, 73–85. https://doi.org/10.1026/0012-1924/a000098 Operario, D., Adler, N. E., & Williams, D. R. (2004). Subjective social status: Reliability and predictive utility for global health. Psychology and Health, 19, 237–246. https://doi.org/10.1080/ 08870440310001638098 Penner, L. A., Dovidio, J. F., Piliavin, J. A., & Schroeder, D. A. (2005). Prosocial behavior: Multilevel perspectives. Annual Review of Psychology, 56, 365–392. https://doi.org/10.1146/ annurev.psych.56.091103.070141 Piff, P. K, Kraus, M. W., Côté, S., Cheng, B. H., & Keltner, D. (2010). Having less, giving more: The influence of social class on prosocial behavior. Journal of Personality and Social Psychology, 99, 771–784. https://doi.org/10.1037/a0020092 Plewa, C., Conduit, J., Quester, P. G., & Johnson, C. (2015). The impact of corporate volunteering on CSR image: A consumer

Ó 2018 Hogrefe Publishing

perspective. Journal of Business Ethics, 127, 643–659. https:// doi.org/10.1007/s10551-014-2066-2 Rodell, J. B., & Lynch, J. W. (2016). Perceptions of employee volunteering: Is it “credited” or “stigmatized” by colleagues? Academy of Management Journal, 59, 611–635. https://doi.org/ 10.5465/amj.2013.0566 Rucker, D. D., Preacher, K. J., Tormala, Z. L., & Petty, R. E. (2011). Mediation analysis in social psychology: Current practices and new recommendations. Social and Personality Compass, 5, 359–371. https://doi.org/10.1111/j.1751-9004.2011.00355.x Schlenker, B. R., Hallam, J. R., & McCown, N. E. (1983). Motives and social evaluation: Actor-observer differences in the delineation of motives for a beneficial act. Journal of Experimental Social Psychology, 19, 254–273. https://doi.org/10.1016/00221031(83)90041-0 Siem, B., & Stürmer, S. (2012). Meeting recipients’ needs in dyadic cross-group helping: When the nature of the helper’s motivation matters. Group Processes & Intergroup Relations, 15, 762– 775. https://doi.org/10.1177/1368430212444323 Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2012). A 21 word solution. Dialogue, 26, 4–7. Simmons, J. P., Nelson, L. D., & Simonsohn, U. (2013, January 18). Life after p-hacking. Meeting of the Society for Personality and Social Psychology, New Orleans, LA, 17–19 January 2013. Retrieved from http://ssrn.com/abstract=2205186 or https:// doi.org/10.2139/ssrn.2205186 Simon, B., Stürmer, S., & Steffens, K. (2000). Helping individuals or group members? The role of individual and collective identification in AIDS volunteerism. Personality and Social Psychology Bulletin, 26, 497–506. https://doi.org/10.1177/ 0146167200266008 Stürmer, S., Rohmann, A., & van der Noll, J. (2016). Mobilizing the global community to combat Ebola: Psychological effects of the Band Aid 30 campaign. The Journal of Social Psychology, 156, 291–304. https://doi.org/10.1080/00224545.2015.1108898 Stürmer, S., Snyder, M., & Omoto, A. M. (2005). Prosocial emotions and helping: The moderating role of group membership. Journal of Personality and Social Psychology, 88, 532–546. https://doi. org/10.1037/0022-3514.88.3.532 Sun, H. J. (2014). A study on the development of public campaign messages for organ donation promotion in Korea. Health Promototion International, 30, 903–918. https://doi.org/ 10.1093/heapro/dau023 Van de Vliert, E., Huang, X., & Levine, R. V. (2004). National wealth and thermal climate as predictors of motives for volunteer work. Journal of Cross Cultural Psychology, 35, 62–73. https:// doi.org/10.1177/0022022103260379 van Leeuwen, E., & Täuber, S. (2010). The strategic side of outgroup helping. In S. Stürmer & M. Snyder (Eds.), The psychology of prosocial behavior: Group processes, intergroup relations, and helping (pp. 81–99). Chichester, UK: WileyBlackwell. Walker, M., Heere, B., Parent, M., & Drane, D. (2010). Social responsibility and the Olympic Games: The mediating role of consumer attributions. Journal of Business Ethics, 95, 659– 680. https://doi.org/10.1007/s10551-010-0445-x Wheeler, R. T. (2009). Nonprofit advertising: Impact of celebrity connection, involvement and gender on source credibility and intention to volunteer time or donate money. Journal of Nonprofit & Public Sector Marketing, 21, 80–107. https://doi. org/10.1080/10495140802111984 Yoon, Y., Zeynep, G. C., & Schwarz, N. (2006). The effect of corporate social responsibility (CSR) activities on companies with bad reputations. Journal of Consumer Psychology, 16, 377–390. https://doi.org/10.1207/s15327663jcp1604_9

Social Psychology (2019), 50(1), 53–66

History Received November 28, 2017 Revision received July 24, 2018 Accepted July 24, 2018 Published online December 10, 2018 Acknowledgments Study 1 was conducted by Janna Schäfer and Study 2 by Christopher R. Gross as part of their master’s theses, both under the joint supervision of Stefan Stürmer and Birte Siem. Authorship Stefan Stürmer and Birte Siem designed the research. The data was collected as part of two master theses (Janna Schäfer collected the data of Study 1, Christopher R. Gross collected the data of Study 2. Birte Siem analyzed the data and wrote the first draft of the article. Stefan Stürmer contributed to later drafts, including the final draft of the article.

Social Psychology (2019), 50(1), 53–66

B. Siem & S. Stürmer, Attribution of Motives to Acts of Helping

ORCID Birte Siem https://orcid.org/0000-0003-4697-7856

Birte Siem Department of Psychology FernUniversität in Hagen Universitätsstraße 33 58084 Hagen Germany birte.siem@fernuni-hagen.de

Ó 2018 Hogrefe Publishing

Instructions to Authors Social Psychology is a publication dedicated to international research in social psychology as well as a forum for scientific discussion and debate. Social Psychology publishes innovative and methodologically sound research and serves as an international forum for scientific discussion and debate in the field of social psychology. Topics include all basic social psychological research themes, methodological advances in social psychology, as well as research in applied fields of social psychology. The journal focuses on original empirical contributions to social psychological research, but is open to theoretical articles, critical reviews, and replications of published research. The journal welcomes original empirical and theoretical contributions to basic research in social psychology, to social psychological methods, as well as contributions covering research in applied fields of social psychology, such as economics, marketing, politics, law, sports, the environment, the community, or health. Preference will be given to original empirical and experimental manuscripts, but theoretical contributions, critical reviews, and replications of published research are welcome as well. Social Psychology aims to increase transparency and openness of the research process and encourages authors to share their data and materials and if possible, pre-register their studies. Social Psychology publishes the following types of article: Original Articles, Research Reports, Replications. Manuscript submission: All manuscripts should in the first instance be submitted electronically at http://www.editorialmanager. com/sp. Detailed instructions to authors are provided at http:// www.hogrefe.com/j/sp Copyright Agreement: By submitting an article, the author confirms and guarantees on behalf of him-/herself and any coauthors that the manuscript has not been submitted or published elsewhere, and that he or she holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches, tables, and electronic supplementary material, and that the article and its contents do not infringe in any way on the rights of third parties. ESM will be published online as received from the author(s) without any conversion, testing, or reformatting. They will not be checked for typographical errors or functionality. The author

Ó 2019 Hogrefe Publishing

indemnifies and holds harmless the publisher from any third-party claims. The author agrees, upon acceptance of the article for publication, to transfer to the publisher the exclusive right to reproduce and distribute the article and its contents, both physically and in nonphysical, electronic, or other form, in the journal to which it has been submitted and in other independent publications, with no limitations on the number of copies or on the form or the extent of distribution. These rights are transferred for the duration of copyright as defined by international law. Furthermore, the author transfers to the publisher the following exclusive rights to the article and its contents: 1. The rights to produce advance copies, reprints, or offprints of the article, in full or in part, to undertake or allow translations into other languages, to distribute other forms or modified versions of the article, and to produce and distribute summaries or abstracts. 2. The rights to microfilm and microfiche editions or similar, to the use of the article and its contents in videotext, teletext, and similar systems, to recordings or reproduction using other media, digital or analog, including electronic, magnetic, and optical media, and in multimedia form, as well as for public broadcasting in radio, television, or other forms of broadcast. 3. The rights to store the article and its content in machinereadable or electronic form on all media (such as computer disks, compact disks, magnetic tape), to store the article and its contents in online databases belonging to the publisher or third parties for viewing or downloading by third parties, and to present or reproduce the article or its contents on visual display screens, monitors, and similar devices, either directly or via data transmission. 4. The rights to reproduce and distribute the article and its contents by all other means, including photomechanical and similar processes (such as photocopying or facsimile), and as part of so-called document delivery services. 5. The right to transfer any or all rights mentioned in this agreement, as well as rights retained by the relevant copyright clearing centers, including royalty rights to third parties. Online Rights for Journal Articles: Guidelines on authors’ rights to archive electronic versions of their manuscripts online are given in the ‘‘Guidelines on sharing and use of articles in Hogrefe journals’’ on the journal’s web page at http://www.hogrefe.com/j/sp September 2016

Social Psychology (2019), 50(1)

Start using strengths today! “The GO-TO book for building character.” Martin E. P. Seligman, the founder of positive psychology

Ryan M. Niemiec

Character Strengths Interventions A Field Guide for Practitioners 2018, xx + 300 pp. US $59.00 / € 46.95 ISBN 978-0-88937-492-8 Also available as eBook This book is the epitome of positive psychology: it takes the “backbone” of positive psychology – character strengths – and builds a substantive bridge between the science and practice. Working with clients’ (and our own) character strengths boosts well-being, fosters resilience, improves relationships, and creates strong, supportive cultures in our practices, classrooms, and organizations. This unique guide brings together the vast experience of the author with the science and the practice of positive psychology in such a way that both new and experienced practitioners will benefit. New practitioners will learn about the core concepts of character and signature strengths and how to fine-tune their approach and troubleshoot. Experienced practitioners will deepen their knowledge about advanced topics such as strengths overuse and collisions, hot button issues, morality, and integrating strengths with savoring,

www.hogrefe.com

flow, and mindfulness. Hands-on practitioner tips throughout the book provide valuable hints on how to take a truly strengths-based approach. The 24 summary sheets spotlighting each of the universal character strengths are an indispensable resource for client sessions, succinctly summarizing the core features of and research on each strength. 70 evidence-based step-by-step activity handouts can be given to clients to help them develop character strengths awareness and use, increase resilience, set and meet goals, develop positive relationships, and find meaning and engagement in their daily lives. No matter what kind of practitioner you are, this one-of-a-kind field guide is a goldmine in science-based applications. You’ll be able to immediately bring the science of well-being into action!

The best ways to support the healthy development of children and adolescents and their families Kristin S. Mathiesen / Ann V. Sanson / Evalill B. Karevold (Editors)

Tracking Opportunities and Problems from Infancy to Adulthood 20 Years With the TOPP Study 2018, x + 272 pp. US $49.80 / € 39.95 ISBN 978-0-88937-543-7 Also available as eBook The unique longitudinal study “Tracking Opportunities and Problems (TOPP)” began following nearly 1,000 children and their families in Norway in 1993. Few studies have ever accumulated such extensive information from such a large number of families. Eight waves of data on many aspects of child and family life have been collected from children aged 18 months to 18 years. The TOPP Study has provided new knowledge about and insight into the precursors, developmental paths and predictors of both good adaptation and mental health problems of children, as well as into parenting and family relationships. The editors have collated the key findings in three parts. Part 1 addressesthe mental health and development of children and adolescents. Part 2 focuses on parents, looking at individual

www.hogrefe.com

parental and family-related factors, including parental couple relationships. Part 3 looks at methodological issues, including the sample, response rate and measurement and analytical approaches. Each chapter reviews the existing knowledge in these areas in relation to the TOPP findings and provides extensive reference lists for those who want to dig deeper. This unique book provides thoughtprovoking insights into the TOPP findings to help guide therapeutic practice, to suggest new avenues of research, to inform teaching, and to shape policy planning and preventive actions. It is thus an invaluable resource for all professionals, researchers, educators, policy makers, and students working with children and adolescents and their families.

Assessment and treatment of Internet addiction “This excellent book is a pleasure to read. At a time when clinicians are scrambling to learn what they can about the rapidly developing problem of Internet addiction, this book offers them an excellent place to start.” Hilarie Cash, PhD, Chief Clinical Officer and Co-Founder of reSTART Life, PLLC, Fall City, WA – the first residential treatment program for Internet addiction in the US

Daria J. Kuss / Halley M. Pontes

Internet Addiction (Series: Advances in Psychotherapy – Evidence-Based Practice – Volume 41) 2019, iv + 86 pp. US $29.80 / € 24.95 ISBN 978-0-88937-501-7 Also available as eBook This book examines how you can identify, assess, and treat Internet addiction in the most effective manner. Internet use has become an integral part of our daily lives, but at what point does it become problematic? What are the different kinds of Internet addiction? And how can professionals best help clients? This compact, evidence-based guide written by leading experts from the field helps disentangle the debates and controversies around Internet addiction, including social media addiction

www.hogrefe.com

and Internet gaming disorder, and outlines the current assessment and treatment methods. The book presents a 12–15 session treatment plan for Internet and gaming addiction using the method and setting with the best evidence: group CBT. Printable tools in the appendix help clinicians implement therapy. This accessible book is essential reading for clinical psychologists, psychiatrists, psychotherapists, counsellors, social workers, teachers, as well as students.