Experimental Psychology 1/2018 by Hogrefe

Volume 65 / Number 1 / 2018

Experimental Psychology

Editor-in-Chief Christoph Stahl Editors Tom Beckers Arndt Bröder Adele Diederich Chris Donkin Gesine Dreisbach Andreas Eder Magda Osman Manuel Perea James Schmidt Samuel Shaki Sarah Teige-Mocigemba

Hogrefe OpenMind Open Access Publishing? It’s Your Choice! Your Road to Open Access Authors of papers accepted for publication in any Hogrefe journal can now choose to have their paper published as an open access article as part of the Hogrefe OpenMind program. This means that anyone, anywhere in the world will – without charge – be able to read, search, link, send, and use the article for noncommercial purposes, in accordance with the internationally recognized Creative Commons licensing standards.

The Choice Is Yours 1. Open Access Publication: The final “version of record” of the article is published online with full open access. It is freely available online to anyone in electronic form. (It will also be published in the print version of the journal.) 2. Traditional Publishing Model: Your article is published in the traditional manner, available worldwide to journal subscribers online and in print and to anyone by “pay per view.” Whichever you choose, your article will be peer-reviewed, professionally produced, and published both in print and in electronic versions of the journal. Every article will be given a DOI and registered with CrossRef.

www.hogrefe.com

How Does Hogrefe’s Open Access Program Work? After submission to the journal, your article will undergo exactly the same steps, no matter which publishing option you choose: peer-review, copy-editing, typesetting, data preparation, online reference linking, printing, hosting, and archiving. In the traditional publishing model, the publication process (including all the services that ensure the scientific and formal quality of your paper) is financed via subscriptions to the journal. Open access publication, by contrast, is financed by means of a one-time article fee (€ 2,500 or US $3,000) payable by you the author, or by your research institute or funding body. Once the article has been accepted for publication, it’s your choice – open access publication or the traditional model. We have an open mind!

Experimental Psychology

Volume 65/Number 1/2018

Editors

C. Stahl (Editor-in-Chief), Köln, Germany T. Beckers, Leuven, Belgium A. Bröder, Mannheim, Germany A. Diederich, Bremen, Germany C. Donkin, Sydney, Australia G. Dreisbach, Regensburg, Germany

A. Eder, Würzburg, Germany M. Osman, London, UK M. Perea, Valencia, Spain J. Schmidt, Ghent, Belgium S. Shaki, Samaria, Israel S. Teige-Mocigemba, Marburg, Germany

Editorial Board

U. J. Bayen, Düsseldorf, Germany H. Blank, Portsmouth, UK J. De Houwer, Ghent, Belgium R. Dell’Acqua, Padova, Italy G. O. Einstein, Greenville, SC, USA E. Erdfelder, Mannheim, Germany M. Goldsmith, Haifa, Israel D. Hermans, Leuven, Belgium R. Hertwig, Berlin, Germany J. L. Hicks, Baton Rouge, LA, USA P. Juslin, Uppsala, Sweden Y. Kareev, Jerusalem, Israel D. Kerzel, Geneva, Switzerland A. Kiesel, Freiburg, Germany K. C. Klauer, Freiburg, Germany R. Kliegl, Potsdam, Germany I. Koch, Aachen, Germany J. I. Krueger, Providence, RI, USA S. Lindsay, Victoria, BC, Canada

E. Loftus, Irvine, CA, USA T. Meiser, Mannheim, Germany K. Mitchell, West Chester, PA, USA N. W. Mulligan, Chapel Hill, NC, USA B. Newell, Sydney, Australia K. Oberauer, Zürich, Switzerland F. Parmentier, Palma, Spain M. Regenwetter, Champaign, IL, USA R. Reisenzein, Greifswald, Germany J. N. Rouder, Columbia, MO, USA D. Shanks, London, UK M. Steffens, Landau, Germany S. Tremblay, Quebec, Canada C. Unkelbach, Köln, Germany M. Waldmann, Göttingen, Germany E. Walther, Trier, Germany P. A. White, Cardiff, UK D. Zakay, Tel Aviv, Israel

Publisher

Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-425, E-mail publishing@hogrefe.com, Web http://www.hogrefe.com

Production

Regina Pinks-Freybott, Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-425, E-mail production@hogrefe.com

Subscriptions

Hogrefe Publishing, Herbert-Quandt-Str. 4, D-37081 Göttingen, Germany, Tel. +49 551 99950-956, Fax +49 551 99950-998

Advertising/Inserts

Melanie Beck, Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, Tel. +49 551 99950-0, Fax +49 551 99950-425, E-mail marketing@hogrefe.com

ISSN

ISSN-L 1618-3169, ISSN-Print 1618-3169, ISSN-Online 2190-5142

Ó 2018 Hogrefe Publishing. The journal as well as the individual contributions to 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, digital, mechanical, photocopying, microﬁlming or otherwise, without prior written permission from the publisher. All rights, including translation rights, are reserved.

Publication

Published in six issues per annual volume. Experimental Psychology is the continuation of Zeitschrift für Experimentelle Psychologie (ISSN 0949-3964), the last annual volume of which (Volume 48) was published in 2001.

Subscription Prices

Annual subscription (2018): Individuals: US $259.00/1185.00; Institutions: US $508.00/1398.00 (postage & handling: US $24.00/118.00). Single issue: US $85.00/166.50 (+ postage & handling).

Payment

Payment may be made by check, international money order, or credit card to Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany, or, for customers in North America, to Hogrefe Publishing, Inc., Journals Department, 7 Bulﬁnch Place, 2nd ﬂoor, Boston, MA 02114, USA.

Electronic Full Text

The full text of Experimental Psychology is available online at http://econtent.hogrefe.com/loi/zea

Abstracting/Indexing Services

Experimental Psychology is abstracted/indexed in Current Contents/Social & Behavioral Sciences (CC/S&BS), Social Science Citation Index (SSCI), Medline, PsyJOURNALS, PsycINFO, PSYNDEX, ERIH, Scopus, and EMCare. Impact Factor (2016): 1.829.

Experimental Psychology (2018), 65(1)

Ó 2018 Hogrefe Publishing

Contents Research Article

Adults’ and Children’s Understanding of How Expertise Inﬂuences Learning Judith H. Danovitch and Christine K. Shenouda

Short Research Articles

Interference in Dutch-French Bilinguals: Stimulus and Response Conﬂict in Intra- and Interlingual Stroop James R. Schmidt, Robert J. Hartsuiker, and Jan De Houwer

A Bird in the Hand Isn’t Good for Long: Action Dynamics Reveal Short-Term Choice Impulses in Intertemporal Choices Stefan Scherbaum, Simon Frisch, and Maja Dshemuchadse

Integrating Orthographic Information Across Time and Space: Masked Priming and Flanker Effects With Orthographic Neighbors Joshua Snell, Daisy Bertrand, Martijn Meeter, and Jonathan Grainger

Shared Processing of Language and Music: Evidence From a Cross-Modal Interference Paradigm Ryan P. Atherton, Quin M. Chrobak, Frances H. Rauscher, Aaron T. Karst, Matt D. Hanson, Steven W. Steinert, and Kyra L. Bowe

Contagion via Magical Thinking and via Mere Proximity: Differences as a Function of Target Type Lennea R. Bower, Zehra F. Peynircioğlu, and Brian E. Rabinovitz

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1)

Research Article

Adults’ and Children’s Understanding of How Expertise Influences Learning Judith H. Danovitch1 and Christine K. Shenouda2 1

Department of Psychological and Brain Sciences, University of Louisville, KY, USA

Department of Psychology, University of Illinois, Champaign, IL, USA

Abstract: Adults and children use information about expertise to infer what a person is likely to know, but it is unclear whether they realize that expertise also has implications for learning. We explore adults’ and children’s understanding that expertise in a particular category supports learning about a closely related category. In four experiments, 5-year-olds and adults (n = 160) judged which of two people would be better at learning about a new category. When faced with an expert and a nonexpert, adults consistently indicated that expertise supports learning in a closely related category; however, children’s judgments were inconsistent and were strongly influenced by the description of the nonexpert. The results suggest that although children understand what it means to be an expert, they may judge an individual’s learning capacity based on different considerations than adults. Keywords: expertise, knowledge, learning

In 2009, as the American auto industry faltered, the United States government created a 50 million dollar program to retrain unemployed auto workers for jobs in the renewable energy sector, such as installing solar panels or servicing wind turbines (Associated Press, 2009). In doing so, policy makers assumed that the auto workers’ expertise in auto manufacturing would enable them to learn relatively quickly and easily about constructing and maintaining other types of complex machines. This assumption is supported by our tacit understanding of expertise. By age 5, children realize that individuals who know about a certain category (e.g., eagles) are likely to also know about closely taxonomically related categories (e.g., other types of birds) and even about more distant categories (e.g., other animals or plants) that belong to the same domain and share the same underlying causal principles (Keil, Stein, Webb, Billings, & Rozenblit, 2008; Lutz & Keil, 2002). Furthermore, as suggested by the example of the auto worker training program, adults appear to intuit that knowledge about a particular topic supports learning novel information in a closely related domain. Thus, an auto worker should find it easier to learn about solar panels and wind turbines than an artist or a zookeeper. The idea that expertise supports learning has substantial research support. Compared to novices, experts are better at perceiving the underlying causal structure of a problem or situation and they can retrieve domain-relevant Ó 2018 Hogrefe Publishing

knowledge more quickly and easily (see Chi, 2006 for a review). Expertise is not only useful for retrieving familiar information; experts can also draw more accurate conclusions about novel exemplars within the same domain. Moreover, these benefits of expertise are apparent even among 7-year-old experts (Gobbo & Chi, 1986). Expertise also increases an individual’s motivation to explore related topics (Silvia, 2008). Therefore, expertise may be viewed as beneficial for learning not only in terms of interpreting and organizing new information, but also in terms of promoting an individual’s desire to learn. By the time they complete preschool, children have already developed an appreciation of others’ expertise and the division of cognitive labor, or the idea that different people know different things. Young children can draw accurate inferences about the range of an individual’s knowledge and use others’ expertise to guide their information-seeking behaviors (Aguiar, Stoess, & Taylor, 2012; Lutz & Keil, 2002; but see Landrum, Mills, & Johnston, 2013 for evidence that they sometimes prioritize other characteristics, such as benevolence). Children also rely on their understanding of expertise to guide decisions about epistemic trust. For example, 3- and 4-year-olds trust a dog expert to correctly label dogs, but they do not believe that a dog expert is more accurate at labeling artifacts than a nondog-expert (Koenig & Jaswal, 2011). However, it is unclear whether children apply their understanding of expertise to Experimental Psychology (2018), 65(1), 1–12 https://doi.org/10.1027/1618-3169/a000387

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

infer that knowledge about one topic can be helpful for learning about another topic. Are children aware that expertise is beneficial for learning new information in the same domain, but not in other domains? Some insights into this question can be gained from research examining children’s understanding of the learning process in general. Children as young as age 4 realize that how well an individual learns novel information depends on the individual’s attention to the information and intention to learn (Sobel, Li, & Corriveau, 2007). By age 6, children remember where they learned new information (Drummey & Newcombe, 2002) and, importantly, they acknowledge that they did not know that information all along (Taylor, Esbensen, & Bennett, 1994). These findings suggest that even before they have much experience with formal education, children realize that knowledge has to be acquired from external sources, and that multiple factors determine how well an individual learns new information. Nevertheless, children may not necessarily view expertise as influencing learning, or they may view other factors, such as attention or memory, as more important for learning than existing knowledge. The current study examines whether children believe that one’s existing knowledge and expertise influence learning about related or unrelated categories. One possibility is that children reason about knowledge acquisition in the same way that they reason about existing knowledge, and therefore they expect someone who knows about a topic to be better at learning about a related topic. For instance, if a child can infer that an eagle expert knows about chickens and dogs, but not about trains and washing machines (Lutz & Keil, 2002), then that child may apply the same reasoning to infer that the eagle expert’s knowledge about eagles (which reflects underlying knowledge about biology) will support the acquisition and retention of information about other animals. Children may also attribute superior learning skills to experts over nonexperts based on a belief that existing expertise and learning ability are both functions of superior intelligence. If this is the case, then children should view experts as more capable of learning novel information than nonexperts, regardless of the experts’ domain of expertise. Although this seems plausible, the fact that young children do not demonstrate a halo effect for expertise (see Koenig & Jaswal, 2011) suggests that it is unlikely children maintain such a domain-general view of the value of expertise for learning. A third possibility is that children view existing knowledge and expertise as detrimental for learning new information. By age 5, children acknowledge that the human

mind has limitations and that typical human beings cannot know everything (Danovitch & Keil, 2008; Lane, Wellman, & Evans, 2010). Thus, children may consider an individual’s capacity for acquiring and storing information to be finite and subsequently reason that experts have less memory capacity or “space” available in their minds for new information. If this is the case, children should indicate that people who have less knowledge have an advantage over people who have more knowledge when it comes to learning and retaining information, regardless of the domain. Similarly, children may view existing knowledge and expertise as potentially interfering with learning. For example, they might believe that an expert would be more likely to confuse information he or she already knows with new information – an intuition that has support in research demonstrating that experts are more prone to falsely recalling information that is related to their domain of expertise (e.g., Castel, McCabe, Roediger, & Heitman, 2007). The experiments described here explore how adults and children attribute learning relative to an individuals’ existing knowledge and expertise, or lack thereof. As an initial foray into this question, these experiments focus on a type of learning that appears to invoke the most consistent intuition among adults: an expert learning about a topic that is closely related to his or her area of expertise.1 In addition, because 5-year-old children are capable of categorizing objects taxonomically and they understand what it means to be an expert, scenarios that involve an expert learning about a closely related category represent a critical test of whether children apply their understanding of categorical relationships and the division of cognitive labor to reasoning about learning outcomes. The current studies focus specifically on dog expertise, a type of expertise that has been used in prior research and that is familiar to young children (e.g., Koenig & Jaswal, 2011). Participants were presented with scenarios where a dog expert and a non-dog-expert studied a category closely taxonomically related to dogs: cats. Participants then indicated which of the two individuals they believed would be more successful in learning about the new category. The description of the non-dog-expert varied across experiments in order to examine how the contrast between expertise in a related category and either the absence of expertise or expertise in an unrelated category influenced participants’ judgments about learning. The experiments presented here explore adults’ and 5-year-old children’s intuitions about the relationship between learning and expertise. We chose 5-year-olds for a number of reasons. First, by age 5, children

In a preliminary study, we found that neither adults (n = 15) nor children aged 4 and 5 (n = 20) had consistent intuitions about an expert’s ability to learn about an unrelated category.

Experimental Psychology (2018), 65(1), 1–12

Ó 2018 Hogrefe Publishing

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

have acquired a basic theory of mind (see Wellman, Cross, & Watson, 2001) and they understand that certain individuals know about familiar topics, while others are ignorant (see Harris, 2012 for a review). Second, 5-year-olds are adept at classifying both familiar and novel objects according to superordinate level categories (e.g., Mervis & Crisafi, 1982; Tager-Flusberg, 1985) and they can use category information to make inductive judgments (see Gelman, 2003). Cate-gorization also appears to be important for understanding the division of cognitive labor (Danovitch, 2013; Danovitch & Noles, 2014) and, thus, it may be critical for linking knowledge about one category with learning about a related category. Third, children in this age group still have limited experience with formal education and the structured study of different domains. Therefore, their beliefs about learning and expertise should reflect their intuitive understanding of how expertise contributes to learning outcomes.

Experiment 1 Experiment 1 examined whether adults and 5-year-old children draw a connection between what a person already knows and their capacity to learn about a closely related category. The experiment presented participants with a scenario where two individuals, a dog expert and a nonexpert who did not know about dogs, both learned about cats. Based on previous research (Koenig & Jaswal, 2011), we expected children to understand what it means to be a dog expert, but it was unclear whether children would extend the dog expert’s knowledge to learning about cats.

Methods Participants Twenty 5-year-old children participated (Mage = 5.61 years, SD = 0.34, range = 5.01–5.99 years; 9 females). Children were recruited from local preschools and a laboratory database in two midsized Midwestern cities. Parents identified one child as Hispanic, and the rest as non-Hispanic. Parents identified 15 of the children as Caucasian-American, 1 child as Asian-American, 1 child as African-American, and 2 children as belonging to more than one ethnic group. Parents of one additional child declined to identify the child’s race or ethnicity. Children were tested individually in a quiet area at their school or in a university laboratory in audio-recorded sessions lasting approximately 10 min. Twenty-three adults (Mage = 20.61 years, SD = 4.03, range = 18–38 years; 9 females) from a large Midwestern university participated for course credit and were tested using the Ó 2018 Hogrefe Publishing

same procedure as the children in individual sessions lasting approximately 10 min. Design, Materials, and Procedure The experimenter presented each participant with photographs of two similar male puppets wearing differentcolored shirts, and introduced them as two of her friends who were adults like the participant’s parents. The experimenter pointed to one of the individuals, placed a picture of a dog next to him, and described the expert (using language based on the expert description in Lutz & Keil, 2002): “This person is a dog expert. He knows all about dogs. He is always around dogs and his job is to work with dogs. He knows all about what kinds of food dogs eat, how many babies they have, and how big they can grow. He knows more about dogs than anyone I know, including this person.” The experimenter then pointed at the second individual and said: “This person is not a dog expert. He doesn’t know much about dogs.” To ensure that participants understood the descriptions, the experimenter removed the dog picture and asked who knew more about dogs and to whom the child would direct a question about dogs. Participants were then told that both individuals were going to have a chance to learn about cats and that both would receive the same book about cats. The experimenter placed two identical miniature books titled “Cats” with a photograph of a cat on the cover and text on the interior pages on each puppet photograph at the same time and stated that each person was told to “read the book and learn all they can about cats. Both of them spend a long time reading all about cats.” The experimenter removed the books and asked an overall knowledge question: “Now, who do you think knows more about cats?” Participants were prompted to explain their answer. Following the explanation, participants completed three questions intended to assess different manifestations of learning: (1) “Who will remember what they read about cats better?”; (2) “Who will do better on a test about cats?”; and (3) “There are many different kinds of cats in the world, and sometimes we see different cats when we go outside. Both of these people are walking down the street one day and they see a cat. Who do you think will know what kind of cat it is?” The identity of the dog expert (i.e., puppet appearance) and the order in which the characters were introduced were counterbalanced between subjects. Experimental Psychology (2018), 65(1), 1–12

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

Explanation Coding Adults’ and children’s answers to the explanation prompt for the overall knowledge question were transcribed from audio recordings and coded into one of five categories. Explanations were categorized as “reference to expertise” when they cited the dog expert’s area of expertise (e.g., “because he’s a dog expert”). This classification was used regardless of whether the child had indicated that the dog expert or the nonexpert learned more about cats. Explanations were classified as “equitable distribution” when children explicitly referred to the distribution of knowledge between the characters, such that if one knows about dogs, then the other should know about cats (“e.g., “[The dog expert] knows a lot about dogs so this guy must know about cats”).2 “Topic similarity” responses referred to the similarity between dogs and cats (e.g., “Because if he knows about dogs, cats are an animal - both are animals”). Explanations were coded as a “post hoc justification” when they cited experience or knowledge that had not been included in the description of the character or events (e.g., “because he’s been seeing a lot of cats,” “because he read more”) or “irrelevant characteristics” when they cited physical features of the character (e.g., “because he looks like he is a little bigger”). There was also an additional category for children who said “I don’t know” or refused to generate an explanation. All explanations were coded by two coders, who were blind to each participant’s other responses (including their response to the overall knowledge question). Inter-rater agreement for coding was very good, Cohen’s κ = .930. Disagreements were resolved via discussion. Explanations were not available for four child participants in Experiment 1 due to technical difficulties.

Results and Discussion Following the introduction of the characters, all participants indicated that the dog expert knew more about dogs and all participants, except for one child, indicated that they would ask the dog expert a question about dogs. This child was excluded from further analyses. For the overall knowledge question, 17 out of 23 adults (74%) chose the dog expert, yet only 2 out of 19 children (11%) did so (see Figure 1). Chi-squared tests indicated that both of these proportions were significantly different from chance (w2 5.26, ps .022) and a Mann-Whitney U-test showed that they were also significantly different from each other, U = 83.00, p < .001. Thus, adults and children exhibited divergent judgments, with the majority of adults indicating that the dog expert would learn more about cats, but the

Figure 1. Children’s and adults’ responses to the overall knowledge question in Experiments 1–4. Error bars indicate the standard error of the mean (SEM). Except for children in Experiments 2 and 4, and adults in Experiment 4, all proportions are significantly different from chance at the p < .05 level.

majority of children indicating that the nonexpert would do so. Initial Cochran’s Q tests showed no significant differences in responses for the three manifestation of learning questions among adults or children, Qs 2.333, ps .311 (see Electronic Supplementary Material, ESM 1). Thus, responses were combined so that participants were assigned 1 point each time they chose the dog expert for each manifestation of learning question, yielding a total score of 0–3. For these items, adults chose the dog expert significantly more often than chance (chance = 1.5), M = 2.48, SD = 1.16, t(22) = 4.04, p = .001. Conversely, children chose the dog expert significantly less often than chance, M = 0.80, SD = 1.11, t(19) = 2.83, p = .011. These responses suggest that adults’ and children’s intuitions about expressions of learning mirrored their intuitions about the retention and application of newly learned information. In explaining their choice for the overall knowledge question, adults who chose the dog expert consistently cited his knowledge about dogs and often made explicit statements about the similarity between cats and dogs (see Table 1 and ESM 2). Likewise, the child who chose the dog expert for the learning question explained his choice by referring to the character’s expertise about dogs. Among the adults and children who chose the nonexpert, the majority explained their choice by focusing on the dog expert’s knowledge, which they appeared to view as detrimental to learning, or they made explicit reference to the equitable distribution of knowledge among the characters (e.g., “[the dog expert] knows a lot about dogs so [the nonexpert] must know about cats”).

Although both dogs and cats are common household pets among the participant population, no participant made references to parents’ or other familiar individuals’ expertise in their explanations.

Experimental Psychology (2018), 65(1), 1–12

Ó 2018 Hogrefe Publishing

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

Table 1. Number of times adults and children cited each type of justification for the overall knowledge question by Experiment and character chosen Experiment 1 Category

Experiment 2

Experiment 3

Dog Expert

Nonexpert

Dog Expert

Nonexpert

Dog Expert

Car Expert

Reference to expertise

Equitable distribution

Adults

Topic similarity Post hoc justification

Reference to expertise

Equitable distribution

Topic similarity

Post hoc justification

Irrelevant characteristics

Total Children

No response

Total

Note. No adult participants cited irrelevant characteristics or did not respond to the question.

The results of Experiment 1 demonstrate that when adults are presented with an expert and a nonexpert, they believe that expertise supports learning about a closely related category. However, children had the opposite intuition: they consistently indicated that the nonexpert would be better than the expert at learning about the new topic. Children’s explanations often focused on the contrast between the experts, suggesting that they may have viewed the expert’s existing knowledge as a liability or the nonexpert’s lack of knowledge about dogs as beneficial for learning. Children may have believed that knowledge about dogs could potentially interfere with learning about cats, or that possessing less knowledge meant that more storage space was available for new information.

Experiment 2 One way of testing whether children’s responses in Experiment 1 were driven by a view of existing knowledge as interfering with learning or of the nonexpert as having more room to store new information is to present the nonexpert as having some, rather than no, knowledge about dogs. If children believe that knowledge about dogs interferes with learning about cats, then presenting both experts as knowing something about dogs might cause children to focus on other characteristics and to potentially shift their attributions of learning in favor of the expert. However, if children focused primarily on the nonexpert as having a greater Ó 2018 Hogrefe Publishing

learning capacity, then presenting him as knowing a little about dogs still leaves him with more storage capacity than the dog expert, in which case children should continue to view him as more capable of learning.

Methods Participants Nineteen 5-year-old children participated (Mage = 5.42 years, SDage = 0.33, range = 5.03–5.95 years; 11 females). Children were recruited from the same communities as in Experiment 1. None of the children were identified by their parents as Hispanic. Parents identified 13 of the children as Caucasian-American, three children as AfricanAmerican, and two children as Asian-American. Parents of one additional child declined to identify the child’s race or ethnicity. Children were tested individually in sessions lasting approximately 10 min. Twenty adults (Mage = 20.63 years, SDage = 4.85, range = 18–40 years, 17 females) from a large university participated for course credit and were tested in individual sessions using the exact same procedure as the children. No child or adult had participated in Experiment 1. Audio recordings were not available for two children due to technical errors. Design, Materials, and Procedure Experiment 2 employed the same design, materials, and procedure as Experiment 1, with two exceptions. First, when introducing the dog expert, the experimenter put a photo-quality picture of 14 dogs of different breeds Experimental Psychology (2018), 65(1), 1–12

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

and sizes next to the dog expert. Second, the experimenter placed a picture of three different dogs next to the nonexpert and described him as follows: “This person knows a little bit about dogs. He has been around a few dogs. He knows a few things about what dogs do.”

Results and Discussion Similar to Experiment 1, all participants indicated that the dog expert knew more about dogs, and all but one child indicated that they would consult the dog expert for a question about dogs. This child was excluded from further analyses. Among adults, 17 out of 20 adults (85%) attributed greater overall knowledge to the dog expert (w2 = 9.800, p = .002; see Figure 1). Children’s choices, on the other hand, were no different from chance (w2 = 0.889, p = .346), with 7 out of 17 children (41%) choosing the dog expert (see Figure 1). A Mann-Whitney U-test showed that adults chose the dog expert significantly more often than children, U = 97.000, p = .004. As in Experiment 1, Cochran’s Q tests showed no significant differences in responses for the manifestation of learning items among either age group, Qs 4.200, ps .122. For these items, adults chose the dog expert at rates significantly different from chance, M = 2.65, SD = 0.93, t(19) = 4.04, p < .001, but children’s responses were not significantly different from chance, M = 1.33, SD = 1.08, t(17) = 0.652, p = .523 (see ESM 1). The distribution of adults’ explanations for the overall knowledge question was very similar to Experiment 1, with most adults citing topic similarity to support their choice of the dog expert (see Table 1 and ESM 2). However, children generated post hoc justifications and cited irrelevant characteristics more often than in Experiment 1. Notably, only one of the seven children who chose the dog expert cited topic similarity, suggesting that even when children made the same judgment about learning as adults, their reasoning did not necessarily follow the same course. The results of Experiment 2 suggest that although adults continued to indicate that the dog expert would demonstrate superior learning compared to the nonexpert, children were equally likely to indicate that the expert or the nonexpert had learned more about the new category overall. In addition, children’s responses to the manifestation of learning questions suggest that children were uncertain which character would demonstrate more knowledge about the novel topic. Even when children indicated that the dog expert would be more successful at learning about cats, very few children explained their reasoning in terms of topic similarity or expertise. That said, the fact that children did not consistently attribute greater learning to the nonexpert suggests that children’s judgments were not based solely on inferences about remaining memory capacity. Experimental Psychology (2018), 65(1), 1–12

The results of Experiment 2 still leave open the possibility that children view existing knowledge – regardless of its quantity – as problematic for acquiring new information in a related category. Experiment 3 further explores this possibility by examining whether children believe that individuals with expertise in a closely related category have an advantage or a disadvantage for learning relative to individuals with expertise in an unrelated category.

Experiment 3 In Experiments 1 and 2, when comparing experts to nonexperts, adults appeared to view existing knowledge about a related topic as beneficial to learning while children appeared to view it as detrimental or irrelevant. Experiment 2 revealed that children’s attribution of a learning advantage to the nonexpert was not related solely to differences in storage capacity. Another potential mechanism underlying children’s response patterns is that children believe knowledge about a category actively interferes with learning about a closely related category. In order to evaluate this possibility, Experiment 3 presented participants with two experts in different categories: a dog expert and a car expert. Because both experts were described as masters of a broad basic-level category, they could be construed as having roughly equivalent amounts of existing knowledge. Thus, this design provides a measure of children’s intuitions about how prior knowledge influences new learning when factors such as memory capacity are removed from the equation. If children can apply their understanding of the division of cognitive labor (e.g., Lutz & Keil, 2002) and knowledge about category membership and similarity to reasoning about learning capacity, then 5-year-olds should indicate that a dog expert would be better at learning about cats than a car expert. However, if children consider expertise in any domain as interfering with learning, then they should choose between the experts at random. Finally, if children view expertise as interfering with learning about a closely related category, then they should attribute superior learning about cats to the car expert.

Methods Participants Twenty-one 5-year-old children participated (Mage = 5.49 years, SD = 0.32, range = 5.02–5.98 years, 9 females). Children were recruited from two midsized Midwestern cities through local preschools and a laboratory database. None of the children were identified as Hispanic. Parents identified 17 of the children as Caucasian-American, 1 child as Asian-American, and 2 children as belonging to more Ó 2018 Hogrefe Publishing

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

than one ethnic group. Parents of one additional child declined to identify the child’s race or ethnicity. Children were tested individually in sessions lasting approximately 10 min. Twenty university students (Mage = 19.5 years, SDage = 1.00, range = 18–21 years, 19 females) participated for course credit and were tested in individual sessions using the same procedure as the children. None of the participants had completed any of the previous experiments. Audio recordings were not available for two children due to technical errors. Design, Materials, and Procedure Experiment 3 employed the same design, materials, and procedure as Experiment 1, except that, when introducing the second individual, the experimenter placed a picture of a car next to the character’s picture and said: “This person is a car expert. He is always around cars and his job is to work with cars. He knows all about how cars work, what kinds of gas they need, and how many companies make cars. He knows more about cars than anyone I know, including this person [pointing to dog expert].” Following the introduction of the experts, children were asked to identify each expert and to indicate who they would consult for a question about dogs and a question about cars.

Results and Discussion Following the introduction of both experts, all participants correctly identified each character’s domain of expertise. One child chose the incorrect character to consult for a question about dogs and a question about cars. This child was excluded from further analyses. For the overall knowledge question, 19 out of 20 adults (95%) and 15 out of 20 children (75%) chose the dog expert (see Figure 1). Chi-squared tests indicated that both of these proportions were significantly different from chance (w2 5.00, ps .025) and a Mann-Whitney U-test showed that they were not significantly different from each other, U = 160.000, p = .08. Based on Cochran’s Q tests indicating no significant difference in responses between items for either age group, Qs 3.500, ps .174, scoring for the manifestation of learning questions followed the same procedure as the prior experiments. Adults chose the dog expert at rates significantly above chance, M = 2.75, SD = 0.55, t(19) = 10.162, p < .001, but children were no different from chance, M = 1.80, SD = 0.95, t(19) = 2.161, p = .175 (see ESM 1).

Ó 2018 Hogrefe Publishing

As in the prior experiments, adults consistently justified their choice of the dog expert in terms of topic similarity (see Table 1 and ESM 2). In contrast, even though the majority of the children in Experiment 3 chose the dog expert, only two children cited topic similarity, and half of the 14 children who chose the dog expert generated post hoc justifications or could not explain their choice at all. This suggests that even when children’s choices were similar to adults, children still struggled to explain why the dog expert should be better able to learn about cats. Likewise, children’s chance responses on the manifestation of learning questions suggested a lack of conviction that either expert would have learned more. The results of Experiment 3 suggest that 5-year-old children grasp the connection between what an individual already knows and that individual’s ability to learn about a new topic that is closely or distantly related to his or her existing expertise. This result is not surprising in light of evidence that children younger than age 5 can make similar kinds of judgments when inferring what experts in different domains are likely to already know (e.g., Lutz & Keil, 2002). However, the fact that children only attributed greater learning about cats to the dog expert when faced with two experts in different fields and that they only did so consistently on the overall knowledge question suggests that the mechanisms underlying children’s judgments may be quite different from those underlying adults’ reasoning. One potential interpretation is that whereas adults always view expertise in a related field as beneficial, children view existing knowledge as detrimental to learning about a new topic, even when those topics are closely taxonomically related. Nevertheless, when children are presented with two individuals who are equally “contaminated” by expertise, they can still rely on their understanding of taxonomic relationships to infer that expertise in a related category is less problematic than expertise in an unrelated category. This interpretation and others are discussed further below.

Comparison Across Experiments 1, 2, and 3 Experiments 1, 2, and 3 systematically varied the description of the non-dog-expert while using the exact same description of the dog expert. Comparing results across experiments reveals how the description of the non-dog-expert influenced participants judgments. Among adults, there were no significant differences in responses between these three experiments on either the overall knowledge question, w2(1, n = 63) = 3.579, p = .167, Cramer’s V = .238, or on the manifestation of learning questions, as revealed by a Kruskal-Wallis test, w2(2, n = 63) = 0.100, p = .951. However,

Experimental Psychology (2018), 65(1), 1–12

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

children showed significantly different responses on both the overall knowledge question, w2(1, n = 57) = 16.727, p < .001, Cramer’s V = .542, and the manifestation of learning questions, as revealed by a Kruskal-Wallis test, w2(2, n = 57) = 7.457, p = .024. Post hoc Mann-Whitney U-tests reveal that there were significant differences between children’s responses to the overall knowledge question in Experiments 1 and 2 (U = 129.00, p = .023) and Experiments 1 and 3 (U = 81.000, p < .001). The difference between the proportions of overall knowledge judgments in Experiments 2 and 3 was also marginally significant (U = 141.000, p = .064). Post hoc t-tests suggest that the difference in children’s responses to the manifestation of learning questions was primarily driven by differences between Experiment 1 and Experiment 3, t(39) = 3.002, p = .005 (all other comparisons were not significant). Taken together, these comparisons suggest that adults’ judgments were largely not affected by the description of the non-dogexpert. Children’s judgments, however, were influenced by the contrast between the dog expert and the other character, particularly in cases where the other character’s expertise was not specified or when the character was described as an expert on an unrelated category. This finding raises the question of what, if any, knowledge children attributed to the non-dog-expert when no information was provided about the non-dog-expert other than his lack of expertise about dogs – a question addressed in Experiment 4.

Experiment 4 Experiment 4 examined whether children’s and adults’ judgments about learning reflected their inferences about each character’s knowledge before the learning activity took place. Although the results of Experiment 3 suggest that 5-year-olds are capable of linking expertise about dogs to learning about cats, the results of Experiments 1 and 2 suggest that this may not be the case when an expert and nonexpert are contrasted. One possible explanation for this pattern is that children assume that the nonexpert knows about cats. For example, in Experiment 1, children may have believed that the nonexpert knew more about cats before the learning activity took place. Thus, children’s judgments about learning in Experiment 1 could reflect attributions of existing knowledge about cats, rather than successful acquisition of knowledge via the learning activity. Experiment 4 also explores whether the medium in which new information is presented influences children’s and adults’ judgments about learning efficacy. In the previous experiments, the characters were described as learning from a book about cats. There is evidence that children may not

Experimental Psychology (2018), 65(1), 1–12

appreciate the reliability of printed text as an information source until they are able to read (Einav, Robinson, & Fox, 2013; Robinson, Einav, & Fox, 2013) and that, among American children, this transition often occurs around age 5 (Corriveau, Einav, Robinson, & Harris, 2014). Although data are not available regarding the reading ability of the 5-year-olds in the current samples, it is likely that some children were pre-readers and some were early readers, and this may have influenced their understanding of the value of reading a text for acquiring information about cats. Thus, in Experiment 4, rather than reading a book about cats, the characters were described as watching a video about cats, a type of learning activity that we expected to be familiar to the children in our sample.

Methods Participants Nineteen 5-year-old children (Mage = 5.51 years, SD = 0.29, range = 5.02–5.95 years, 6 females) were recruited from a midsized city through local schools and a laboratory database. Parents identified 17 of the children as CaucasianAmerican and 1 child as African-American. None of the children were identified as Hispanic. Parents of one additional child declined to identify the child’s race or ethnicity. Children were tested individually in sessions lasting approximately 10 min. Eighteen university students (Mage = 20.7, SDage = 1.45, range = 18–24 years, 9 females) participated for course credit and were tested in individual sessions using the same procedure as the children. None of the participants had completed any of the previous experiments. Design, Materials, and Procedure Experiment 4 began with a description of the expert and nonexpert identical to Experiment 1, using the same character and dog images. Because children consistently indicated that the dog expert knew more about dogs in the previous experiments and in order to avoid influencing participants’ judgments of existing cat knowledge, the first question participants were asked in Experiment 4 was “who knows more about cats?” This was followed by the question “who knows more about dogs?” that served as an attention and memory check. Next, participants were asked “who knows more about cars?” and “if you had a question about dogs, who would you ask?” Participants then heard a description of the learning activity that was identical to Experiment 1, except that instead of “reading about cats,” the characters were described as “watching a video about cats” and the miniature books about cats were replaced with an image of a computer monitor with the same photograph of a cat that

Ó 2018 Hogrefe Publishing

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

was used in the book covers. Children were then asked the overall knowledge and manifestation of learning questions using the same phrasing as the prior experiments, except that the word “read” was replaced by “watched in the video” in the first question (e.g., “Who do you think will remember what they watched in the video about cats better?”).

non-dog-expert for the overall knowledge question spontaneously indicated that they were unsure about their choice, but that they wanted to be consistent with their prior attribution of cat knowledge to the non-dog-expert. This suggests that responding to the initial cat knowledge question may have influenced subsequent decisions about learning, and that these data should be interpreted with caution.

Results and Discussion

Comparison Between Experiments 1 and 4

For the initial question about cat knowledge, all child participants and 15 out of 18 adult participants (83%) indicated that the non-dog-expert knew more about cats, with a Mann-Whitney U-test showing that the difference between age groups approached significance, U = 142.500, p = .067. All participants also indicated that the dog expert knew more about dogs and that they would consult the dog expert for information about dogs. Fifteen children (79%) indicated that the non-dog-expert would know more about cars, a proportion that significantly differed from chance, w2 = 6.368, p = .012. Ten adults (56%) also did so, but this proportion did not differ from chance, w2 = 0.222, p = .637, nor from children’s responses, U = 131.000, p = .134. Thus, both children and adults consistently attributed prior knowledge about cats to the non-dog-expert. Most of the children also attributed superior knowledge about cars to the non-dog-expert, whereas adults did not have consistent intuitions (see ESM 1). This supports that children as well as adults tend to assume that knowledge is equitably distributed, such that they inferred the non-dog-expert would know more about topics other than dogs. For the overall knowledge question, 7 children (37%) and 11 adults (61%) indicated that the dog expert would know more about cats after watching the video (see Figure 1). Each age group’s proportion of responses did not differ from chance, w2s 1.316, ps .251, nor from each other, U = 129.500, p = .145. Cochran’s Q tests showed no significant differences in responses for the manifestation of learning items among either age group, Qs 2.167, ps .338. Neither adults (M = 1.78, SD = 1.35) nor children (M = 1.21, SD = .976) attributed learning to either expert at rates significantly different from chance, ts 1.292, ps .213. McNemar tests showed that both children, p = .016 (twotailed), and adults, p = .008 (two-tailed), were significantly more likely to select the dog expert for the overall knowledge question than for the initial cat knowledge question. This shift may reflect beliefs that the dog expert is more intelligent or competent overall, or uncertainty about whether the characters’ prior knowledge would influence learning. Notably, some adults who chose the

Because the descriptions of the characters were identical, comparing beliefs about the non-dog-expert’s initial knowledge in Experiment 4 with responses to the overall learning question in Experiment 1 can provide insight into judgments about the role of prior knowledge when determining success at learning. Mann-Whitney U-tests showed no significant difference between children’s responses across experiments, U = 171.000, p = .162, but there was a significant difference in adults’ responses, U = 88.500, p < .001, with most adults attributing baseline knowledge about cats to the non-dog-expert in Experiment 4 but superior learning outcomes to the dog expert in Experiment 1. This suggests that there is a developmental shift, where children attribute greater knowledge about cats to the non-dog-expert both before and after the learning activity, yet adults intuit that the dog expert is more capable of learning about cats, even if the non-dog-expert is initially assumed to know more about cats.

Ó 2018 Hogrefe Publishing

General Discussion In four experiments, children and adults judged whether an expert on dogs would be more or less successful than a non-dog-expert in learning about cats. Across variations in the non-dog-expert’s description, adults consistently treated expertise as advantageous for learning new material in a related category (i.e., they believed that the knowledge, skills, and experiences related to being a dog expert would result in the acquisition of greater knowledge about cats and a greater capacity to apply that knowledge). In contrast, 5-year-olds’ judgments were affected by the description of the non-dog-expert’s knowledge. When the non-dog-expert was described as not knowing about dogs or having relatively limited knowledge about dogs (Experiments 1, 2, and 4), children sometimes indicated that he would be more capable of learning than the dog expert. However, when the non-dog-expert was presented as an expert in an entirely different domain (cars; Experiment 3), children consistently attributed greater learning to the dog expert. Children’s justifications for their judgments suggest that

Experimental Psychology (2018), 65(1), 1–12

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

they consider a number of different, and potentially competing, attributes when judging an individual’s ability to learn and that these may not always be attributes that adults would prioritize. In addition, the results of Experiment 4 suggest that children may be relying on inferences about an individual’s existing knowledge when judging the capacity to learn. Children exhibit an impressive capacity for accurately inferring what experts know, even when dealing with distantly related categories or unfamiliar phenomena (Aguiar et al., 2012; Danovitch & Keil, 2004; Lutz & Keil, 2002). However, in the present study, children only appeared to apply their understanding of domains of expertise when they were asked to evaluate the learning potential of two experts. When the learning capacity of a nonexpert was contrasted with that of an expert, children responded as if existing expertise might potentially interfere with new learning. In Experiment 1, children favored the individual whose existing knowledge was not specified over the dog expert, but in Experiment 2, when both individuals knew something about dogs, children were uncertain who would learn better about cats. Only in Experiment 3, when the characters had equivalent amounts of existing knowledge about different topics, did children shift to a different heuristic and attribute greater learning to the dog expert than the car expert, albeit only for the overall knowledge question. Importantly, even in Experiment 3, children rarely cited topic similarity as the basis for attributing superior learning to the dog expert, suggesting that children had difficulty discerning that knowledge about dogs would be applicable to learning about cats. The current results contribute to mounting evidence that neither adults nor children perceive expertise in one domain as indicative of superior overall competence or intelligence (see Danovitch & Keil, 2007; Koenig & Jaswal, 2011). Although it may seem plausible that an individual who becomes an expert in a particular area has stronger comprehension skills or a larger memory capacity than the average nonexpert, there was no evidence of this belief among child participants in the current study. Even among participants who indicated that the dog expert would be better at learning about cats, no adult or child cited the dog expert’s general intelligence when explaining their judgment. These findings raise interesting questions about whether there are any skills (e.g., being a strong reader), experiences (e.g., exposure to diverse exemplars), or types of expertise (e.g., expertise in a superordinate level category) that would prompt children or adults to view an individual as more capable of learning across related or unrelated domains. At age 5, children do not seem to appreciate that an expert about a general domain (e.g., animals) is likely to have a better understanding of the principles underlying that domain than an individual Experimental Psychology (2018), 65(1), 1–12

with a specialized, narrow area of expertise (e.g., poodles; Landrum & Mills, 2015). Likewise, even though 5-year-olds are capable of clustering expert knowledge according to shared underlying principles, they prefer to cluster knowledge by topic (Danovitch & Keil, 2004). In the current studies, children may have had difficulty discerning that expertise about dogs entails an understanding of the underlying principles that would also be applicable to, and therefore facilitate learning about cats, whereas adults may have more readily made that connection. More generally, together with prior studies, the current results suggest that children have a more fragile and fragmented notion of expertise than adults, and that this may underlie developmental differences in judgments about the ease of learning about related categories. There are a number of additional factors that may explain differences in children’s and adults’ judgments of how expertise influences learning in the current studies. First, although dogs, cats, and cars were chosen as the target topics of expertise and learning because they are common and familiar to children and adults in this population, familiarity may have also had unintended consequences for judgments about learning and expertise. For example, the popular media often portray interest and liking of dogs and cats as being exclusive of each other (e.g., classifying individuals as a “dog person” or a “cat person”). Consequently, children, and to some extent, adults, maybe have been prone to believe that the dog expert would dislike cats and show less interest in learning about cats. Additional studies where the expert is described as knowing and learning about less popular and value-laden basic-level categories (e.g., an expert about ducks learning about chickens) or a broader range of categories are needed to better understand to what extent the current results generalize across exemplars and domains. Second, the nature of the learning activity may have had differential effects on children’s and adults’ judgments about learning. As noted earlier, 5-year-olds who were prereaders may have had trouble appreciating the value of the book as an information source in the first three experiments. Moreover, both children and adults may have differed in their attention to how closely the way in which the expert initially acquired his knowledge aligned with the nature of the learning activity. In the current experiments, how the dog expert initially learned about dogs was not specified, but the dog expert was described as having frequent exposure to dogs and working with dogs. This description may have led some participants to assume that he acquired his expertise through personal experience and that his experience-based expertise would not be as relevant to learning the type of information that is found in books or videos. Although we believe this is unlikely given children’s weak understanding of the learning process in Ó 2018 Hogrefe Publishing

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

general (Sobel & Letourneau, 2015), it would be informative to further explore how the means by which an individual has acquired expertise in the past influences children’s and adults’ judgments about how effectively that individual will learn from different types of sources in the future. Third, the overall knowledge and manifestation of learning questions in the current experiments emphasize learning outcomes, rather than the learning process. Recent evidence suggests that 5-year-olds struggle to define learning and that, even when they do, they have a very rudimentary understanding that learning is a process (Sobel & Letourneau, 2015). The design of the current experiments was based on evidence that young children consider demonstrations of knowledge (e.g., the ability to sing a song) as evidence of learning (Sobel, 2015). However, this design leaves open the question of whether children understand that expertise can support the learning process. For example, had participants been asked to predict which character would learn about cats more quickly, they may have demonstrated different intuitions. Finally, judgments about learning may have been influenced by the degree to which an individual is perceived to be interested in and motivated to learn about a new topic. This factor was intentionally kept neutral in the present study, but experts report that their knowledge about a topic often leads them to generate new questions, and motivates them to learn more (Silvia, 2008). Hence, judgments about learning could be influenced by beliefs about an individual’s motivation and interest in learning. It may be that adults view expertise in one category as an indicator of interest and motivation to study closely related categories, yet children take a more narrow view of how expertise influences motivation to learn. For instance, children may view expertise about dogs as sign of an exclusive interest that extends neither to closely related topics, such as cats, nor to distantly related topics such as cars. Taken together, the current results suggest that children’s reasoning about the relationship between knowledge and learning is influenced not only by their understanding of categorical relationships and expertise, but also by other factors, such as a view of the mind as having a limited capacity to store information and assumptions about prior knowledge. Conversely, when adults determine how well a person will learn about a particular topic, established expertise in a related domain seems to take precedence over other considerations. Much of the research on children’s understanding of expertise and information sources has relied on forced-choice paradigms where children choose between two or more sources who have expertise in different domains (e.g., Lutz & Keil, 2002), yet the current results raise questions about whether the design of these tasks has led researchers to overestimate children’s understanding

Ó 2018 Hogrefe Publishing

and other ways in which prior findings may have been influenced by task demands. For instance, when faced with an expert and a nonexpert, children may have to sort through competing intuitions about which individual would be more knowledgeable about other categories and even adults may be prone to infer that knowledge is somewhat equitably distributed among individuals. Providing children with choices between experts in distinct domains may also have limited ecological validity: in reality, children often lack information about an individual’s existing knowledge (e.g., when meeting someone new) and they may have to decide between consulting with or sharing information with someone whose area of expertise is known to them and someone whose knowledge status is less clear. The current findings suggest that, unlike adults, children do not always view expertise as a clear indicator of one’s learning capacity for a related domain, and that, in some situations, they may infer that nonexperts have an advantage when it comes to learning. Acknowledgments We thank the staff, parents, and students at participating schools for their support. Thanks also to members of the KID lab for their assistance, and to Nicholaus Noles for helpful feedback. Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1618-3169/a000387 ESM 1. Data (.xls) Raw data for Experiments 1–4. ESM 2. Data (.xls) Explanation transcription and coding scheme for Experiments 1–3.

References Aguiar, N. R., Stoess, C. J., & Taylor, M. (2012). The development of children’s ability to fill the gaps in their knowledge by consulting experts. Child Development, 83, 1368. https:// doi.org/10.1111/j.1467-8624.2012.01782.x Associated Press. (2009, May 21). Auto recovery leader announces $50 million in aid. Retrieved from http://www.fox28.com/story/ 10402224/auto-recovery-leader-announces-50-million-in-aid Castel, A. D., McCabe, D. P., Roediger, H. L., & Heitman, J. L. (2007). The dark side of expertise: Domain-specific memory errors. Psychological Science, 18, 3–5. https://doi.org/10.1111/ j.1467-9280.2007.01838.x. Chi, M. T. H. (2006). Two approaches to the study of experts’ characteristics. In K. A. Ericsson, N. Charness, P. J. Feltovich, & R. R. Hofmann (Eds.), The Cambridge handbook of expertise and expert performance (pp. 21–30). Cambridge, UK: Cambridge University Press.

Experimental Psychology (2018), 65(1), 1–12

J. H. Danovitch & C. K. Shenouda, Understanding Expertise and Learning

Corriveau, K. H., Einav, S., Robinson, E. J., & Harris, P. L. (2014). To the letter: Early readers trust print-based over oral instructions to guide their actions. British Journal of Developmental Psychology, 32, 345–358. https://doi.org/10.1111/bjdp. 12046 Danovitch, J. H. (2013). Understanding expertise: The contribution of social and cognitive processes to social judgments. In M. R. Banaji & S. A. Gelman (Eds.), Navigating the social world: What infants, children, and other species can teach us (pp. 225–229). New York, NY: Oxford University Press. Danovitch, J. H., & Keil, F. C. (2004). Should you ask a fisherman or a biologist? Developmental shifts in ways of clustering knowledge. Child Development, 75, 918–931. https://doi.org/ 10.1111/j.1467-8624.2004.00714.x Danovitch, J. H., & Keil, F. C. (2007). Choosing between hearts and minds: Children’s understanding of moral advisors. Cognitive Development, 22, 110–123. https://doi.org/10.1016/ j.cogdev.2006.07.001 Danovitch, J. H., & Keil, F. C. (2008). Young humans: The role of emotions in children’s evaluation of moral reasoning abilities. Developmental Science, 11, 33–39. https://doi.org/10.1111/ j.1467-7687.2007.00657.x Danovitch, J. H., & Noles, N. S. (2014). Categorization ability, but not theory of mind, contributes to children’s developing understanding of expertise. In P. Bello, M. Guarini, M. McShane, & B. Scassellati (Eds.), Proceedings of the 36th Annual Conference of the Cognitive Science Society (pp. 2097–2102). Austin, TX: Cognitive Science Society. Drummey, A. B., & Newcombe, N. S. (2002). Developmental changes in source memory. Developmental Science, 5, 502–513. https://doi.org/10.1111/1467-7687.00243 Einav, S., Robinson, E. J., & Fox, A. (2013). Take it as read: Origins of trust in knowledge gained from print. Journal of Experimental Child Psychology, 114, 262–274. https://doi.org/10.1016/ j.jecp.2012.09.016 Gelman, S. A. (2003). The essential child: Origins of essentialism in everyday thought. New York, NY: Oxford University Press. Gobbo, C., & Chi, M. T. H. (1986). How knowledge is structured and used by expert and novice children. Cognitive Development, 1, 221–237. https://doi.org/10.1016/S0885-2014(86)80002-8 Harris, P. L. (2012). Trusting what you’re told: How children learn from others. Cambridge, MA: Belknap of Harvard University Press. Keil, F. C., Stein, C. M., Webb, L., Billings, V., & Rozenblit, L. (2008). Discerning the division of cognitive labor: An emerging understanding of how knowledge is clustered in other minds. Cognitive Science, 32, 259–300. https://doi.org/10.1080/ 03640210701863339 Koenig, M. A., & Jaswal, V. K. (2011). Characterizing children’s expectations about expertise and incompetence: Halo or pitchfork effects? Child Development, 82, 1634–1647. https:// doi.org/10.1111/j.1467-8624.2011.01618.x Lane, J. D., Wellman, H. M., & Evans, E. M. (2010). Children’s understanding of ordinary and extraordinary minds. Child Development, 81, 1475–1489. https://doi.org/10.1111/j.1467-8624. 2010.01486.x Landrum, A. R., & Mills, C. M. (2015). Developing expectations regarding the boundaries of expertise. Cognition, 134, 215–231. https://doi.org/10.1016/j.cognition.2014.10.013

Experimental Psychology (2018), 65(1), 1–12

Landrum, A. R., Mills, C. M., & Johnston, A. M. (2013). When do children trust the expert? Benevolence information influences children’s trust more than expertise. Developmental Science, 16, 622–638. https://doi.org/10.1111/desc.12059 Lutz, D. J., & Keil, F. C. (2002). Early understanding of the division of cognitive labor. Child Development, 73, 1073–1084. https:// doi.org/10.1111/1467-8624.00458 Mervis, C. B., & Crisafi, M. A. (1982). Order of acquisition of subordinate-, basic-, and superordinate-level categories. Child Development, 53, 258–266. Robinson, E. J., Einav, S., & Fox, A. (2013). Reading to learn: Prereaders’ and early readers’ trust in text as a source of knowledge. Developmental Psychology, 49, 505. https://doi.org/ 10.1037/a0029494 Silvia, P. J. (2008). Interest – The curious emotion. Current Directions in Psychological Science, 17, 57–60. https://doi.org/ 10.1111/j.1467-8721.2008.00548.x Sobel, D. M. (2015). Can you do it? How preschoolers judge whether others have learned. Journal of Cognition and Development, 16, 492–508. https://doi.org/10.1080/15248372. 2013.815621 Sobel, D. M., & Letourneau, S. M. (2015). Children’s developing understanding of what and how they learn. Journal of Experimental Child Psychology, 132, 221–229. https://doi.org/ 10.1016/j.jecp.2015.01.004 Sobel, D. M., Li, J., & Corriveau, K. H. (2007). “They danced around in my head and then I learned them”: Children’s developing conceptions of learning events. Journal of Cognition and Development, 8, 345–369. https://doi.org/10.1080/ 15248370701446806 Tager-Flusberg, H. (1985). Basic level and superordinate level categorization by autistic, mentally retarded, and normal children. Journal of Experimental Child Psychology, 40, 450–469. https://doi.org/10.1016/0022-0965(85)90077-3 Taylor, M., Esbensen, B. M., & Bennett, R. T. (1994). Children’s understanding of knowledge acquisition: The tendency for children to report that they have always known what they have just learned. Child Development, 65, 1581–1604. https://doi. org/10.1111/j.1467-8624.1994.tb00837.x Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of-mind development: The truth about false belief. Child Development, 72, 655–684. http://doi.org/10.1111/14678624.00304 Received August 30, 2016 Revision received June 30, 2017 Accepted September 8, 2017 Published online February 8, 2018

Judith Danovitch Department of Psychological and Brain Sciences 317 Life Sciences University of Louisville Louisville, KY 40292 USA j.danovitch@louisville.edu

Ó 2018 Hogrefe Publishing

Research Domain Criteria (RDoC) – interdisciplinary constructs of psychopathology Contents and topics include • Psychopathology research in the spirit of the Research Domain Criteria (RDoC) initiative • Misunderstanding RDoC • The psychophysiology of anxiety and mood disorders: the RDoC challenge • Challenges of fear conditioning research in the age of RDoC • Effects of an anxiety-specific psychometric factor on fear conditioning and fear generalization • Trait worry and neural correlates of emotion regulation • Assessing hedonic bias in emotional scene memory: implications for clinical research • Frontostriatal connectivity during reward anticipation: a neurobiological mechanism cutting across alcohol use disorder and depression? • Dissociating pathological buying from obsessive compulsive symptoms using delay discounting • Bridging the gaps between basic science and cognitive-behavioral treatments for anxiety disorders in routine care: current status and future demands • Identifying patterns in complex field data: clustering heart rate responses of agoraphobic patients undertaking situational exposure

Alfons O. Hamm (Editor)

Mechanisms of Mental Disorders Zeitschrift für Psychologie, Vol. 225/3 2017, iv + 128 pp., large format US $49.00 / € 34.95 ISBN 978-0-88937-544-4 Our understanding of the mechanisms guiding adaptive behavior and normal psychological functioning and of how these processes become disrupted in various forms of mental disorders has increased but translating the results of basic psychological research into clinical practice is still in the early stages. One problem for this translation is that the classification of mental disorders is still almost exclusively descriptive. The Research Domain Criteria (RDoC) initiative aims to redress this by promoting the development of an interdisciplinary science

www.hogrefe.com

of psychopathology that consists of dimensional constructs integrating models and findings from psychology, biology, and behavioral neuroscience. This compilation focuses on research in experimental psychopathology that investigates such constructs. It includes theoretical and empirical work on different psychological constructs (such as fear learning and mental imagery) that are relevant for a better understanding of the mechanisms of mental disorders across a large spectrum of categorical diagnoses.

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!

Short Research Article

Interference in Dutch–French Bilinguals Stimulus and Response Conﬂict in Intra- and Interlingual Stroop James R. Schmidt,1 Robert J. Hartsuiker,2 and Jan De Houwer1 1

Department of Experimental Clinical and Health Psychology, Ghent University, Belgium

Department of Experimental Psychology, Ghent University, Belgium

Abstract: In the present manuscript, we investigate the source of congruency effects in a group of Dutch–French bilinguals. In particular, participants performed a color-identification Stroop task, in which both (first language) Dutch and (second language) French distracting color words were presented in colors. The typical finding is impaired responding when the word and color are incongruent (e.g., “red” in blue) relative to congruent (e.g., “red” in red). This congruency effect is observed for both first and second language distracting color words. The current experiment used a 2-to-1 keypress mapping manipulation, which allows one to separate stimulus conflict (i.e., conflict between word and color meanings) and response conflict (i.e., conflict between potential responses). For both the first and second language, both stimulus and response conflict were observed. These results suggest that second language words influence semantic and response processing similarly to first language words, rather than having diminished semantic and/or response influences. Keywords: bilingualism, Stroop effect, stimulus conflict, response conflict, semantics, response selection

Considerable research has focused on bilingual cognition (for reviews, see Bialystok, Craik, & Luk, 2012; Koda, 1996; Werker & Byers-Heinlein, 2008). One particular focus is on how second language words influence cognitive processing in a qualitatively different (or not) way than first language words. In the present report, we explore the source of crosstalk between languages within a bilingual version of the color-word Stroop paradigm. In particular, we investigate the extent to which foreign words might have diminished impact on semantic identification and/or response decision processes, or to what extent first and second language words might influence cognitive processing in a similar fashion. In the Stroop task (Stroop, 1935; for a review, see MacLeod, 1991), participants are tasked with the goal of identifying the print color of a color word, while ignoring the meaning of the word itself (e.g., say “green” to the word “red” printed in green). The congruency (or Stroop) effect is the observation that participants are typically slower and less accurate to incongruent trials (e.g., the word “green” printed in yellow), where the meaning of the word and color mismatch, relative to congruent trials Ó 2018 Hogrefe Publishing

(e.g., “green” in green), where the meaning of the word and color match. One question of interest in the Stroop literature is the origin of the conflict. That is, what produces the conflict between a color word and an incongruent color? On the one hand, conflict could occur between the meaning of the word and of the color (e.g., lexical-semantic representations), which we term here stimulus conflict (Glaser & Glaser, 1989; Mackinnon, Geiselman, & Woodward, 1985; Stirling, 1979). An alternative possibility is that the response engendered by the word and the response engendered by the color compete for selection, which we term here response conflict (Klein, 1964; Posner & Presti, 1987). As will be discussed shortly, the general consensus is that both stimulus and response conflict contribute to the standard Stroop effect (Augustinova & Ferrand, 2014; Augustinova, Silvert, Ferrand, Llorca, & Flaudias, 2015; Ferrand & Augustinova, 2014). One clear line of evidence for both stimulus and response contributions to the Stroop effect comes from 2-to-1 mapping experiments. For instance, De Houwer (2003; see also, A. T. Chen, Bailey, Tiernan, & West, 2011; Experimental Psychology (2018), 65(1), 13–22 https://doi.org/10.1027/1618-3169/a000384

De Houwer, 2004; Hasshim & Parris, 2015; Jongen & Jonkman, 2008; van Veen & Carter, 2005) presented participants with a 2-to-1 Stroop task, in which participants responded to two colors for each key. For instance, a given participant might have been instructed to respond to the colors blue and yellow with the left key and the colors green and red with the right key, as illustrated in Figure 1. This produces three conditions, rather than just two. First, there are identity trials (e.g., “blue” in blue), which are typical congruent trials in which the word matches the color. The response to the word, by extension, also matches the response to the color. Second are same response trials (e.g., “blue” in yellow), which are incongruent in meaning (i.e., blue and yellow are different colors) but mapped to the same response key (i.e., the responses for blue and yellow are both the left key). A difference between identity and same response trials therefore indicates stimulus conflict, and not response conflict. Third are different response trials (e.g., “blue” in green), in which the word mismatches the meaning of the color (as in same response trials), but the assigned responses also mismatch (i.e., the response keys for blue and green are different). Any difference between same and different response trials thus indicates response conflict. Evidence for both stimulus and response conflict was observed, with same response trials being slower than identity trials (stimulus conflict), but faster than different response trials (response conflict). Not all stimulus types produce both stimulus and response conflict. For instance, color associates also produce a congruency effect, with incongruent color associates (e.g., “sky” in red) impaired relative to congruent color associates (e.g., “sky” in blue). Following debate about whether this effect was due to stimulus or response conflict (Glaser & Glaser, 1989; Klein, 1964; Mackinnon et al., 1985; Posner & Presti, 1987; Stirling, 1979), Schmidt and Cheesman (2005) used the 2-to-1 mapping procedure and found exclusively stimulus conflict for color associates (cf. Risko, Schmidt, & Besner, 2006). This was interpreted as indicating that associates spread activation to related concepts in semantics, producing semantic conflict with the target color concept, but are not potent enough to indirectly bias a potential response (e.g., “sky” facilitating “blue” strongly enough to retrieve a D-key response linked to “blue,” which also applies in a verbal task where “sky” is not a potential response). Similarly, response conflict can be observed for distracting stimuli where stimulus conflict would be impossible, such as in the Simon task (Simon, Craft, & Webster, 1973; Simon & Rudell, 1967); there can be no stimulus conflict between a color and a location as these involve different stimulus dimensions. Thus, both stimulus and response conflict contribute to performance in an intralingual (within language) Stroop procedure, though both types of interference are not necessarily Experimental Psychology (2018), 65(1), 13–22

J. R. Schmidt et al., Bilingual Stroop Effect

Identity “blue” in blue blue

yellow F

—

green

red J

Same res ponse “yellow” in blue

—

Different res ponse “green” in blue Figure 1. Illustration of the 2-to-1 mapping procedure. In addition to stimulus-/response-compatible identity trials, and stimulus-/ response-incompatible different response trails, there are also stimulus-incompatible but response-compatible same response trials.

observable for all stimulus types. In sum, stimulus and response conflict do not necessarily develop in parallel for any given association (see also, Schmidt, Crump, Cheesman, & Besner, 2007; Schmidt & De Houwer, 2012). The Stroop effect has also been used extensively to study interference between two languages (interlingual) in bilingual participants (Altarriba & Mathis, 1997; Atalay & Misirlisoy, 2012; H. C. Chen & Ho, 1986; Dalrymple-Alford, 1968; Dyer, 1971; La Heij et al., 1990; Mägiste, 1984, 1985; Preston & Lambert, 1969; Smith & Kirsner, 1982; Tzelgov, Henik, & Leiser, 1990). It is known from this research that a Stroop effect can be observed with both color-word distracters of the first language (L1) and of the second language (L2). For instance, a native English speaker who also speaks French will be impaired by incongruent French color words, in addition to incongruent English color words. However, the standard finding is that the effect for L2 words is smaller than that for L1 words. For instance, the native English speaker described above will be more impaired by a trial such as “yellow” in green than by a trial such as “jaune” (French for “yellow”) in green (for a review, see MacLeod, 1991). Aside from the issue of the overall size of the congruency effect in native and foreign languages is the source of the conflict. The key question of the present manuscript is whether L2 distracting words engender both stimulus and response conflict or only one of the two. This can have important implications for theorizing about language cognition, as much debate centers on how language lexicons are connected to each other and to semantics. For instance, consider the account provided by Kroll and Stewart (1994), presented in Figure 2. According to this account, the L1 and L2 lexical representations for words are connected. For instance, “zwart” and “noir” (respectively, Dutch and French for “black”) are connected in memory. In addition, both words are connected to a single semantic representation for black. It is additionally assumed in this model that: (a) L1 words are more strongly connected to semantic concepts (i.e., heavy overlearning) than L2 words, and (b) L2 words are more strongly connected to L1 words Ó 2018 Hogrefe Publishing

J. R. Schmidt et al., Bilingual Stroop Effect

zwart geel groen bruin

noir jaune vert marron

Semantics black yellow green brown Figure 2. Illustration of the Kroll and Stewart (1994) model of the connections between first (L1) and second (L2) lexicons and semantics. Note the asymmetries in the connections between lexicons and to semantics.

(i.e., learned as translations of the already well-known first language). Of course, there are many alternative versions of the model presented in the figure (e.g., which involve different connection strengths between memory stores). However, the key question of the current report is where conflict occurs within the system: at the level of semantics, at the level of responses, or both. In that vein, the present work made use of the 2-to-1 mapping design1 that included color words from both the dominant L1 language (Dutch) and a nondominant L2 language (French). Three possible results might occur. Of course, the first is that both stimulus and response conflict will be observed for L2 words, just as with L1 words. In this case, we would expect the same pattern of results for L2 French color words (i.e., identity < same responses < different response). However, conflict effects from L2 words are generally smaller than those from L1 words, which might mean that one and/or both of the conflict components are decreased for L2 words. A second possibility, therefore, is that foreign color words produce exclusively stimulus conflict. For instance, Altarriba and Mathis (1997) argued that foreign color words are directly linked to semantics, even at early stages of learning a new language. According to this view (which contrasts sharply with that presented in Figure 2), the foreign word “noir” interferes with the semantic identification of yellow print color (i.e., slower answer to question “What color is that?”). That is, presentation of the word “noir” will activate the semantic representation for black, and the yellow print color activates the semantic representation for yellow, leading to conflict in semantics (cf. Schmidt, Cheesman, & Besner, 2013). However, depending on further assumptions one makes, “noir,” according to this 1

account, will not interfere afterward when a response (keypress) is being selected. If so, then L2 color words act like color associates to the L1 equivalents (e.g., “noir” as an associate to “black”). Like with color associates, the assumption would be that “noir” is unable to retrieve the response linked to black, even though (unlike with color associates) “noir” is a direct lexical translation of “black.” If this is the case, then we should expect a difference between identity and same response trials for L2 color words, but no difference between same and different response trials. A third possibility is that foreign color words might produce exclusively response conflict. For instance, it could be the case that “noir” interferes with what response needs to be made (i.e., slower answer to the question “What key do I need to press?” or, in verbal naming, “What do I need to verbalize?”). For this to occur, the foreign distracting word would need to be able to automatically bias a response without interfering with identification of the stimulus color itself. For instance, it might be that stimulus conflict occurs exclusively within semantics, to which L2 words are weakly connected (Kroll & Stewart, 1994), as illustrated earlier in Figure 2. Instead, L2 words might be quickly and automatically translated to their L1 lexical equivalents (e.g., because the words were learned as translations), allowing an indirect biasing of responses via encoded lexical-response instruction memories. As an added consideration, either or both of the stimulus and response conflict effects for L2 color words might depend on the type of color word. Cognates are translation equivalents with similar spellings in both languages, such as “bleu” (blue) in French and “blauw” in Dutch, which are typically similar because of a shared etymology. Of course, for cognates there is a compatibility in pronunciation (in the case of a verbal task), in addition to orthographic (spelling) similarities between the words. Non-cognates, in contrast, are dissimilar, such as “jaune” (yellow) in French and “geel” in Dutch. For non-cognates, there is little or no overlap in pronunciation/spelling. Indeed, similarity between the words in the two languages does matter for the amount of interference observed (e.g., Dyer, 1971), with larger interference effects for cognates. Because effects for cognates might be due to processes that are less interesting for our present purposes (see Costa, Miozzo, & Caramazza, 1999), such as priming across languages due to spelling similarities (or even just first letter priming), we opted to use Dutch–French non-cognates (Dutch/French: zwart/noir [black], groen/ vert [green], bruin/marron [brown], and geel/jaune [yellow]).

It is relevant to point out that this design makes use of a keypress rather than verbal modality. Added differences between languages might be observable that are specific to a verbal response modality (e.g., conflict during articulation), which we discuss in the Discussion section.

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 13–22

Method Participants Ninety-three Ghent University undergraduates (71 female, 22 male) participated in the study in exchange for €5. We aimed for a large sample relative to previous experiments with the 2-to-1 mapping procedure because we were testing for potential language differences of unknown size. The exact sample size was determined by the number of participants who signed up during the allotted testing time. On the recruitment website, we explicitly solicited participants who were native Dutch speakers with some familiarity with French. Language questionnaires (to be discussed shortly) were used to confirm the fit of participants with these criteria. Average language metric scores are presented in Table 1. All participants seemed to sufficiently fit our language dominance criteria, so no participants were removed from the sample. To broadly characterize the sample, English is learned as the second foreign language in schools and practically everyone in Flanders develops near native-speaker proficiency in the language.2 French, on the other hand, while learned earlier on in school, is not nearly as well developed in the Flemish population. This was borne out in the language metrics. While participants self-rated their French proficiency relatively moderately (6.05 on a 0–10 scale), their more objective French test scores (see description of the LEXTALE_FR later) were quite low. Indeed, even the highest scorers in the sample were nowhere near the range of scores for native speakers. Thus, our sample was generally familiar with French, but had only weak French skills. See the Appendix for more detailed information on the language demographics.

Apparatus and Materials The main part of the experiment was programmed in E-Prime 2 (Psychology Software Tools, Pittsburgh, PA) and conducted on a standard PC. Responses were made with the “F” (left) and “J” (right) keys on an AZERTY keyboard with the two index fingers. Prior to the computer portion of the experiment, participants were also given a short pen-and-paper survey to fill out. This included the LEXTALE_FR (Brysbaert, 2013) with Dutch-language instructions. In this test, participants are presented with a list of 84 French-looking words, only about 2/3 of which are actual French words (e.g., “église”), whereas the remaining 1/3 are not (e.g., “metter”). The participants are informed to select the words that they are fairly certain are actual French words. Correct “hits” are rewarded with one point, and incorrect “false alarms” are penalized by two points. 2

J. R. Schmidt et al., Bilingual Stroop Effect

Table 1. Mean language scores with standard errors Mean

Years French

7.93 years

0.144

French Level

6.05 (0–10)

0.137

Score

5.23*

0.911

LEXTALE_FR

LEAP-Q Dominance Dutch

1.02**

0.015

Dominance French

3.10**

0.061

Order Dutch

1.09**

0.039

2.29**

0.056

Order French Dutch Use (%)

71.30

1.680

French Use (%)

7.40

0.970

Acquisition

1.05 years

0.190

Fluent

4.73 years

0.250

Reading

6.09 years

0.150

Fluent Read

8.13 years

0.210

9.22 years

0.300

Dutch

French Acquisition Fluent

13.06 years

0.370

Reading

11.24 years

0.220

Fluent Read

13.96 years

0.250

Note. *1st percentile L1, 48th L2, **ranks from 1st up. All values rounded to the shown decimal place.

Random guessing will therefore produce a score of around zero, with higher scores for better hit-to-false alarm ratios. The questionnaire also asks for gender, native language, years of French training in school, and a self-rating of French knowledge ranging from 0 (= almost none) to 10 (= perfect). Appended to this were a subset of questions from the Dutch for Belgium version of the Language Experience and Proficiency Questionnaire (LEAP-Q; Marian, Blumenfeld, & Kaushanskaya, 2007). In particular, the first three questions were retained, which asked, respectively, for a list of languages in order of dominance, a list of languages in order of acquisition, and the percentage with which the participant used each of their languages in the recent period. Also retained from the LEAP-Q were two boxes, one for Dutch and one for French, asking for the age the participant began acquiring the language, became fluent in the language, began learning to read in the language, and became fluent in reading the language. The purpose of these questionnaires was primarily to assure participants had the correct language dominance, but we also consider correlations of these metrics with the observed congruency effects. Finally, as an addition to these two questionnaires, participants were asked to give the French translations of the four Dutch color words used

For this reason, we also ensured that the French stimuli were non-cognates with English color words.

Experimental Psychology (2018), 65(1), 13–22

Ó 2018 Hogrefe Publishing

J. R. Schmidt et al., Bilingual Stroop Effect

in the experiment. This was to get a general idea of how familiar the stimuli were to participants (see Appendix for a summary) and to make sure the participants knew the correct translation of each of the colors.

respond in 2,000 ms, then the message “Fout” (“False/ Error”) or “Te Traag” (“Too slow”), respectively, appeared in red (255,0,0) for 1,000 ms before the next trial.

Design

Results

During the main part of the experiment, participants were presented with the Dutch and French color words for “black,” “green,” “brown,” and “yellow” (Dutch/French: “zwart/noir,” “groen/vert,” “bruin/marron,” and “geel/ jaune,” respectively). Notably, these four words are noncognates, unlike several other color words (e.g., “blauw/ bleu” [blue], “rood/rouge” [red], etc.). The corresponding print colors were black (0,0,0), green (0,128,0), brown (139,69,19), and yellow (255,215,0), corresponding to “black,” “green,” “saddle brown,” and “gold” in the standard E-Prime color palette. For each participant, two colors were mapped to the left key (e.g., black and yellow) and two to the right key (e.g., brown and green). Which colors were mapped to which keys and in which combinations was fully counterbalanced across participants (i.e., six factorial combinations) on the basis of participant number. These manipulations allow for two within factors. The first is distracter language (Dutch vs. French). The second is congruency: identity when the word and color match, same response when the mismatching word and color are mapped to the same key, and different response when the mismatching word and color are mapped to different keys. In total there were three larger blocks of trials, separated by a 5-s pause. Each of the larger blocks consisted of two smaller subblocks. In each subblock, each of the eight words was presented once each in all four colors (32 trials), selected randomly without replacement. Thus, there were 192 experimental trials in total across the six subblocks. The main phase of the experiment was also preceded by a practice block. Similar to the main phase, the practice block had two subblocks of 32 trials each. However, the color words were replaced with the stimulus “xxxx,” randomly presented eight times in each of the colors per subblock.

Procedure After completing the survey questions on pen and paper (see above), the main experiment began on the computer. Stimuli were presented on a white (255,255,255) screen in 18 pt., bold Courier New font. On each trial, participants were first presented with a fixation “+” in gray (128,128,128) for 250 ms. This was followed by a blank screen for 250 ms. Next, the colored word was presented until a response was registered or 2,000 ms elapsed. The next trial began immediately following a correct response. If the participant made an error or failed to Ó 2018 Hogrefe Publishing

Both mean correct response time and percentage error data were assessed for the computer portion of the task. For response times, only correct responses were considered, but no other trims were made. For error percentages, trials in which participants failed to respond before the trial ended were excluded (0.3% of trials). Raw data and participant mean are available in the Electronic Supplementary Materials, ESM 1–3.

Response Times The correct response time data are presented in Figure 3. To analyze response times, we conducted a language (Dutch vs. French) by congruency (identity vs. same response vs. different response) within-subjects repeated-measures analysis of variance (ANOVA). The main effect of congruency was significant, F(2, 184) = 22.840, MSE = 1,486, p < .001, ηp2 = .20. However, there was no main effect of language, F(1, 92) = 0.183, MSE = 1,525, p = .670, ηp2 < .01, indicating no overall difference in response speed to Dutch and French word trials. Most importantly, the interaction between language and congruency was not significant, F(2, 184) = 0.703, MSE = 1,239, p = .497, ηp2 < .01. Despite this lack of an interaction, we conducted planned comparisons on each language separately. For Dutch color words, there was both a significant stimulus conflict effect (same response – identity), t(92) = 2.409, SEdiff = 6, p = .018, η2 = .06, and response conflict effect (different response – same response), t(92) = 3.348, SEdiff = 5, p = .001, η2 = .11. Similarly for French color words, both stimulus conflict, t(92) = 2.322, SEdiff = 5, p = .022, η2 = .06, and response conflict, t(92) = 2.182, SEdiff = 5, p = .032, η2 = .05, were observed. There was no evidence for any differences in the magnitude of the stimulus conflict effect, t(92) = 0.319, SEdiff = 8, p = .750, η2 < .01, or response conflict effect, t(92) = 0.903, SEdiff = 7, p = .369, η2 < .01, across languages, though at least a numerical trend for smaller effects in French (particularly for response conflict).

Percentage Errors The percentage error data are presented in Figure 4. We again conducted a language (Dutch vs. French) by congruency (identity vs. same response vs. different response) within-subjects repeated-measures ANOVA. Experimental Psychology (2018), 65(1), 13–22

J. R. Schmidt et al., Bilingual Stroop Effect

620

identity

same response

different response

Response Time (ms)

610

600 590 580 570

560 550 Dutch (L1)

French (L2)

Language

Correlations

Figure 3. Response times with standard errors for Dutch and French color words.

identity

same response

different response

Percentage Error

η2 = .17. For French color words, there was neither a stimulus conflict effect, t(92) = 2.632, SEdiff = 0.8, p = .243, η2 = .07, nor a response conflict effect, t(92) = 1.309, SEdiff = 0.7, p = .194, η2 = .02. There was no evidence for any difference in the magnitude of the stimulus conflict effect across languages, t(92) = 0.132, SEdiff = 1.1, p = .895, η2 < .01. However, the response conflict effect was marginally larger in Dutch than in French, t(92) = 1.978, SEdiff = 1.0, p = .051, η2 = .04. Similar to the response times, then, there were some hints of larger effects for L1 color words, particularly for response conflict. However, despite the relatively large sample size, these differences were not sufficiently robust.

12 10 8

6 4

2 0 Dutch (L1)

French (L2)

Language Figure 4. Percentage errors with standard errors for Dutch and French color words.

The main effect of congruency was significant, F(2, 184) = 7.195, MSE = 24, p < .001, ηp2 = .07. However, there was no main effect of language, F(1, 92) = 0.036, MSE = 14, p = .850, ηp2 < .01, indicating no overall difference in error rates between Dutch and French word trials. Most importantly, the interaction between language and congruency was not significant, F(2, 184) = 2.285, MSE = 28, p = .105, ηp2 = .02, albeit with a trend toward a larger effect for Dutch. Again, we conducted planned comparisons on each language separately. For Dutch color words, there was no stimulus conflict effect (same response – identity), t(92) = 2.344, SEdiff = 0.8, p = .281, η2 = .06, but there was a significant response conflict effect (different response – same response), t(92) = 4.266, SEdiff = 0.7, p = .001, Experimental Psychology (2018), 65(1), 13–22

As a supplementary analysis, we consider how the measures of language level correlate with the stimulus and response conflict effects for both languages. Although the language measures correlate with each other well in the intuitive fashion (data available from the lead author on request), there was little evidence for a relationship between any of the language metrics with any of the observed congruency effects. The nonparametric Spearman’s ρ correlations are presented in Table 2 (results were similar with the parametric Pearson’s r). As can be observed, none of the performance (response time or error) measures correlated with years of French training, self-rated French level, or LEXTALE_FR score. With the LEAP-Q, percentage of Dutch and French language use also did not correlate with any performance measures. Age of (speaking) acquisition and fluency, and age of reading acquisition and fluency for both languages were not related to the response time or error effects after a Holm-Bonferroni correction for multiple comparisons. Without a correction, some correlations were significant at the α = .05 level, but this should, of course, be interpreted with caution (e.g., the largest correlation is between age of beginning to learn to read French and the stimulus conflict effect in Dutch, which seems difficult to interpret). More generally, the relative lack of strong correlations between the conflict effects and the language skill metrics is probably not surprising, given that there were little overall differences in the conflict effects across languages to begin with.

Discussion In the present report, we investigated for the first time the source of L2 (in addition to L1) congruency effects with a 2-to-1 keypress mapping procedure. Most importantly, the experiment revealed both stimulus and response Ó 2018 Hogrefe Publishing

J. R. Schmidt et al., Bilingual Stroop Effect

Table 2. Correlations with stimulus and response conflict effects Dutch

French

Stimulus RT

Response ERR

Stimulus ERR

Response ERR

ERR

LEXTALE_FR Years French

.007

.017

.153

.099

.029

.163

.137

.124

French Level

.099

.094

.022

.017

.018

.116

.117

.007

Score

.096

.036

.078

.161

.081

.176

.175

.138

%Dutch Use

.002

.090

.041

.052

.109

.060

.165

.021

%French Use

.025

.023

.046

.002

.011

.173

.062

.060

LEAP-Q

Dutch Acquisition

.059

.039

.085

.237

.045

.033

.185

.023

Fluent

.022

.065

.084

.283

.193

.140

.005

.039

Reading

.119

.061

.023

.265

.063

.076

.240

.039

Fluent Read

.146

.078

.059

.197

.084

.079

.083

.007

Acquisition

.044

.239

.102

.271

.024

.100

.079

.097

Fluent

.032

.210

.070

.110

.054

.101

.130

.031

Reading

.066

.321

.189

.175

.036

.032

.035

.002

Fluent Read

.112

.217

.116

.013

.114

.010

.176

.021

French

Notes. Italic = p < .05. No tests significant with Holm-Bonferroni correction.

conflict for a second language (French), just as with the first language (Dutch) in response times. This is contrary to the hypothesis that second language color words act as mere associates for the first language translations, as color associates (e.g., “sky”) produce stimulus conflict alone (Schmidt & Cheesman, 2005). That is, the results are not consistent with the notion that second language color words do not bias a potential response. Similarly, the current results are inconsistent with the notion that foreign language color words only influence response selection (e.g., because they are not strongly enough connected to semantics). That is, the results are not consistent with the notion that second language color words retrieve the response associated with the first language (e.g., via lexical translation), but do not activate semantics (or at least sufficiently to produce stimulus conflict). Such a notion would also assume that retrieval of a response can bypass semantics (i.e., that “noir” can retrieve the black response without a mediation through semantics), which may or may not be plausible. Also interesting, there were no sizeable differences in the observed congruency effects across languages, with the exception of some numerical trends and a marginally larger response conflict effect in errors for L1 color words. As previously discussed, past reports have observed smaller interference effects from a second language than from a first language. Indeed, Mägiste (1984, 1985) argued that the amount of conflict is proportional to mastery of a language (see also, Brauer, 1998). This may seem Ó 2018 Hogrefe Publishing

inconsistent with the present report, if not for a few added considerations. First, the asymmetry between first and second language congruency effects partially depends on the response language (Atalay & Misirlisoy, 2012; Dyer, 1971; Preston & Lambert, 1969; Tzelgov et al., 1990). For instance, with verbal Dutch responses, Dutch color-word interference would increase and French color-word interference would decrease. The reverse would be true with verbal French responses. In the current experiment, keypress responses were used, which are not inherently compatible with either language. Thus, any observed asymmetry should not be expected to be particularly large. Future research directly comparing keypress and verbal response modalities might test this notion more directly (though unfortunately the 2-to-1 mapping procedure cannot be used with verbal naming responses). There were some small hints of a larger congruency effect for L1 words in the current experiment, even marginally so in the error data (i.e., for response conflict). It might be supposed that Dutch-speaking participants subvocally name the colors in Dutch, making keypress responses more compatible with Dutch than French words. If this is true, however, it might also be surprising that an asymmetry with larger L1 conflict effects was not observed. On the other hand, the larger asymmetry in verbal experiments might have to do with the much stronger stimulus-response compatibility for color naming. Future research might investigate these possibilities more Experimental Psychology (2018), 65(1), 13–22

closely (e.g., by having participants subvocally name in Dutch vs. French). As one caveat with the present report, however, it is worth stressing that the present investigation made use of keypress responses, rather than verbal. This was necessary for the 2-to-1 mapping procedure. It remains possible, therefore, that additional differences between languages might be observable that are specific to the verbal response modality. For instance, during phonetic or articulatory planning, an L2 word like “jaune” might interfere less than an L1 word like “geel” because the former does not correspond to a potential response in the (Dutch-language) response set. Indeed, this could additionally explain the dependence of the asymmetry in L1 and L2 Stroop effects on the response language discussed earlier. Future research may therefore be directed at disentangling these issues further. For both the stimulus and response conflict effects, congruent trials were compared with incongruent trials. The conflict effects observed may therefore be in part due to incongruent-trial interference, and may in part be due to congruent-trial facilitation (Hasshim & Parris, 2014). Development of an appropriate neutral control condition (relative to which facilitation and interference can be measured), however, is a notoriously difficult task in Stroop and other research domains (Jonides & Mack, 1984; MacLeod, 1991). Future research may nevertheless aim to tease these subcomponents of the Stroop effect further apart in both L1 and L2 speakers. Above caveats aside, the present results suggest that L2 color words influence cognitive processing in much the same way as L1 color words. L2 color words produce both stimulus conflict and response conflict. Interestingly, this is even true for the non-cognates used in the present report. That is, foreign words (e.g., “marron”) that look quite different from the native language equivalent (e.g., “bruin”) automatically interfere with both stimulus and response selection. Future research might aim to investigate the extent to which the same is true for participants with considerably lower L2 language knowledge (e.g., no direct formal training). For instance, in the Kroll and Stewart (1994) model discussed above, it is assumed that the connections to semantics do increase with increasing proficiency. Despite the case that French proficiency in our sample was relatively weak, it could be the case that participants were proficient enough to induce stimulus conflict. On the other hand, others have argued for relatively early semantic mediation (e.g., Duyck & De Houwer, 2008). The approach to studying stimulus and response conflict in two languages novelly introduced here might therefore be extended further to very early language learning to help in discriminating between these competing ideas.

Experimental Psychology (2018), 65(1), 13–22

J. R. Schmidt et al., Bilingual Stroop Effect

Acknowledgments This research was supported by Grant BOF16/MET_V/002 of Ghent University to Jan De Houwer and by the Interuniversity Attraction Poles Program initiated by the Belgian Science Policy Office (IUAPVII/33). Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1618-3169/a000384 ESM 1. Data (xls) Raw data of the study. ESM 2. Tables and Figures (ods) Mean participants. ESM 3. Data (emrg) E-Prime data.

References Altarriba, J., & Mathis, K. M. (1997). Conceptual and lexical development in second language acquisition. Journal of Memory and Language, 36, 550–568. Atalay, N. B., & Misirlisoy, M. (2012). Can contingency learning alone account for item-specific control? Evidence from withinand between-language ISPC effects. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, 1578–1590. https://doi.org/10.1037/a0028458 Augustinova, M., & Ferrand, L. (2014). Social priming of dyslexia and reduction of the Stroop effect: What component of the Stroop effect is actually reduced? Cognition, 130, 442–454. https://doi.org/10.1016/j.cognition.2013.11.014 Augustinova, M., Silvert, L., Ferrand, L., Llorca, P. M., & Flaudias, V. (2015). Behavioral and electrophysiological investigation of semantic and response conflict in the Stroop task. Psychonomic Bulletin & Review, 22, 543–549. https://doi.org/10.3758/ s13423-014-0697-z Bialystok, E., Craik, F. I. M., & Luk, G. (2012). Bilingualism: Consequences for mind and brain. Trends in Cognitive Sciences, 16, 240–250. https://doi.org/10.1016/j.tics.2012.03.001 Brauer, M. (1998). Stroop interference in bilinguals: The role of similarity between the two languages. In A. F. Healy & L. E. Bourne (Eds.), Foreign language learning: Psycholinguistic studies on training and retention (pp. 317–337). Hillsdale, NJ: Erlbaum. Brysbaert, M. (2013). LEXTALE_FR: A fast, free, and efficient test to measure language proficiency in French. Psychologica Belgica, 53, 23–37. Chen, A. T., Bailey, K., Tiernan, B. N., & West, R. (2011). Neural correlates of stimulus and response interference in a 2–1 mapping Stroop task. International Journal of Psychophysiology, 80, 129–138. https://doi.org/10.1016/j.ijpsycho.2011. 02.012 Chen, H. C., & Ho, C. (1986). Development of Stroop interference in Chinese-English bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 397–401. Costa, A., Miozzo, M., & Caramazza, A. (1999). Lexical selection in bilinguals: Do words in the bilingual’s two lexicons compete for selection? Journal of Memory and Language, 41, 365–397.

Ó 2018 Hogrefe Publishing

J. R. Schmidt et al., Bilingual Stroop Effect

Dalrymple-Alford, E. C. (1968). Interlingual interference in a colornaming task. Psychonomic Science, 10, 215–216. De Houwer, J. (2003). On the role of stimulus-response and stimulus-stimulus compatibility in the Stroop effect. Memory & Cognition, 31, 353–359. De Houwer, J. (2004). Spatial Simon effects with nonspatial responses. Psychonomic Bulletin & Review, 11, 49–53. Duyck, W., & De Houwer, J. (2008). Semantic access in secondlanguage visual word processing: Evidence from the semantic Simon paradigm. Psychonomic Bulletin & Review, 15, 961–966. Dyer, F. N. (1971). Color-naming interference in monolinguals and bilinguals. Journal of Verbal Learning and Verbal Behavior, 10, 297–302. Ferrand, L., & Augustinova, M. (2014). Differential effects of viewing positions on standard versus semantic Stroop interference. Psychonomic Bulletin & Review, 21, 425–431. https://doi.org/10.3758/s13423-013-0507-z Glaser, W. R., & Glaser, M. O. (1989). Context effects in Stroop-like word and picture processing. Journal of Experimental Psychology: General, 118, 13–42. Hasshim, N., & Parris, B. A. (2014). Two-to-one color-response mapping and the presence of semantic conflict in the Stroop task. Frontiers in Psychology, 5, Article 1157. https://doi.org/ 10.3389/fpsyg.2014.01157 Hasshim, N., & Parris, B. A. (2015). Assessing stimulus-stimulus (semantic) conflict in the Stroop task using saccadic two-toone color response mapping and preresponse pupillary measures. Attention Perception & Psychophysics, 77, 2601–2610. Jongen, E. M. M., & Jonkman, L. M. (2008). The developmental pattern of stimulus and response interference in a color-object Stroop task: An ERP study. BMC Neuroscience, 9, 82. https://doi.org/10.1186/1471-2202-9-82 Jonides, J., & Mack, R. (1984). On the cost and benefit of cost and benefit. Psychological Bulletin, 96, 29–44. Klein, G. S. (1964). Semantic power measured through the interference of words with color-naming. American Journal of Psychology, 77, 576–588. Koda, K. (1996). L2 word recognition research: A critical review. Modern Language Journal, 80, 450–460. Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33, 149–174. La Heij, W., Debruyn, E., Elens, E., Hartsuiker, R., Helaha, D., & Vanschelven, L. (1990). Orthographic facilitation and categorical interference in a word-translation variant of the Stroop task. Canadian Journal of Psychology, 44, 76–83. Mackinnon, D. P., Geiselman, R. E., & Woodward, J. A. (1985). The effects of effort on Stroop interference. Acta Psychologica, 58, 225–235. MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychological Bulletin, 109, 163–203. Mägiste, E. (1984). Stroop tasks and dichotic translation: The development of interference patterns in bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 304–315. Mägiste, E. (1985). Development of intralingual and interlingual interference in bilinguals. Journal of Psycholinguistic Research, 14, 137–154. Marian, V., Blumenfeld, H. K., & Kaushanskaya, M. (2007). The Language Experience and Proficiency Questionnaire (LEAP-Q): Assessing language profiles in bilinguals and multilinguals. Journal of Speech Language and Hearing Research, 50, 940–967. https://doi.org/10.1044/1092-4388(2007/067)

Ó 2018 Hogrefe Publishing

Posner, M. I., & Presti, D. E. (1987). Selective attention and cognitive control. Trends in Neurosciences, 10, 13–17. Preston, M. S., & Lambert, W. E. (1969). Interlingual interference in a bilingual version of the Stroop color-word task. Journal of Verbal Learning and Verbal Behavior, 8, 295–301. Risko, E. F., Schmidt, J. R., & Besner, D. (2006). Filling a gap in the semantic gradient: Color associates and response set effects in the Stroop task. Psychonomic Bulletin & Review, 13, 310–315. Schmidt, J. R., & Cheesman, J. (2005). Dissociating stimulusstimulus and response-response effects in the Stroop task. Canadian Journal of Experimental Psychology, 59, 132–138. Schmidt, J. R., Cheesman, J., & Besner, D. (2013). You can’t Stroop a lexical decision: Is semantic processing fundamentally facilitative? Canadian Journal of Experimental Psychology, 67, 130–139. https://doi.org/10.1037/a0030355 Schmidt, J. R., Crump, M. J. C., Cheesman, J., & Besner, D. (2007). Contingency learning without awareness: Evidence for implicit control. Consciousness and Cognition, 16, 421–435. Schmidt, J. R., & De Houwer, J. (2012). Contingency learning with evaluative stimuli: Testing the generality of contingency learning in a performance paradigm. Experimental Psychology, 59, 175–182. Simon, J. R., Craft, J. L., & Webster, J. B. (1973). Reactions toward stimulus source: Analysis of correct responses and errors over a five-day period. Journal of Experimental Psychology, 101, 175–178. Simon, J. R., & Rudell, A. P. (1967). Auditory S-R compatibility: Effect of an irrelevant cue on information processing. Journal of Applied Psychology, 51, 300–304. Smith, M. C., & Kirsner, K. (1982). Language and orthography as irrelevant features in color-word and picture-word Stroop interference. The Quarterly Journal of Experimental Psychology, 34A, 153–170. Stirling, N. (1979). Stroop interference: An input and an output phenomenon. The Quarterly Journal of Experimental Psychology, 31, 121–132. Stroop, J. R. (1935). Studies on interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–661. Tzelgov, J., Henik, A., & Leiser, D. (1990). Controlling Stroop interference: Evidence from a bilingual task. Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 760–771. van Veen, V., & Carter, C. S. (2005). Separating semantic conflict and response conflict in the Stroop task: A functional MRI study. Neuroimage, 27, 497–504. Werker, J. F., & Byers-Heinlein, K. (2008). Bilingualism in infancy: First steps in perception and comprehension. Trends in Cognitive Sciences, 12, 144–151. https://doi.org/10.1016/ j.tics.2008.01.008

Received June 9, 2017 Revision received August 1, 2017 Accepted August 1, 2017 Published online February 8, 2018

James R. Schmidt Ghent University Henri Dunantlaan 2 9000 Ghent Belgium james.schmidt@ugent.be

Experimental Psychology (2018), 65(1), 13–22

Appendix Language Demographics There was relatively little variability in the language demographics of the participants. The vast majority of participants rated their order of language acquisition as Dutch, followed by French (followed by English), but rated their language dominance as Dutch, followed by English, followed by French in third, and indicated Dutch as their native language. Most critically, all but two participants ranked Dutch as their dominant language. One of the remaining two indicated that both Dutch and Turkish were joint dominant languages, and the other indicated Russian as the dominant language. Both, however, ranked Dutch as second and French as fourth. One other participant rated both Dutch and French as equally dominant (though with Spanish and English as native but not dominant languages), but the results on the LEXTALE_FR did not support this. Though this participant had the highest score in the sample (tied with another) of 32, this only corresponds to the 4th percentile for native French speakers. Additionally, this participant indicated that they used Dutch much more frequently (75%) than French (5%). Another two participants indicated French as a joint first language (one with Dutch and the other with Turkish) who were also among the higher scorers on the LEXTALE_FR (27 and 32, respectively). However, none of these participants rated French as their dominant language and, as already mentioned, none of the participants in the sample had particularly

J. R. Schmidt et al., Bilingual Stroop Effect

convincing LEXTALE_FR scores. On average, participants self-rated their French fluency at 6.3 on a 0–10 scale. In contrast, the average LEXTALE_FR score was only 5.8 (out of a maximum score of 56), corresponding to the 1st percentile for L1 users and 48th percentile for L2 users. All participants gave a higher percentage for Dutchlanguage use than for French, with one exception. One participant rated Dutch and French both 50%, though, strangely, this participant rated French as their third most dominant language, self-rated their French knowledge as 6 (average), and only scored 1 on the LEXTALE_FR. All in all, the participants in the sample seemed to fit the basic requirements of dominant Dutch and less dominant French. Thus, no participants were trimmed. Trimming of the potentially problematic participants mentioned above had only negligible impact on the results reported below. The vast majority of participants correctly translated “noir(e)” (90/93), “vert(e)” (88/93), and “jaune” (83/93). For “brown,” most participants indicated “brun(e)” (66/ 93), with only a few indicating “marron” (7/93, two of which wrote both “marron” and “brun”). “Brun” is more similar to the Dutch “bruin” (and English “brown”), but is only a partially correct translation for the color brown.3 Any incorrect answers (including misspellings) were pointed out to the participant before starting the experiment, including the semi-mistranslations of “brown.” Note that though very few participants indicated “marron,” most participants did seem to recognize this word as soon as it was presented to them by the experimenter (this was not tested systematically, however).

“Brun” is primarily used for hair colors, and derivatives (e.g., “ours brun” [brown bear], which has brown hair, or “bière brune” [brown beer], which also relates to hair in the same as was “bière blonde” [blonde beer]). “Brun” can also be used to refer to specific shades or in an “artistic” context (e.g., like referring to blue as “azure” in English). “Marron” is the more standard color name for brown.

Experimental Psychology (2018), 65(1), 13–22

Ó 2018 Hogrefe Publishing

Short Research Article

A Bird in the Hand Isn’t Good for Long Action Dynamics Reveal Short-Term Choice Impulses in Intertemporal Choices Stefan Scherbaum,1 Simon Frisch,1 and Maja Dshemuchadse2 1

Department of Psychology, Technische Universität Dresden, Germany

Faculty of Social Sciences, Hochschule Zittau/Görlitz, Germany

Abstract: Folk wisdom tells us that additional time to make a decision helps us to refrain from the first impulse to take the bird in the hand. However, the question why the time to decide plays an important role is still unanswered. Here we distinguish two explanations, one based on a bias in value accumulation that has to be overcome with time, the other based on cognitive control processes that need time to set in. In an intertemporal decision task, we use mouse tracking to study participants’ responses to options’ values and delays which were presented sequentially. We find that the information about options’ delays does indeed lead to an immediate bias that is controlled afterwards, matching the prediction of control processes needed to counter initial impulses. Hence, by using a dynamic measure, we provide insight into the processes underlying short-term oriented choices in intertemporal decision making. Keywords: process, dynamics, delay discounting, intertemporal choice, decision making, mouse tracking

Many significant decisions in our lives are intertemporal ones where we have to choose between smaller but sooner and larger but later gains, for instance, when we have to decide whether we save our money for pension or buy the most recent smartphone instead. When making such decisions, people are usually advised to resist first impulses and sleep on them for a night. The wisdom behind this strategy is that short-sighted behavior gives place to longterm oriented behavior when one resists first impulses long enough. Why is that? On the one hand, a number of theories on impulsiveness in intertemporal decisions assume that the accumulation of information about the two choice alternatives might be biased such that it unduly weighs both options’ delays resulting in suboptimal, often short-sighted decisions (cf. Kim & Lee, 2011). From this perspective, waiting for a longer time would support the accumulation of evidence about the options’ values which would drive the system toward a more value- instead of delay-oriented decision. However, other theorists in the field argue that information accumulation alone does not provide an exhaustive description of the processes leading to an intertemporal decision. Instead, these authors claim that additional control processes are necessary to protect us from short-sighted, Ó 2018 Hogrefe Publishing

unduly time-oriented decisions (Figner et al., 2010; Hare, Camerer, & Rangel, 2009; Harris, Hare, & Rangel, 2013; Peters & Büchel, 2011). As these control processes need time to engage, this perspective would argue that biding one’s time can improve decision making because it gives control processes time to suppress the initial behavioral impulse to respond to the delays of the options and hence seize the earliest opportunity. Here, we provide evidence for the latter explanation by tracing participants’ mouse movements in a decision task. In intertemporal decisions, the question when a reward is delivered is a crucial one, as human beings discount the objective values of the options by their delay. In the laboratory, this delay discounting effect is studied by having participants choose between different variations of sooner but smaller (SS) and later but larger (LL) rewards. In these tasks, participants often violate assumptions of normative decision theories (Samuelson, 1937) by choosing the smaller sooner option more often than these theories would predict – especially when it is offered without any delay (e.g., Ainslie, 1975; Laibson, 1997). Recent process models that conceive of intertemporal decisions as a process of sequential information sampling Experimental Psychology (2018), 65(1), 23–31 https://doi.org/10.1027/1618-3169/a000385

(Dai & Busemeyer, 2014; Gold & Shadlen, 2007; Rodriguez, Turner, & McClure, 2014) provide an explanation for this behavior: these models are based on the idea that a decision maker accumulates evidence in favor of each of the two options and that an option is chosen as soon as its amount of accumulated evidence exceeds a threshold. In such models, short-sighted behavior results from an accumulation process that overweighs the delay information relative to the value information. Such overweighing might result, for example, from an initial bias to choose the SS option or from a faster accumulation of the delay information (favoring the SS option) compared to the value information (favoring the LL option; Scherbaum, Dshemuchadse, & Goschke, 2012). In both cases, the answer sequential-sampling models provide to the question why we should wait when making decisions is that the passage of time prolongs the accumulation of evidence for both options which results in a more accurate representation of their overall attractiveness and, hence, prevents premature decisions that are primarily influenced by the options’ delays. While this explanation seems to provide a parsimonious explanation for short-sighted behavior, it is not unchallenged. For example, researchers in the field of self-control (Hare et al., 2009; Harris et al., 2013; Peters & Büchel, 2011) would argue that prolonged information accumulation alone does not suffice to explain why the passage of time helps to overcome impulsivity. Instead, they assume that the cognitive system incorporates additional control processes which prevent impulsive choices by actively suppressing the impact of the delay information on the decision process. These control processes need cognitive resources and time to kick in: it has been shown that limiting cognitive resources increases the rate of short-sighted decisions in healthy subjects (Deck & Jahedi, 2015; Hinson, Jameson, & Whitney, 2003; but see Kurth-Nelson, Bickel, & Redish, 2012) and a number of clinical studies demonstrate that patients suffering from mental diseases associated with low self-control (e.g., addiction or attention deficit hyperactivity disorder) show an increased rate of short-sighted decisions compared to healthy participants (e.g., Bickel & Marsch, 2001; Kräplin et al., 2014; Wittmann & Paulus, 2007). Accordingly, the control explanation states that the passage of time allows control processes to set in which, then, suppress the initial impulse to make short-sighted decisions that are primarily driven by the options’ delays. In order to distinguish the two explanations – the sequential-sampling explanation and the control explanation – we constructed an intertemporal choice task in which we manipulated the order of presentation for the delay and value information systematically. More precisely, our paradigm consisted of blocks of decisions across which Experimental Psychology (2018), 65(1), 23–31

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

one source of information (e.g., the options’ delays) was held constant and was hence known in advance while the other source of information (e.g., the options’ values) was varied from decision to decision and hence became known only at the beginning of each trial. We expected these differences in the time course of decision making to distinctively influence the final choice of participants. According to the sequential-sampling explanation, short-sighted choices are the result of a biased process of information accumulation. Hence, showing options’ delays first should induce a bias toward the SS option (favored by the delay information) compared to when showing the options’ values first. Hence, showing options’ delays first leads to more SS choices. In contrast, the control explanation attributes short-sighted choices to an initial behavioral impulse – a strong response to the options’ delays – that is overruled and suppressed later through control processes. As a consequence, showing the options’ delays first should initially induce a bias toward the SS option. However, this bias is then overcome by control processes: since the options’ delays are kept constant across subsequent trials and because a fair amount of time passes before the value information is presented, control processes should have sufficient time to set in and suppress this initial bias to the SS option. Hence, showing options’ delays first should – in the end – lead to more LL choices than showing the value information first. Crucially, the two explanations make also diverging predictions with regard to the time course of the decision-making process. Hence, studying the time course will provide more direct evidence for one or the other explanation. To study the time course of intertemporal decision making, we recently used mouse movements (Dshemuchadse, Scherbaum, & Goschke, 2012) and showed that analyzing the temporal patterns of influence for different sources of information can provide additional insight into the underlying decision processes. Considering differences in this time course of decision making in our paradigm, the sequential-sampling explanation predicts a bias by the information shown first and hence, mouse movements should also show this bias – irrespective of whether the information presented first were values or delays. In contrast, the control explanation predicts a bias induced by the value information when it is shown first, but no such bias for the delay information when it is shown first. This is because the initial impulsive influence of the options’ delays should already be suppressed by control processes when the actual decision process starts in each choice trial. Inversely however, the control explanation predicts a strong response to the delay information when it is presented second, because control processes then need time to kick in and suppress the delay information. Ó 2018 Hogrefe Publishing

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

Taken together, this study aims to distinguish two explanations of why the passage of time leads to less short-sighted intertemporal decisions: while the sequential-sampling explanation predicts (1) more discounting when the options’ delays are presented first and (2) a constant influence on choice movements by the information that is shown first, the control explanation predicts (1) less discounting when the options’ delays are presented first and (2) a lack of influence on choice movements by the options’ delays (but not the options’ values) when they are shown first as well as a strong response to the options’ delays (but not the options’ values) when they are shown second.

Methods Participants Twenty-four right-handed students (23 female, Mage = 23.54 years) of the Technische Universität Dresden took part in the experiment. Participants gave informed consent to the study and received class credit or 5.00 € payment.

Apparatus and Stimuli Stimuli were presented in white on a black background on a 1700 CRT screen running at a resolution of 1,280 1,024 pixels (75 Hz refresh frequency). The decision options were presented in the vertical center of the left and the right half of the screen. As targets for mouse movements, response boxes were presented at the top left and top right of the screen. We used Psychophysics Toolbox 3 (Brainard, 1997; Pelli, 1997) in Matlab 2006b (Mathworks Inc., Natick, MA, USA) as presentation software, running on a Windows XP SP2 personal computer. Participants performed their responses with a standard computer mouse (Logitech Wheel Mouse USB). The mouse speed was reduced to ¼ in the systems settings and nonlinear acceleration was switched off. Mouse movement trajectories were sampled with a frequency of 100 Hz and recorded from presentation of the complete choice options (including both, times and values) until the cursor reached the response boxes and the trial ended.

Trials were grouped into mini-blocks of 14 trials (see Figure 1). In the time first condition, options’ delays remained constant within a mini-block and values were varied trial-wise; in the value first condition, options’ values remained constant and times were varied trial-wise. At the start of each mini-block, the constant information was presented for 5 s (see Figure 1). Each trial consisted of three stages. Stage 1: participants had to click into a red box at the bottom of the screen (within a time limit of 2 s) to provide comparable starting positions across trials. Stage 2: two response boxes at the right and left upper corner of the screen were presented and participants had to start the mouse movement upwards within a time limit of 2 s. The movement criterion was fulfilled when participants moved at least 4 pixels in each of two consecutive time steps. We chose to force participants to be already in motion before the information relevant for this trial appeared to ensure that they did not decide first and then only executed the final movement (Scherbaum, Dshemuchadse, Fischer, & Goschke, 2010). Stage 3: after starting their movement, the trial-wise information was presented and participants had to choose one of the two options. The trial ended after moving the cursor into one of the response boxes or after a time limit of 2.5 s. The assignment of SS or LL options to the left or right response box was balanced across participants. If participants missed any time limit of one of the three stages, the current trial was aborted and the next trial started automatically with the presentation of the start box (see Figure 1).

Design We orthogonally varied the delay of the SS option (0 and 10 days), the time interval between the options (1, 7, 13, 19, 25, 31, and 37 days – called intervals in the following), the value of the SS option in percent of the value of the LL option (5, 15, 30, 50, 70, 85, and 95% – called differences in the following), and the value of the LL option (20.- and 40.-Euro). The conditions time first and value first were varied between blocks, with their order balanced across participants. Altogether, this yielded 2 blocks (time first, value first) and 196 trials (14 mini-blocks) per block. The order of mini-blocks as well as the order of trials within a mini-block was randomized.

Procedure Participants were asked to decide on each trial which of two options they preferred: a sooner/smaller (SS) or a later/larger (LL) option. Participants were instructed to respond to the hypothetical choices as if they were real choices (cf. Lagorio & Madden, 2005). Ó 2018 Hogrefe Publishing

Data Processing Trials missing a deadline (3.5%) and trials with response times below 0.3 s (6.06%) were excluded as outliers. Mouse movements were aligned for a common starting position (horizontal midline) and each movement was Experimental Psychology (2018), 65(1), 23–31

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

Figure 1. Experimental setup. At the beginning of each mini-block of 14 trials, the first information (depending on condition value first or time first) was presented. To start a trial, participants had to click into a start box. After starting the movement upwards, the second information was presented and participants had to choose between options by moving the mouse into one of the response boxes at the top of the screen.

time-normalized to 100 equal time slices for statistical comparability.

Results We first analyzed the results for the final choice. The sequential-sampling explanation predicted more discounting when options’ delays were presented first, while the control explanation predicted more discounting when the options’ delays were presented second. To test these predictions, we calculated the probability of choosing the LL option for the two conditions time first and value first. A t-test yielded a significant difference t(23) = 4.07, p < .01, d = 0.83, 95% CI [0.06, 0.16]. As predicted by the control explanation, the value first condition showed more discounting [P(LL) = 0.61, SE = 0.034] than the time first condition [P(LL) = 0.72, SE = 0.031; see Figure 2]. Next, we analyzed the dynamics of the decision process and hence the timing of the different sources of influence on mouse movements (for basic mouse data, i.e., raw mouse movements, area under the curve, and maximum deviation, please see the Electronic Supplementary Material, ESM 1). The sequential-sampling explanation predicted the information shown first to exert an initial bias on choice movements (irrespective of delay or values shown first), while the control explanation predicted a bias only for the value information. To test these predictions, we applied continuous multiple-regression analyses to the 1

Figure 2. Probability to choose the LL option as a function of condition and intervals between the SS and the LL option. Error bars mark standard errors.

mouse movements (Scherbaum et al., 2010) in a four-step procedure. First, we calculated the movement angle relative to the Y-axis for each time step of a movement as our dependent measure.1 The movement angle was smoothed by a Gaussian of 10 time steps.2 Second, we coded two binary predictors (as 1 and 1 for comparable beta-weights) for all trials and each participant. For the time of delivery, the predictor intervals coded if the interval between SS and LL was larger than the median; for the reward value, the predictor differences coded if the difference between

The movement angle reflects the instantaneous tendency of the movement more precisely as it integrates the movement on the X/Y plane into a single measure. We apply temporal smoothing for three reasons (similar to spatial smoothing in functional magnetic resonance imaging [fMRI] analysis; e.g., Mikl et al., 2008): first, to increase signal-to-noise ratio. Second, to improve the validity of statistical tests (increased normality of error distributions). Third, to accommodate slight temporal variations between subjects.

Experimental Psychology (2018), 65(1), 23–31

Ó 2018 Hogrefe Publishing

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

(A)

Figure 3. Results of continuous regression analysis on mouse movements. (A) The time first condition. (B) the value first condition. (C) The contrast value first – time first indicates an advantage for time information (intervals, red, dashed lines) only when presented later in the trial, but not when presented first, in contrast to the value information (differences, blue, dasheddotted lines).

(C)

(B)

SS and LL was larger than the median. Third, we computed multiple regressions with these predictors on the movement angle for each time slice (100 time slices, hence 100 multiple regressions), yielding two time-varying betaweights (2 weights 100 time slices) for each participant. We tested each time point of beta-weights with t-tests against zero (which would represent no influence, compare Dshemuchadse et al., 2012). To reduce the error by multiple comparisons, we only accepted significant segments of more than 10 consecutive time steps (see Dale, Kehoe, & Spivey, 2007; Scherbaum, Gottschalk, Dshemuchadse, & Fischer, 2015 for Monte Carlo analyses on this issue). According to the sequential sampling explanation, we expected the intervals to show a constant bias in the time first condition, and the differences to show a constant bias in the value first condition. In contrast, according to the control explanation, we expected such a bias only for differences in the value first condition, and no bias in the time first condition. The results for the time first condition (Figure 3A) indicated that intervals showed an earlier influence (time steps 22–99) than differences (time steps 47–99). In contrast, the results for the value first condition (Figure 3B) indicated that differences showed an immediate influence (time steps 1–99) in contrast to intervals (time steps 59–99). Thus, the information presented first influenced the initial movements differently. The strength of the initial influence of intervals in the time first condition was less reliable compared to the initial influence of differences in the value first condition. In contrast, later in the course of the trial, intervals showed a stronger peak in influence in the value first compared to the time first condition. To statistically check for the described differences between conditions, we calculated the contrast value first – time first for both beta-weights. Again, we tested each time point of the contrasted beta-weights with t-tests Ó 2018 Hogrefe Publishing

against zero with a significant criterion of 10 consecutive time steps. The contrast supported our observation of an initial influence of differences in the value first condition compared to the time first condition (significant contrast for differences, time steps 5–59; see Figure 3C) and the observation of a much weaker initial influence of the intervals in the time first condition compared to the value first condition, but a stronger influence of intervals at the end of the trial in the value first condition compared to the time first condition (significant contrast for intervals, time steps 73–86; see Figure 3C). This difference in peaks was not found for differences. Taken together, this indicates two points. First, the options’ values bias the mouse movement from the start on when values are available from the beginning. This initial bias is not observed for options’ delays when delays are available from the beginning. Second, the delays show a stronger effect on movement trajectories when delays are presented second. This late response is not present for the options’ values when presented as second information. The results hence match the predictions of the control explanation, but not those of the sequential-sampling explanation.

Discussion Our aim was to distinguish two explanations for why the passage of time supports making less short-term oriented choices in intertemporal decision making. One theoretical perspective – the sequential-sampling explanation – states that the more time one has, the more information is accumulated supporting a more well-informed decision. A different theoretical perspective – the control explanation – states that an initial tendency to respond to the options’ Experimental Psychology (2018), 65(1), 23–31

delays induces an immediate impulse to take the sooner option. Only when enough time is given, this impulse can be countered by control processes which, then, open up the way for a more long-term oriented decision. To distinguish the two explanations, we used a novel variant of a standard intertemporal choice paradigm where we combined the sequential presentation of choice-relevant information (i.e., options’ delays and values) with mouse tracking. When options’ values were presented first, we found stronger discounting at the level of choice outcomes. At the level of mouse movements, we observed a moderate yet persistent bias of the value information (which was presented first) as well as a strong response to the delay information (which was timed to its onset later in the trial). In contrast, when options’ delays were presented first, we found less discounting at the level of choices. Importantly, we neither observed an initial bias by the delay information (which was presented first) nor a strong response to the presentation of the value information (which was presented later in the trial). These results support the predictions resulting from the control explanation which assumes that short-sighted choices result from an initial impulse to respond to the options’ delays which is subsequently suppressed by control processes. Our interpretation of the results reflecting the influence of control integrates well into research showing that cognitive control processes also need time to kick in and are also impaired in patients suffering from a lack of control (Ridderinkhof, van den Wildenberg, Wijnen, & Burle, 2004), similar to intertemporal choices (Bickel & Marsch, 2001; Kräplin et al., 2014, 2015). Our findings add important new insights for a number of process-oriented models of intertemporal choice that have gained popularity in recent years. While most of these models assume a linear (Dai & Busemeyer, 2014; Rodriguez et al., 2014) or at least a monotonically rising (Scherbaum et al., 2012) accumulation process over the course of the decision, the temporal continuity of the mouse-tracking methodology allows to investigate these assumptions directly. With regard to value information, our data confirmed a monotonically rising information accumulation process which was reflected in an initial and prevailing effect on decision trajectories when values were presented first and a slowly rising effect when values were presented second in our paradigm. The information on delays, however, shows a different pattern: instead of rising monotonically, it seems to accumulate quickly when it is presented second but its influence also vanishes rapidly as soon as it has passed a peak. This may also be the reason why we observed only a weak influence of the delay information when it remained constant across trials in our paradigm (i.e., when it was presented first). Again, this pattern of results can be explained by an undue impulse Experimental Psychology (2018), 65(1), 23–31

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

to respond to the options’ delays which can be corrected, however, when the cognitive system is provided with sufficient time. All in all, this evidence strengthens earlier findings which indicate that the delay information may be processed differently than the value information in intertemporal decision making (Scherbaum et al., 2012). One might ask whether the missing influence of the delay information might be caused by a lack of power, since delays show weaker beta-weights compared to values across all conditions. We think that two points argue against such an interpretation. First, if we look at the strongest peaks of delays (β = 0.191) and values (β = 0.373), we can see that values are about twice as strong as delays. However, looking at the interval in which delays first show a significant influence, we find a beta-weight (β = 0.011) that is less than a third of the beta-weight for values (β = 0.035). Hence, delays show a numerically weaker influence which is not only a matter of power and significance thresholds. Second, and in line with the first point, we calculated the power of our study for the prediction of sequential-sampling models: Time and values should follow the same time course in principle and hence we should have found an influence for delays in the same magnitude of values (although corrected by the overall strength, delay/value = 0.5). Hence, we took the beta for values at the first point of significance for the delay influence (β = 0.035) and calculated the beta for the delay influence that could have been expected under this assumption. For the resulting β of 0.017, we calculated a power of 0.91 (Using G-Power 3, Faul, Erdfelder, Lang, & Buchner, 2007). Hence, under the assumption of sequential-sampling models, but considering a general asymmetry between the influence of values and delay, we should have found significant betas for the delays with a sufficient power of 90% (Cohen, 1988). Considering other approaches to delay discounting, the findings of this study could also be interpreted within the framework of heuristics in decision making as it has been recently applied to intertemporal choices (Ericson, White, Laibson, & Cohen, 2015). Similar to these approaches, we focused on the psychological mechanisms underlying a decision instead of economic theory (Kahneman, Slovic, & Tversky, 1982). However, while models based on heuristics parsimoniously explain fundamental findings in intertemporal choice (e.g., Scholten & Read, 2010), they are limited to choice outcomes – similar to classical (i.e., static) discounting models. Hence, our study adds new information to heuristics-based models by focusing on the dynamics of the mechanisms underlying the choice process on the micro-level that most heuristics only describe on the macro-level (e.g., the dynamics of calculating and using weighted average or percentage differences in intertemporal decisions, Ericson et al., 2015). Ó 2018 Hogrefe Publishing

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

Concerning the underlying brain systems, our findings could easily be integrated into dual systems theories (McClure, Laibson, Loewenstein, & Cohen, 2004; e.g., Metcalfe & Mischel, 1999) by describing the temporal dynamics of the interaction between an initially delayresponsive system and a control system. However, the results leave open whether control processes inhibit a delay-responsive impulsive system or simply modulate the (initially delay-responsive) value signal in a common valuation system (Harris et al., 2013; Peters & Büchel, 2011). In the latter interpretation, our results are also compatible with more holistic theories of brain systems (Ballard & Knutson, 2009; Kable & Glimcher, 2007). By focusing on the dynamics of the decision process, we were able to provide insights into the processes leading to short-sighted choices. This showcases how a dynamic approach can complement information from outcomebased approaches, for example, classical intertemporal choice studies (for a review, see Frederick, Loewenstein, & O’Donoghue, 2002), or structurally oriented (neuroimaging) approaches with lower temporal resolution (e.g., McClure et al., 2004). However, various outcome-based and step-wise process theories of intertemporal choice (Killeen, 2009; Loomes, 2010; e.g., Scholten & Read, 2006; Stewart, Chater, & Brown, 2006) make predictions how the different properties of choice options may be compared by the cognitive system in order to derive a decision. By manipulating the order of information presentation our experimental manipulation may have implied specific strategies to process the time and value information in a certain order, which has been shown to influence intertemporal decisions (Zauberman, Kim, Malkoc, & Bettman, 2009). On the one hand, these experimental constraints in the availability of information mean that our results about the decision process are limited with respect to distinguishing between different outcome-based theories (e.g., with respect to attribute-wise or option-wise comparisons). It could be possible that our procedure, presenting options attribute-wise, suggested a certain decision strategy. However, these constraints also underline the dire necessity to take the context of intertemporal choices – like the order in which choice-relevant information is presented – into account if we wish to reach a true and deep understanding of the decision process (Lempert & Phelps, 2016). From a methodological point of view, mouse movements are a promising tool to study process dynamics of decision making (cf. O’Hora, Dale, Piiroinen, & Connolly, 2013 for a recent example of advanced analyses). Nevertheless, one might object that they capture the perceptual and cognitive effects on the decision process we are interested in only indirectly as they are, in the end, a purely motor-based measure. However, recent applications of the mouse Ó 2018 Hogrefe Publishing

tracking methodology to intertemporal choices (Dshemuchadse et al., 2012), delay discounting decisions (Scherbaum, Dshemuchadse, Leiberg, & Goschke, 2013), impulsive decisions (Travers, Rolison, & Feeney, 2016), semantic and numerical processing (Dale et al., 2007; Dshemuchadse, Grage, & Scherbaum, 2015; Spivey, Grosjean, & Knoblich, 2005), and cognitive control processes (Frisch, Dshemuchadse, Görner, Goschke, & Scherbaum, 2015; Scherbaum et al., 2010, 2015) impressively show that mouse tracking is sensitive to a variety of cognitive processes and contextual influences. Nevertheless, other dynamic tools, for example, frequency tagged EEG (e.g., Müller, Andersen, & Keil, 2007) or eye tracking (Franco-Watkins, Mattson, & Jackson, 2015) will contribute to a more complete picture of process dynamics in decision making, as they have done in other related fields like the study of cognitive control processes (Scherbaum, Fischer, Dshemuchadse, & Goschke, 2011). The folk wisdom that the passage of time leads to better decisions is well known. At least for intertemporal decisions, it seems that taking one’s time can help to resist the first impulse to take the bird in the hand – but to wait for the two birds in the bush instead.

Acknowledgments This research was partly supported by the German Research Council (DFG) (Grant SFB 940/1 2012). We thank Romy Schneider, Marta Kristlib, and Luise Ristein for support in collecting the data. Primary data and analysis functions can be found at https://osf.io/rxe73. Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1618-3169/a000385 ESM 1. Figures (.pdf) Mouse movements as a function of time, condition, and choice.

References Ainslie, G. (1975). Specious reward: A behavioral theory of impulsiveness and impulse control. Psychological Bulletin, 82, 463–495. https://doi.org/10.1037/h0076860 Ballard, K., & Knutson, B. (2009). Dissociable neural representations of future reward magnitude and delay during temporal discounting. NeuroImage, 45, 143–150. https://doi.org/ 10.1016/j.neuroimage.2008.11.004 Bickel, W. K., & Marsch, L. A. (2001). Toward a behavioral economic understanding of drug dependence: Delay discounting processes. Addiction, 96, 73–86. https://doi.org/10.1046/ j.1360-0443.2001.961736.x

Experimental Psychology (2018), 65(1), 23–31

Brainard, D. H. (1997). The psychophysics toolbox. Spatial Vision, 10, 433–436. https://doi.org/10.1163/156856897X00357 Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (Vol. 2, 2nd Rev ed.). Hillsdale, NJ: Erlbaum. Dai, J., & Busemeyer, J. R. (2014). A probabilistic, dynamic, and attribute-wise model of intertemporal choice. Journal of Experimental Psychology: General, 143, 1489–1514. https:// doi.org/10.1037/a0035976 Dale, R., Kehoe, C., & Spivey, M. J. (2007). Graded motor responses in the time course of categorizing atypical exemplars. Memory and Cognition, 35, 15–28. https://doi.org/10.3758/BF03195938 Deck, C., & Jahedi, S. (2015). The effect of cognitive load on economic decision making: A survey and new experiments. European Economic Review, 78, 97–119. https://doi.org/ 10.1016/j.euroecorev.2015.05.004 Dshemuchadse, M., Grage, T., & Scherbaum, S. (2015). Action dynamics reveal two components of cognitive flexibility in a homonym relatedness judgment task. Frontiers in Cognition, 6, 1244. https://doi.org/10.3389/fpsyg.2015.01244 Dshemuchadse, M., Scherbaum, S., & Goschke, T. (2012). How decisions emerge: Action dynamics in intertemporal decision making. Journal of Experimental Psychology: General, 142, 151–185. https://doi.org/10.1037/a0028499 Ericson, K. M., White, J., Laibson, D., & Cohen, J. (2015). Money earlier or later? Simple heuristics explain intertemporal choices better than delay discounting does. Psychological Science, 26, 826–833. https://doi.org/10.1177/0956797615572232 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 Figner, B., Knoch, D., Johnson, E. J., Krosch, A. R., Lisanby, S. H., Fehr, E., & Weber, E. U. (2010). Lateral prefrontal cortex and self-control in intertemporal choice. Nature Neuroscience, 13, 538–539. https://doi.org/10.1038/nn.2516 Franco-Watkins, A. M., Mattson, R. E., & Jackson, M. D. (2015). Now or later? Attentional processing and intertemporal choice. Journal of Behavioral Decision Making, 29, 206–217. https:// doi.org/10.1002/bdm.1895 Frederick, S., Loewenstein, G., & O’Donoghue, T. (2002). Time discounting and time preference: A critical review. Journal of Economic Literature, 40, 351–401. https://doi.org/10.1257/jel. 40.2.351 Frisch, S., Dshemuchadse, M., Görner, M., Goschke, T., & Scherbaum, S. (2015). Unraveling the sub-processes of selective attention: Insights from dynamic modeling and continuous behavior. Cognitive Processing, 16, 377–388. https://doi.org/ 10.1007/s10339-015-0666-0 Gold, J. I., & Shadlen, M. N. (2007). The neural basis of decision making. Annual Review of Neuroscience, 30, 535–574. https://doi.org/10.1146/annurev.neuro.29.051605.113038 Hare, T. A., Camerer, C. F., & Rangel, A. (2009). Self-control in decision-making involves modulation of the vmPFC valuation system. Science, 324, 646–648. https://doi.org/10.1126/science. 1168450 Harris, A., Hare, T., & Rangel, A. (2013). Temporally dissociable mechanisms of self-control: Early attentional filtering versus late value modulation. The Journal of Neuroscience, 33, 18917–18931. https://doi.org/10.1523/JNEUROSCI.5816-12.2013 Hinson, J. M., Jameson, T. L., & Whitney, P. (2003). Impulsive decision making and working memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 298–306. https://doi.org/10.1037/0278-7393.29.2.298 Kable, J. W., & Glimcher, P. W. (2007). The neural correlates of subjective value during intertemporal choice. Nature Neuroscience, 10, 1625–1633. https://doi.org/10.1038/nn2007

Experimental Psychology (2018), 65(1), 23–31

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

Kahneman, D., Slovic, P., & Tversky, A. (1982). Judgment under uncertainty: Heuristics and biases. Cambridge, UK: Cambridge University Press. Killeen, P. R. (2009). An additive-utility model of delay discounting. Psychological Review, 116, 602–619. https://doi.org/10.1037/ a0016414 Kim, S., & Lee, D. (2011). Prefrontal cortex and impulsive decision making. Biological Psychiatry, 69, 1140–1146. https://doi.org/ 10.1016/j.biopsych.2010.07.005 Kräplin, A., Behrendt, S., Scherbaum, S., Dshemuchadse, M., Bühringer, G., & Goschke, T. (2015). Increased impulsivity in pathological gambling: Considering nicotine dependence. Journal of Clinical and Experimental Neuropsychology, 37, 367–378. https://doi.org/10.1080/13803395.2015.1018145 Kräplin, A., Dshemuchadse, M., Behrendt, S., Scherbaum, S., Goschke, T., & Bühringer, G. (2014). Dysfunctional decisionmaking in pathological gambling: Pattern specificity and the role of impulsivity. Psychiatry Research, 215, 675–682. https:// doi.org/10.1016/j.psychres.2013.12.041 Kurth-Nelson, Z., Bickel, W., & Redish, A. D. (2012). A theoretical account of cognitive effects in delay discounting. The European Journal of Neuroscience, 35, 1052–1064. https://doi.org/ 10.1111/j.1460-9568.2012.08058.x Lagorio, C. H., & Madden, G. J. (2005). Delay discounting of real and hypothetical rewards III: Steady-state assessments, forcedchoice trials, and all real rewards. Behavioural Processes, 69, 173–187. https://doi.org/10.1016/j.beproc.2005.02.003 Laibson, D. (1997). Golden eggs and hyperbolic discounting. The Quarterly Journal of Economics, 112, 443–477. https://doi.org/ 10.1162/003355397555253 Lempert, K. M., & Phelps, E. A. (2016). The malleability of intertemporal choice. Trends in Cognitive Sciences, 20, 64–74. https://doi.org/10.1016/j.tics.2015.09.005 Loomes, G. (2010). Modeling choice and valuation in decision experiments. Psychological Review, 117, 902–924. https://doi. org/10.1037/a0019807 McClure, S. M., Laibson, D. I., Loewenstein, G., & Cohen, J. D. (2004). Separate neural systems value immediate and delayed monetary rewards. Science, 306, 503–507. https://doi.org/ 10.1126/science.1100907 Metcalfe, J., & Mischel, W. (1999). A hot/cool-system analysis of delay of gratification: Dynamics of willpower. Psychological Review, 106, 3–19. https://doi.org/10.1037/0033-295X.106.1.3 Mikl, M., Marecek, R., Hlustík, P., Pavlicová, M., Drastich, A., Chlebus, P., . . . Krupa, P. (2008). Effects of spatial smoothing on fMRI group inferences. Magnetic Resonance Imaging, 26, 490–503. https://doi.org/10.1016/j.mri.2007.08.006 Müller, M. M., Andersen, S. K., & Keil, A. (2007). Time course of competition for visual processing resources between emotional pictures and foreground task. Cerebral Cortex, 18, 1892–1899. https://doi.org/10.1093/cercor/bhm215 O’Hora, D., Dale, R., Piiroinen, P. T., & Connolly, F. (2013). Local dynamics in decision making: The evolution of preference within and across decisions. Scientific Reports, 3, Article 2210. https://doi.org/10.1038/srep02210 Pelli, D. G. (1997). The VideoToolbox software for visual psychophysics: Transforming numbers into movies. Spatial Vision, 10, 437–442. https://doi.org/10.1163/156856897X00366 Peters, J., & Büchel, C. (2011). The neural mechanisms of intertemporal decision-making: Understanding variability. Trends in Cognitive Sciences, 15, 227–239. https://doi.org/10.1016/j.tics. 2011.03.002 Ridderinkhof, K., van den Wildenberg, W. P., Wijnen, J., & Burle, B. (2004). Response inhibition in conflict tasks is revealed in delta plots. In M. Posner (Ed.), Cognitive neuroscience of attention (pp. 369–377). New York, NY: Guilford Press.

Ó 2018 Hogrefe Publishing

S. Scherbaum et al., Action Dynamics of Intertemporal Choices

Rodriguez, C. A., Turner, B. M., & McClure, S. M. (2014). Intertemporal choice as discounted value accumulation. PLoS One, 9, e90138. https://doi.org/10.1371/journal.pone.0090138 Samuelson, P. A. (1937). A note on measurement of utility. The Review of Economic Studies, 4, 155–161. https://doi.org/ 10.2307/2967612 Scherbaum, S., Dshemuchadse, M., Fischer, R., & Goschke, T. (2010). How decisions evolve: The temporal dynamics of action selection. Cognition, 115, 407–416. https://doi.org/10.1016/ j.cognition.2010.02.004 Scherbaum, S., Dshemuchadse, M., & Goschke, T. (2012). Building a bridge into the future: Dynamic connectionist modeling as an integrative tool for research on intertemporal choice. Frontiers in Cognition, 3, 514. https://doi.org/10.3389/fpsyg. 2012.00514 Scherbaum, S., Dshemuchadse, M., Leiberg, S., & Goschke, T. (2013). Harder than expected: Increased conflict in clearly disadvantageous intertemporal choices in a computer game. PLoS One, 8, e79310. https://doi.org/10.1371/journal.pone. 0079310 Scherbaum, S., Fischer, R., Dshemuchadse, M., & Goschke, T. (2011). The dynamics of cognitive control: Evidence for withintrial conflict adaptation from frequency-tagged EEG. Psychophysiology, 48, 591–600. https://doi.org/10.1111/j.14698986.2010.01137.x Scherbaum, S., Gottschalk, C., Dshemuchadse, M., & Fischer, R. (2015). Action dynamics in multitasking: The impact of additional task factors on the execution of the prioritized motor movement. Frontiers in Cognition, 6, 934. https://doi.org/ 10.3389/fpsyg.2015.00934 Scholten, M., & Read, D. (2006). Discounting by intervals: A generalized model of intertemporal choice. Management Science, 52, 1424–1436. https://doi.org/10.1287/mnsc.1060. 0534

Ó 2018 Hogrefe Publishing

Scholten, M., & Read, D. (2010). The psychology of intertemporal tradeoffs. Psychological Review, 117, 925–944. https://doi.org/ 10.1037/a0019619 Spivey, M. J., Grosjean, M., & Knoblich, G. (2005). Continuous attraction toward phonological competitors. Proceedings of the National Academy of Sciences of the United States of America, 102, 10393–10398. https://doi.org/10.1073/pnas.0503903102 Stewart, N., Chater, N., & Brown, G. D. A. (2006). Decision by sampling. Cognitive Psychology, 53, 1–26. https://doi.org/ 10.1016/j.cogpsych.2005.10.003 Travers, E., Rolison, J. J., & Feeney, A. (2016). The time course of conflict on the Cognitive Reflection Test. Cognition, 150, 109–118. https://doi.org/10.1016/j.cognition.2016.01.015 Wittmann, M., & Paulus, M. P. (2007). Decision making, impulsivity and time perception. Trends in Cognitive Sciences, 12, 7–12. https://doi.org/10.1016/j.tics.2007.10.004 Zauberman, G., Kim, B. K., Malkoc, S. A., & Bettman, J. R. (2009). Discounting time and time discounting: Subjective time perception and intertemporal preferences. Journal of Marketing Research, 46, 543–556. https://doi.org/10.1509/ jmkr.46.4.543 Received December 9, 2016 Revision received May 5, 2017 Accepted August 21, 2017 Published online February 8, 2018 Stefan Scherbaum Department of Psychology Technische Universität Dresden Zellescher Weg 17 01062 Dresden Germany stefan.scherbaum@tu-dresden.de

Experimental Psychology (2018), 65(1), 23–31

Short Research Article

Integrating Orthographic Information Across Time and Space Masked Priming and Flanker Effects With Orthographic Neighbors Joshua Snell,1 Daisy Bertrand,2 Martijn Meeter,3 and Jonathan Grainger1 1

Laboratoire de Psychologie Cognitive, Aix-Marseille University & CNRS, Marseille, France

Centre d’Etudes et de Recherche en Gestion d’Aix-Marseille, Aix-Marseille University, Aix-en-Provence, France

Vrije Universiteit Amsterdam, LEARN ! Research Institute, Amsterdam, The Netherlands

Abstract: Research has suggested that the word recognition process is influenced by the integration of orthographic information across words. The precise nature of this integration process may vary, however, depending on whether words are in temporal or spatial proximity. Here we present a lexical decision experiment, designed to compare temporal and spatial integration processes more directly. Masked priming was used to reveal effects of temporal integration, while the flanker paradigm was used to reveal effects of spatial integration. Primes/flankers were high-frequency orthographic neighbors of the target (blue-blur) or unrelated control words (head-blur). We replicated prior observations of inhibition in trials where the neighbor was used as a masked prime, while facilitation was observed in trials where the neighbor was presented as flanker. We conclude that sub-lexical orthographic information is integrated both temporally and spatially, but that spatial information is used to segregate lexical representations activated by spatially distinct sources. Keywords: reading, orthographic processing, masked priming, flanker paradigm, spatial integration, temporal integration

The processing of orthographic information during reading involves both the temporal and spatial integration of information. Temporal integration of orthographic information concerns the accumulation over time of information extracted from the same spatial location, and is typically evaluated by presenting successive orthographic stimuli (words and nonwords) at the same location (Grainger & Jacobs, 1999). Spatial integration of orthographic information concerns the combination of information extracted from different word locations, at the same point in time (e.g., Dare & Shillcock, 2013). In the present study we investigate the mechanisms that may underlie these integration processes, and in particular, to what extent they may differ. The masked priming paradigm (Forster & Davis, 1984) has been the paradigm of choice for investigating the temporal integration of information during single word reading. Brief presentation of the prime stimulus is thought to prevent it from being perceived as a distinct perceptual event (Humphreys, Evett, & Quinlan, 1990) hence

Experimental Psychology (2018), 65(1), 32–39 https://doi.org/10.1027/1618-3169/a000386

facilitating integration of information across prime and target (Grainger & Jacobs, 1999). Temporal integration of orthographic information can then be investigated by manipulating the orthographic overlap across prime and target stimuli (e.g., Ferrand & Grainger, 1992; Forster & Davis, 1984; Humphreys et al., 1990). More recently, spatial integration of orthographic information has been revealed in a paradigm introduced by Dare and Shillcock (2013), the flanking letters lexical decision (FLLD) task, whereby a central target stimulus is flanked by two letters on each side, separated from the target by a space (e.g., “ro rock ck”). Here, spatial integration is investigated by means of manipulating the orthographic overlap between the target word and the two flanking stimuli. In the present study, we focus on one manipulation that has produced contrasting effects in the masked priming and flanker paradigms. The manipulation in question is one where primes/flankers can be orthographic neighbors of target words (e.g., blue-blur) or unrelated words

Ó 2018 Hogrefe Publishing

J. Snell et al., Integrating Orthographic Information Across Time and Space

(e.g., head-blur). Prior research has revealed inhibitory effects of orthographic neighbor primes in masked priming (e.g., Davis & Lupker, 2006; De Moor & Brysbaert, 2000; Segui & Grainger, 1990). On the contrary, orthographic neighbor flanking stimuli have been found to facilitate target word processing (Snell, Vitu, & Grainger, 2017). The inhibitory effects of word neighbor primes found with masked priming have been taken as evidence for competitive processes operating between lexical representations (lexical competition) during visual word recognition (Segui & Grainger, 1990). In support of this interpretation, Jacobs and Grainger (1992) demonstrated that lateral inhibition across co-activated lexical representations in the interactive-activation model (McClelland & Rumelhart, 1981) accurately simulated inhibitory priming effects from orthographic neighbors. It is further known that these effects are affected by word frequency and -lexicality: the strongest inhibition is obtained with a combination of high-frequency prime words and low-frequency target words (Segui & Grainger, 1990), while nonword neighbor primes either generate facilitatory priming or null effects (Forster & Davis, 1991; Forster, Davis, Schoknecht, & Carter, 1987; Van Heuven, Dijkstra, Grainger, & Schriefers, 2001). The inhibitory effects of neighbor primes but concurrent facilitatory effects of nonword neighbor primes suggest that the temporal integration of orthographic information takes place both at the sub-lexical level as well as the lexical level. Following this reasoning, considering that orthographic neighbors facilitated target processing in the flanker paradigm (Snell, Vitu, et al., 2017), Snell et al. concluded that the spatial integration of orthographic information operates at the sub-lexical level but not beyond. We further elaborate on this reasoning in the section “Discussion”. The facilitatory parafoveal-on-foveal effect reported by Snell, Vitu, et al. (2017) speaks against a single-channel “one-word-at-a-time” approach to word identification and reading (e.g., Grainger, Dufau, & Ziegler, 2016; Reichle, Pollatsek, & Rayner, 2006). According to Grainger et al. (2016), orthographic information provided by flanking stimuli is integrated into a single channel that outputs a unique word identity. Given a flanker condition “bl blur ue,” the flanking letters “bl” and “ue” should combine with orthographic information extracted from the target “blur” and provide evidence for the competing word “blue,” leading to inhibition and not to the facilitation observed by Snell, Vitu, et al. (2017). Instead, their results suggest that despite the spatial integration of sub-lexical orthographic information, the lexical representations that are consequently activated continue to be processed independent from one

another – as long as these are associated with spatially distinct sources (see also Snell, Meeter, & Grainger, 2017). On the other hand, one could argue that this pattern was obtained because orthographic information concerning the competing word was split across the left and right flankers in the Snell et al. experiment, whereas in masked priming the competing word was intact. This caused individual flankers to bear no lexical status (e.g., neither “ro” nor “ck” in “ro rock ck” is a word), as such possibly activating sub-lexical integration processes but not lexical integration processes. It is therefore important to examine effects of word neighbor flankers when these are intact, such as in the example “blue blur blue” – while ensuring, crucially, that no facilitation is obtained with the same stimuli and participants in the masked priming paradigm. This was the primary goal of the present study.

Method Participants Thirty-two students from Aix-Marseille University gave informed consent to participate in this experiment and received €4. All participants reported to be native to the French language, non-dyslexic, and had normal or corrected-to-normal vision. All participants were naïve to the purpose of the experiment.

Materials Using the same procedure as in Snell, Vitu, et al. (2017), we retrieved a list of 74 triplets (target word (e.g., “brut”), orthographic neighbor (e.g., “bout”), and orthographically unrelated control word (e.g., “noix”) from the French Lexicon Project database (Ferrand et al., 2010). All words consisted of four letters, were nonconjugated, and contained no diacritics. Word pairings were chosen such that orthographic neighbors and control words had a lower lexical decision time (LDT) than their respective target word (for targets, neighbors, and controls, the mean LDT was 671 ms, 618 ms, and 615 ms, respectively).1 Targets and neighbors only differed in an inner-positioned letter. In a similar fashion we retrieved a list of 74 pseudoword triplets from the French Lexicon Project pseudoword database (Ferrand et al., 2010). These were filler stimuli, not to be included in data analyses. We present the complete stimulus list in the Appendix.

Following Snell, Vitu, et al. (2017) we selected stimuli based on the LDT measure because it more directly reflects the speed with which words become active and reach recognition. Words with a low-LDT value are activated faster, and as such exert stronger inhibition on lexical competitors; hence the choice for low-LDT primes and high-LDT targets.

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 32–39

J. Snell et al., Integrating Orthographic Information Across Time and Space

Figure 1. Overview of the trial procedure in the flanker setting (top) and masked priming setting (bottom). The size of stimuli relative to the screen is exaggerated in these examples.

Design We used a 2 2 2 factorial design, with word lexicality (word/pseudoword), trial type (masked priming/flanking), and relatedness of the prime/flanker (neighbor/control) as factors. Participants were Latin-squared into two groups, such that every stimulus was presented twice to each participant (once in the neighbor condition and once in the control condition) and in both trial types per two participants. The experiment thus consisted of 296 trials per participant (148 of which were included in the analyses), and these were presented in randomized order.

Apparatus The experiment was implemented with OpenSesame (Mathôt, Schreij, & Theeuwes, 2012) and presented on a 1,024 768 px, 150 Hz computer monitor (Dell, Trinitron series, Dell Inc., Austin, TX, USA). Participants were seated in a comfortable office chair at a distance of 50 cm from the display, so that each character space subtended 0.40 degrees of visual angle. Responses were collected with a keyboard.

Procedure Before commencing the experiment, participants received instructions both verbally by the experimenter and visually onscreen. Participants were instructed to fixate in between two centrally located vertical fixation bars that were presented throughout the experiment. Figure 1 shows the procedure for each trial type. Both trial types would start with a 500 ms mask consisting of four hashmarks. In masked priming trials, the mask would be replaced by the 2

neighbor/control for 70 ms, followed by the target word. In the flanker trials, the mask would be replaced by the target, with the neighbor/control being presented left and right of the target (separated by a single character space). Following Snell, Vitu, et al. (2017), all words were presented in 18-point monospaced font (droid sans mono; the default monospaced font in OpenSesame) and in lowercase. The target would stay onscreen until participants pressed a leftor right-handed key for pseudoword or word, respectively. Participants were instructed to respond as quick and accurate as possible, and the maximum allowed response time (RT) was 1,800 ms after the target onset. Participants received feedback in the form of a briefly presented centrally located green or red dot, for correct and incorrect responses, respectively. The next trial began immediately after the 600 ms feedback signal.

Results Only correctly answered trials (93.14%) were included in the analysis of response times (RT).2 For our analyses of RTs and error rates we used linear mixed-effect models (LMMs) with items and participants as crossed random effects (Baayen, 2008). To meet the models’ assumption that the data were normally distributed, RTs were inversetransformed ( 1,000/RT) prior to the analyses. The models were fitted with the lmer function from the lme4 package (Bates, Maechler, Bolker, & Walker, 2015) in the R statistical computing environment. Following Barr, Levy, Scheepers, and Tily (2013) we determined the maximal random effect structure permitted by the data. This led us to include by-item and by-participant random intercepts, as well as by-item and by-participant random slopes

The raw data files can be accessed online at https://osf.io/tq38d/.

Experimental Psychology (2018), 65(1), 32–39

Ó 2018 Hogrefe Publishing

J. Snell et al., Integrating Orthographic Information Across Time and Space

RT ~ Relatedness + (1+Relatedness|Participant) + (1+Relatedness|Item) Figure 2. The typical LMM structure used to analyze the present data. Here, RT is the dependent variable, whereas the model terms are presented to the right of the tilde (“ ”) character. The first term (Relatedness) is a fixed effect. The next two terms are random effects, with the random factors (respectively, Participant and Item) being presented right of the bar (“|”) character. The expression to the left of the bar indicates the inclusion of random intercept (“1”) and random slope (“Relatedness”).

(Figure 2). We report regression coefficients (b), standard errors (SE) and t-values. Fixed effects were deemed reliable if |t| > 1.96 (Baayen, 2008). Logistic LMMs (fitted with the glmer function) were used to analyze the error rates. Below we present separate analyses for flanker trials and masked priming trials, followed by a direct comparison of the two trial types.

Table 1. Response times (ms) and error rates (probability)

Flanker Trials

Lastly, there was a noteworthy main effect of trial type on RTs, with increased RTs in the flanker setting compared to the masked priming setting (b = 0.06, SE = 0.01, t = 5.67). This suggests that flankers generally perturbed target processing more than primes.

We replicated the finding of Snell, Vitu, et al. (2017) that target processing is facilitated by orthographic neighbor flankers, as RTs were significantly shorter in the neighbor condition as compared to the control condition (b = 0.03, SE = 0.01, t = 2.54; condition means in Table 1). The error rate did not differ significantly between conditions (b = 0.25, SE = 0.19, z = 1.36).

Masked Priming Trials Whereas our neighbor stimuli were found to facilitate target processing in the flanker condition, the opposite pattern was found in the masked priming trials. An inhibitory effect was found in the error rates, with significantly more errors following neighbor primes than control primes (b = 0.43, SE = 0.22, z = 1.98). The pattern of RTs followed the same direction numerically (see Table 1), but did not reach significance (b = 0.02, SE = 0.01, t = 1.27).3

Comparison of Trial Types To compare the two trial types directly,4 we entered the interaction of Relatedness Trial type in a separate model. The effect in RTs of prime/flanker on target processing turned out to interact significantly with trial type (b = 0.05, SE = 0.02, t = 3.15), thus confirming the significance of the opposite pattern of effects found in the two trial types (Table 1). No significant interaction was established in the error rates (b = 0.19, SE = 0.28, z = 0.69; Table 1). 3

Response times

Error rates

Condition

Neighbor

Control

Neighbor

Control

Flanker trials

731 (171)

742 (164)

.072 (.045)

.059 (.050)

Masked priming

707 (161)

697 (149)

.058 (.044)

.039 (.048)

Note. Values in between parentheses indicate standard deviations.

Discussion A lexical decision experiment examined the effects of orthographic neighbors on target word recognition when the neighbors were either presented as masked primes immediately before the target word at the same location, or presented as flanking words simultaneously with, and to the left and to the right of the target word. The general aim was to compare the temporal integration of orthographic information as revealed by masked priming, with the spatial integration of orthographic information as revealed by the flanker task. Insight into the respective natures of these different types of integration further provides a means to test two opposing accounts of word identification and reading: a single-channel, one-word-at-a-time account (e.g., Grainger et al., 2016; Reichle et al., 2006) and a multichannel, parallel word identification account (e.g., Snell, Meeter, et al., 2017). According to the single-channel model proposed by Grainger et al. (2016), orthographic processing operates in parallel across multiple words during sentence reading (cf., Engbert, Nuthmann, Richter, & Kliegl, 2005; Reilly & Radach, 2006), and the orthographic information extracted from different words is integrated into a single processing

It should be noted that inhibitory prime effects are not always established in RT data. Zimmerman and Gomez (2012) have argued that the amount of attentional resources spent on processing of the prime directly affects the chance of finding an inhibitory effect, and that longer prime durations might as such lead to stronger inhibitory effects. We acknowledge that such a comparison is complicated by the many differences between the two paradigms (we further elaborate on these differences in the section “Discussion”), but believe that a direct comparison is nonetheless relevant in the context of the present study.

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 32–39

J. Snell et al., Integrating Orthographic Information Across Time and Space

channel that outputs a unique word identity. We reasoned that if this were the case, then the presence of an orthographic neighbor as flanking stimulus should lead to inhibition of target word processing, mimicking the effects seen when orthographic neighbors are presented as masked prime stimuli (Davis & Lupker, 2006; De Moor & Brysbaert, 2000; Segui & Grainger, 1990). Snell, Vitu, et al. (2017) put this reasoning to test and, on the contrary, found that parafoveal orthographic neighbors facilitated target word processing both in a sentence reading setting as well as in the flanker task (e.g., “bl blur ue”). When used as flanker, however, the orthographic neighbor was divided either side of the target, contrary to the use of complete prime stimuli in masked priming experiments. The results of the present experiment show that presenting whole-word flankers on either side of the target (e.g., “blue blur blue”) produces a similar facilitatory effect, the size of the effect being 11 ms compared with the 14 ms effect in the Snell, Vitu, et al. (2017) study.5 Crucially, in the present study, the same participants showed an inhibitory priming effect with the same stimuli when these were presented as primes and targets in a masked priming procedure. Why then do orthographic neighbor flanking stimuli facilitate target word identification in the flanker task? The answer offered by Snell, Vitu, et al. (2017) is that orthographic information extracted from distinct spatial locations is integrated sub-lexically (see also Angele, Tran, & Rayner, 2013; Grainger, Mathôt, & Vitu, 2014; Snell, Vitu, et al., 2017), hence facilitating target word recognition when there is orthographic overlap. What is novel in Snell, Meeter, et al.’s (2017) account is that spatial information is used to keep track of which activated word representation belongs to which spatial location, hence enabling parallel higher-level processing of multiple stimuli. The fact that this parallel processing is geared to output several distinct word identities means that flanker and target stimuli do not interfere at the level of lexical processing and beyond. Thus, whereas sub-lexical orthographic information is integrated across spatially and temporally distinct stimuli, lexical integration takes place within- rather than across spatial locations. On a methodological note, it is important to consider the various differences between the masked priming and flanker trials – in particular with respect to the availability of the prime/flanker stimulus during target processing – and whether or not such differences may have contributed to the opposing (facilitatory vs. inhibitory) effects obtained in each respective setting. Concretely, one might argue that the neighbor could have been processed to a further extent in flanker trials than in masked priming trials, given that the 5

neighbor was only available for 70 ms in the latter trial type whereas it was available during the whole stimulusresponse interval in the former trial type. On the other hand, one might argue that the constraints imposed by visual acuity cause foveal processing of the prime stimulus to be of higher quality than parafoveal processing of the flanker stimulus, as such compensating for their different presentation time. Importantly, we opted to keep flanking stimuli onscreen rather than to have them disappear after 70 ms because the offset of these stimuli would have directed attention away from the fovea (similar to a stimulus onset). Crucially, even if flankers were processed to a further degree than primes, this should have then only increased the effects that were established here. Indeed, it is clear that the potentially increased processing of flankers compared to primes did not lead to inhibition, as might otherwise be expected following deeper integration of information between orthographic neighbors. In sum, the present results underline the idea that the integration of orthographic information from multiple words can impact the recognition process in various ways. The outcome of this integration process seems to depend strongly on the words’ spatial locations, in line with the idea that readers keep track of which word belongs to which position: when word representations are tied to the same spatial location, the integration of information is carried on to the lexical level, where lexical competition perturbs the recognition process. In contrast, when word representations are tied to different spatial locations, this segregation allows for parallel independent lexical processing, resulting in stronger activation and faster word recognition.

Acknowledgment This research was supported by Grants ANR-15-CE330002-01 and ANR-11-LABX-0036 (BLRI) from the French National Research Agency (ANR).

References Angele, B., Tran, R., & Rayner, K.. Parafoveal-foveal overlap can facilitate ongoing word identification during reading: Evidence from eye movements. Journal of Experimental Psychology: Human Perception and Performance, 39, 526–538. https://doi. org/10.1037/a0029492 Baayen, R. (2008). Analyzing Linguistic Data: A practical introduction to statistics. Cambridge, UK: Cambridge University Press. Barr, D., Levy, R., Scheepers, C., & Tily, H. (2013). Random effects structure for confirmatory hypothesis testing: Keep it maximal. Journal of Memory and Language, 68, 255–278. https://doi.org/ 10.1016/j.jml.2012.11.001

We further note the presence of a small speed-accuracy trade-off in the present study, with a nonsignificant increase in errors arising in the presence of related flankers.

Experimental Psychology (2018), 65(1), 32–39

Ó 2018 Hogrefe Publishing

J. Snell et al., Integrating Orthographic Information Across Time and Space

Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects Models using lme4. Journal of Statistical Software, 67, 1–48. https://doi.org/10.18637/jss.v067.i01 Dare, N., & Shillcock, R. (2013). Serial and parallel processing in reading: Investigating the effects of parafoveal orthographic information on nonisolated word recognition. The Quarterly Journal of Experimental Psychology, 66, 417–428. https://doi. org/10.1080/17470218.2012.759979 Davis, C., & Lupker, S. (2006). Masked inhibitory priming in English: Evidence for lexical inhibition. Journal of Experimental Psychology: Human Perception and Performance, 32, 668–687. https://doi.org/10.1037/00961523.32.3.668 De Moor, W., & Brysbaert, M. (2000). Neighborhood-frequency effects when primes and targets are of different lengths. Psychological Research, 63, 159–162. https://doi.org/10.1007/ PL00008174 Engbert, R., Nuthmann, A., Richter, E., & Kliegl, R. (2005). SWIFT: A dynamical model of saccade generation during reading. Psychological Review, 112, 777–813. https://doi.org/10.1037/ 0033-295X.112.4.777 Ferrand, L., & Grainger, J. (1992). Phonology and orthography in visual word recognition: Evidence from masked nonword priming. The Quarterly Journal of Experimental Psychology, 45, 353–372. https://doi.org/10.1080/02724989208250619 Ferrand, L., New, B., Brysbaert, M., Keuleers, E., Bonin, P., Méot, A., . . . Pallier, C. (2010). The French Lexicon Project: Lexical decision data for 38,840 French words and 38,840 pseudowords. Behavior Research Methods, 42, 488–496. https://doi. org/10.3758/BRM.42.2.488 Forster, K., & Davis, C. (1984). Repetition priming and frequency attenuation in lexical access. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 680–698. https://doi.org/10.1037/0278-7393.10.4.680 Forster, K., & Davis, C. (1991). The density constraint on formpriming in the naming task: Interference effects from a masked prime. Journal of Memory and Language, 30, 1–25. https://doi. org/10.1016/0749-596X(91)90008-8 Forster, K., Davis, C., Schoknecht, C., & Carter, R. (1987). Masked priming with graphemically related forms: Repetition or partial activation? The Quarterly Journal of Experimental Psychology, 39, 211–251. https://doi.org/10.1080/ 14640748708401785 Grainger, J., Dufau, S., & Ziegler, J. (2016). A vision of reading. Trends in Cognitive Sciences, 20, 171–179. https://doi.org/ 10.1016/j.tics.2015.12.008 Grainger, J., & Jacobs, A. (1999). Temporal integration of information in orthographic priming. Visual Cognition, 6, 461– 492. https://doi.org/10.1080/135062899395064 Grainger, J., Mathôt, S., & Vitu, F. (2014). Test of a model of multiword reading: Effects of parafoveal flanking letters on foveal word recognition. Acta Psychologica, 146, 35–40. https://doi. org/10.1016/j.actpsy.2013.11.014 Humphreys, G., Evett, L., & Quinlan, P. (1990). Orthographic processing in visual word identification. Cognitive Psychology, 22, 517–560. https://doi.org/10.1016/0010-0285(90)90012-S

Ó 2018 Hogrefe Publishing

Jacobs, A., & Grainger, J. (1992). Testing a semistochastic variant of the interactive activation model in different word recognition experiments. Journal of Experimental Psychology, Human Perception and Performance, 20, 1311–1334. https://doi.org/ 10.1037/0096-1523.18.4.1174 Mathôt, S., Schreij, D., & Theeuwes, J. (2012). OpenSesame: An open-source, graphical experiment builder for the social sciences. Behavior Research Methods, 44, 314–324. https:// doi.org/10.3758/s13428-011-0168-7 McClelland, J., & Rumelhart, D. (1981). An interactive activation model of context effects in letter perception: Part I. An account of basic findings. Psychological Review, 88, 375–407. https:// doi.org/10.1037/0033-295X.84.5.413 Reichle, E., Pollatsek, A., & Rayner, K. (2006). E-Z Reader: A cognitive-control, serial-attention model of eye movement behavior during reading. Cognitive Systems Research, 7, 4–22. https://doi.org/10.1016/j.cogsys.2005.07.002 Reilly, R., & Radach, R. (2006). Some empirical tests of an interactive activation model of eye movement control in reading. Cognitive Systems Research, 7, 34–55. https://doi. org/10.1016/j.cogsys.2005.07.006 Snell, J., Meeter, M., & Grainger, J. (2017). Evidence for simultaneous syntactic processing of multiple words during reading. PLoS One, 12, e0173720. https://doi.org/10.1371/journal. pone.0173720 Snell, J., Vitu, F., & Grainger, J. (2017). Integration of parafoveal orthographic information during foveal word reading: Beyond the sub-lexical level? The Quarterly Journal of Experimental Psychology, 70, 1984–1994. https://doi.org/10.1080/17470218. 2016.1217247 Segui, J., & Grainger, J. (1990). Priming word recognition with orthographic neighbors: Effects of relative prime-target frequency. Journal of Experimental Psychology: Human Perception and Performance, 16, 65–76. https://doi.org/10.1037/ 0096-1523.16.1.65 Van Heuven, W., Dijkstra, T., Grainger, J., & Schriefers, H. (2001). Shared neighborhood effects in masked orthographic priming. Psychonomic Bulletin & Review, 8, 96–101. https://doi.org/ 10.1080/17470218.2013.850521 Zimmerman, R., & Gomez, P. (2012). Drawing attention to primes increases inhibitory word priming effects. The Mental Lexicon, 7, 119–146. https://doi.org/10.1075/ML.7.2.01zim Received May 21, 2017 Revision received September 1, 2017 Accepted September 1, 2017 Published online February 8, 2018 Joshua Snell Laboratoire de Psychologie Cognitive Aix-Marseille University 3 place Victor Hugo 13331 Marseille France joshua.snell@hotmail.com

Experimental Psychology (2018), 65(1), 32–39

J. Snell et al., Integrating Orthographic Information Across Time and Space

Appendix Table A1. Stimuli used in the present experiment Words Target

Neighbor

Aile Aire

Pseudowords Control

Target

Neighbor

Control

Aide

Fond

Orve

Ouve

Atal

Aile

Moto

Rine

Rone

Abac

Base

Baie

Rond

Afit

Adit

Spen

Bise

Bile

Poux

Stin

Suin

Moce

Bond

Bord

Arme

Bral

Bril

Uste

Boxe

Boue

Rail

Puif

Poif

Lage

Brin

Brun

Taxe

Nile

Nole

Carc

Brut

Bout

Noix

Psat

Psut

Ulle

Cage

Cave

Soif

Guve

Guge

Insi

Case

Cape

Midi

Assa

Asta

Itre

Chic

Choc

Peau

Tuce

Tute

Glin

Chut

Chat

Lire

Uise

Uide

Alon

Clou

Chou

Haie

Vret

Vuet

Gnal

Cote

Code

Pays

Oche

Ocle

Fida

Crue

Crie

Clan

Olde

Olle

Ahui

Cuve

Cube

Poil

Sabe

Sube

Rêt

Dame

Date

Noir

Vave

Vate

Murf

Dune

Dupe

Sort

Vure

Vuie

Imci

Fade

Face

Joli

Lube

Luve

Giot

Fine

Fixe

Drap

Nouf

Noif

Pala

Flic

Fric

Onue

Oque

Gami

Flot

Foot

Abri

Jine

Jive

Momb

Flux

Feux

Amie

Cude

Cuse

Gnai

Gage

Gaie

Unir

Tese

Tose

Haid

Gaie

Gare

Bouc

June

Juse

Tipt

Gras

Gros

Dent

Vore

Vose

Jaut

Huer

Hier

Golf

Itie

Ilie

Ranu

Joue

Joie

Rang

Erle

Erme

Vonc

Juge

Jupe

Gris

Coui

Cofi

Alve

Laid

Lard

Peur

Adre

Adie

Ofci

Lave

Lame

Cuir

Uite

Uire

Phol

Lion

Lien

Bref

Vare

Vace

Noui

Logo

Loto

Nier

Cona

Cena

Didi

Luge

Loge

Pain

Oute

Oste

Misi

Lune

Luxe

Fort

Soge

Sige

Trit

Mime

Mine

Sauf

Spel

Suel

Wadu

Mont

Mort

Vive

Arne

Arie

Mout

Moue

Mode

Pair

Onsa

Ofsa

Dute

Muse

Mule

Nord

Chen

Cien

Gord

Nerf

Neuf

Plat

Geuf

Gerf

Acci

Noce

Note

Juin

Iage

Inge

Stou

Ocre

Ogre

Aigu

Amit

Anit

Greu

Onde

Onze

Mari

Jote

Jore

Clai

Page

Pape

Film

Sute

Supe

Dida

Paix

Prix

Bleu

Fien

Fren

Solu (Continued on next page)

Experimental Psychology (2018), 65(1), 32–39

Ó 2018 Hogrefe Publishing

J. Snell et al., Integrating Orthographic Information Across Time and Space

Table A1. (Continued) Words

Pseudowords

Target

Neighbor

Control

Target

Neighbor

Control

Paon

Pion

Fier

Stre

Stue

Ilci

Pieu

Pneu

Tour

Arni

Arti

Ouge

Pipe

Pile

Trou

Dave

Dade

Gonc

Pire

Pure

Thon

Arve

Aive

Ucun

Pont

Port

Mare

Euve

Eule

Drif

Porc

Parc

Test

Buge

Bune

Tion

Port

Part

Beau

Poge

Poce

Arut

Pote

Pose

Char

Nobe

Noge

Phif

Pure

Puce

Loin

Ique

Idue

Atat

Race

Rage

Mois

Gict

Gint

Nala

Raie

Rate

Doux

Ajet

Anet

Frum

Rame

Rare

Bloc

Jave

Jace

Rodi

Ride

Rire

Long

Nora

Noma

Eige

Rime

Rive

Solo

Nent

Nept

Jurf

Robe

Rose

Saut

Oile

Oble

Daud

Roue

Robe

Bain

Ince

Inse

Psou

Rude

Ride

Kilo

Mune

Muve

Bara

Ruse

Rude

Foin

Vact

Valt

Orio

Sain

Sein

Clef

Bome

Boce

Drai

Sale

Sage

Coin

Ogne

Ogme

Fauf

Scie

Soie

Jury

Cavu

Catu

Phre

Sein

Soin

Papa

Nire

Nure

Spho

Soja

Soda

Lent

Alse

Alme

Cigi

Taux

Toux

Bois

Enre

Ente

Vima

Toit

Tort

Grec

Suve

Sule

Apit

Trac

Troc

Sens

Zote

Zode

Afil

Troc

Truc

Lieu

Oire

Oige

Satu

Vent

Vert

Bras

Imin

Itin

Stet

Vice

Vite

Tard

Igle

Igue

Uant

Vide

Vice

Abus

Fuve

Fube

Clat

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 32–39

Alternatives to traditional self-reports in psychological assessment “A unique and timely guide to better psychological assessment.” Rainer K. Silbereisen, Research Professor, Friedrich Schiller University Jena, Germany Past-President, International Union of Psychological Science

Tuulia Ortner / Fons J. R. van de Vijver (Editors)

Behavior-Based Assessment in Psychology Going Beyond Self-Report in the Personality, Affective, Motivation, and Social Domains (Series: Psychological Assessment – Science and Practice – Vol. 1) 2015, vi + 234 pp. US $63.00 / € 44.95 ISBN 978-0-88937-437-9 Also available as eBook Traditional self-reports can be an unsufficiant source of information about personality, attitudes, affect, and motivation. What are the alternatives? This first volume in the authoritative series Psychological Assessment – Science and Practice discusses the most influential, state-of-the-art forms of assessment that can take us beyond self-report. Leading scholars from various countries describe the theo-

www.hogrefe.com

retical background and psychometric properties of alternatives to self-report, including behavior-based assessment, observational methods, innovative computerized procedures, indirect assessments, projective techniques, and narrative reports. They also look at the validity and practical application of such forms of assessment in domains as diverse as health, forensic, clinical, and consumer psychology.

Assessment methods in health psychology “This book is an excellent overview of measurement issues that are central to health psychology.” David French, PhD, Professor of Health Psychology, University of Manchester, UK

Yael Benyamini / Marie Johnston / Evangelos C. Karademas (Editors)

Assessment in Health Psychology (Series: Psychological Assessment – Science and Practice – Vol. 2) 2016, vi + 346 pp. US $69.00 / € 49.95 ISBN 978-0-88937-452-2 Also available as eBook

Assessment in Health Psychology presents and discusses the best and most appropriate assessment methods and instruments for all specific areas that are central for health psychologists. It also describes the conceptual and methodological bases for assessment in health psychology, as well as the most important current issues and recent progress in methods. A unique feature of this book, which brings together leading authorities on health psychology assessment, is its emphasis on the bidirectional link between theory and practice. Assessment in Health Psychology is addressed to masters and doctoral students in health psychology, to all

www.hogrefe.com

those who teach health psychology, to researchers from other disciplines, including clinical psychology, health promotion, and public health, as well as to health policy makers and other healthcare practitioners. This latest volume in the series Psychological Assessment – Science and Practice provides a thorough and authoritative record of the best available assessment tools and methods in health psychology, making it an invaluable resource both for students and academics as well as for practitioners in their daily work.

Short Research Article

Shared Processing of Language and Music Evidence From a Cross-Modal Interference Paradigm Ryan P. Atherton,1 Quin M. Chrobak,1 Frances H. Rauscher,1 Aaron T. Karst,1 Matt D. Hanson,2 Steven W. Steinert,3 and Kyra L. Bowe1 1

Department of Psychology, University of Wisconsin-Oshkosh, Oshkosh, WI, USA

The University at Albany, State University of New York, Albany, NY, USA

Eastern Michigan University, Ypsilanti, MI, USA

Abstract: The present study sought to explore whether musical information is processed by the phonological loop component of the working memory model of immediate memory. Original instantiations of this model primarily focused on the processing of linguistic information. However, the model was less clear about how acoustic information lacking phonological qualities is actively processed. Although previous research has generally supported shared processing of phonological and musical information, these studies were limited as a result of a number of methodological concerns (e.g., the use of simple tones as musical stimuli). In order to further investigate this issue, an auditory interference task was employed. Specifically, participants heard an initial stimulus (musical or linguistic) followed by an intervening stimulus (musical, linguistic, or silence) and were then asked to indicate whether a final test stimulus was the same as or different from the initial stimulus. Results indicated that mismatched interference conditions (i.e., musical – linguistic; linguistic – musical) resulted in greater interference than silence conditions, with matched interference conditions producing the greatest interference. Overall, these results suggest that processing of linguistic and musical information draws on at least some of the same cognitive resources. Keywords: language/memory interactions, music cognition, working memory

The perception of music is a complex task that involves multiple brain regions. Although research has begun to establish the neural correlates of music perception (e.g., Brown, Martinez, & Parsons, 2006; Koelsch et al., 2002), far less is known regarding the cognitive processes that are involved. For example, although there is a wealth of research exploring how information that is the direct focus of attention is processed and stored in immediate memory (see Posner & Snyder, 2004 for a review), only a few studies have explored these issues in regard to musical stimuli. Moreover, the research that has been conducted on this topic has produced conflicting results in terms of the cognitive systems that are involved (e.g., Deutsch, 1970; Semal, Demany, Ueda, & Halle, 1996; Williamson, Baddeley, & Hitch, 2010). The leading model that describes the processing and storage of information in immediate memory is the multicomponent working memory model (e.g., Baddeley & Hitch, 1974). Of particular interest to the present study

Experimental Psychology (2018), 65(1), 40–48 https://doi.org/10.1027/1618-3169/a000388

is the phonological loop, which has two components, the phonological store and the articulatory rehearsal process. The phonological store temporarily holds phonological information, while the articulatory rehearsal process continually refreshes information (by a process of subvocalization) until it is no longer needed. Importantly, the phonological store can be directly accessed by phonological information, whereas other inputs (e.g., visual) need to first be recoded by the articulatory rehearsal process (Baddeley, Lewis, & Vallar, 1984). Research on these systems has revealed a number of well-known findings regarding the duration of phonological information without rehearsal (e.g., Baddeley et al., 1984), the overall capacity of the phonological store (e.g., Baddeley, Thomson, & Buchanan, 1975), and the very nature of the information being rehearsed by the articulatory rehearsal process (e.g., Wilson & Emmorey, 1998). However, it is unclear to what extent these same components may contribute to the storage of purely musical information. At a broad level, the investigation of the type

Ó 2018 Hogrefe Publishing

R. P. Atherton et al., Language, Music, and Working Memory

of stimuli that are processed within working memory is of vital importance to our understanding of how this system functions. For example, if various types of stimuli can be processed by the phonological loop, what commonalities might these stimuli share?

Previous Research Deutsch (1970) sought to explore whether basic sensory information (in this case tonal information) was processed by the same cognitive system that processes phonological information. Participants in this study were presented with an initial tone, followed by an intervening tone or word (i.e., spoken numbers such as “three”), and then finally a comparison tone. Participants were then asked to indicate whether the comparison tone was the same pitch as the initial tone. Overall, performance was significantly disrupted if the initial and comparison tones were separated by intervening tones (Condition A). By contrast, intervening words had a negligible impact on participants’ ability to remember the initial tones, regardless of whether participants were required to ignore the intervening words (Condition B) or subsequently recall them (Condition C). Deutsch concluded that basic sensory information and phonological information have separate storage areas in immediate memory, as the latter type of intervening stimulus (i.e., words) failed to interfere with the to-be-remembered tonal information. Unfortunately, a methodological confound associated with this study may have contributed to the results reported. Specifically, the pitch of the initial tone was not matched with the pitch of the intervening words (even though they were matched between initial tones and intervening tones). As a result, the dissimilar pitch between the initial tone and the intervening word likely mitigated any potential interference in those conditions. Semal et al. (1996) used the same basic paradigm to examine the speech specificity hypothesis. One version of this hypothesis suggests that pitch is stored in two separate systems: one system for the pitch associated with language and a separate system for pitch associated with nonspeech sounds. In contrast to the rationale behind the Deutsch (1970) study, the authors were specifically interested in the processing of pitch information (as opposed to generic sensory information, which can exist in a variety of modalities; e.g., visual). In Experiment 1, participants heard an initial tone and then heard either intervening tones or words that varied in terms of their pitch similarity to the original stimulus. Participants then had to decide whether the comparison tone was the same pitch as the original. Experiment 2 was comparable to Experiment 1 with the 1

primary exception being that the initial and comparison stimuli were words. Critically, when the comparison word was presented, it was always the same word that participants had been presented with initially. Participants were tasked with indicating whether or not the pitch of the comparison word was the same as pitch of the initial word. Overall, Semal et al. (1996) found that both the intervening tones and the intervening speech sounds interfered with the pitch recollection of tones (Experiment 1) and of words (Experiment 2). Based on these findings the authors rejected the speech specificity hypotheses, in essence arguing instead for the shared storage and processing of pitch information – regardless of the source. It is important to note, however, that the stimuli used in this study were matched in terms of frequency, they still consisted of simple tones that lacked the systematic complexity that constitutes “music.” At a fundamental level, language and music are both complex sensory inputs that consist of larger units (e.g., words and chords) constructed from a series of intricately arranged smaller units (e.g., letters and notes).1 If the question at hand is whether music and language share processing resources in working memory, then it is essential to use stimuli that are functionally equivalent. To our knowledge, Williamson et al. (2010) were the first researchers to use nonlanguage stimuli arranged within a musical context to investigate issues related to the potential shared processing of linguistic and musical stimuli. Specifically, they explored the “pitch-proximity” effect, the finding that tones close together (i.e., pitch-proximal) are recalled less accurately than tones that are further apart (i.e., pitchdistal). This finding is akin to the phonological similarity effect (Conrad & Hull, 1964), wherein phonologically similar words are recalled less well than phonologically dissimilar words. For our purposes, the critical experiment reported in this paper was Experiment 3. Here participants heard groups of tones or letters and were asked to serially recall what they heard. Importantly, the tones consisted of frequencies characteristic of musical notes within the context of an actual musical key, in this case C Major (i.e., C4, D4, E4, G4, and B4 – all notes found in the key of C Major). If there is shared storage between musical and phonological information, pitch proximity should affect recall of both types of information. Data supported the shared storage of phonological and musical information in working memory, as nonmusician participants demonstrated a pitch-proximity effect that was comparable to the observed phonological similarity effect: tones that were proximal (i.e., similar or “close together”) were more difficult to accurately recall than sounds that were distal (i.e., different or “further apart”). Overall, based on the finding of similar effects for both phonological and musical stimuli, Williamson et al.

Analyses were run with and without these participants, and the overall pattern of results was consistent regardless.

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 40–48

(2010) concluded that musical and phonological information may be stored in similar areas of the brain. The results from both Semal et al. (1996) and Williamson et al. (2010) seem to support the notion of shared overlap in processing between language and music, though as noted earlier – it is unclear the extent to which the stimuli used in Semal et al. (1996) could be considered musical. However, both share a similar methodological limitation related to their inability to directly assess the potential interfering effect of musical stimuli on language processing. In Semal et al. (1996), the participants’ goal was to remember the pitch of the original stimuli that was presented to them. It was found that intervening tones and language both interfered with participants’ memory for pitch – and this was the case whether the initial stimulus was presented as a basic tone (Experiment 1) or as a word (Experiment 2). However, the experimental design did not assess participants’ memory for the language stimuli themselves (i.e., memory for the actual word). A similar argument can be made in terms of the results reported by Williamson et al. (2010). The primary focus of that study was to demonstrate a pitch-proximity effect for both musical and language stimuli. However, the pitch-proximity procedure reported in Williamson et al. (2010) does not allow one to directly assess the potential interfering effect of music on language, and vice versa, as the two classes of stimuli were not intermixed. Indeed, the authors suggest that the pattern of results that was reported could also occur if two separate processing systems existed and operated based on similar principles. Thus, while previous research is suggestive of shared processing, we believe that the methodological limitations of these studies warranted additional research using ecologically valid nonlanguage stimuli and methods that directly compared the potential interfering effects of each type of stimulus on one another.

The Current Investigation In order to further explore the possibility of shared overlap in processing between language and music stimuli, participants were presented with an initial stimulus (word or chord), followed by intervening stimuli (words, chords, or silence), and then asked if a comparison stimulus was the same as or different from the originally presented stimulus. This arrangement creates experimental conditions where the initial and intervening stimuli are either matched (language – language), mismatched (language – music), or neither (language – silence). The unique contribution of the current study is that it directly compared the crossdomain interfering effects of musical and language stimuli while using nonlanguage stimuli organized within a musical context. In order to equate the language and music stimuli on a structural level, the language stimuli (initial, Experimental Psychology (2018), 65(1), 40–48

R. P. Atherton et al., Language, Music, and Working Memory

Figure 1. This is one version of the circle of fifths, showing a visual representation of the relationship between the 12 tones of the chromatic scale. Each pitch is separated by seven semitones.

intervening, and comparison) consisted of three-letter words and the musical stimuli (initial, intervening, and comparison) consisted of three note chords. Importantly, trying to equate the two different types of stimuli is only the first step in constructing valid materials. It is equally important that the acoustic relationship between the initial and intervening stimuli be comparable on matched trials across both modalities (i.e., language – language and music – music). Previous research on the processing of language by the phonological loop has frequently manipulated this relationship by making the intervening stimuli either phonologically similar or dissimilar to the initial stimulus. However, how would the intervening stimuli be determined when the initial stimulus was a chord? Importantly, the use of intervening chords in the current experiment was systematically determined with reference to the Circle of Fifths (see Figure 1) – which visualizes the acoustic signature of chords on the harmonic scale – with some chords being more similar, and others being less so. Similarly, it was necessary to equate the comparison lures across stimuli in terms of how similar they were to the initial comparison stimuli – a task also accomplished using the Circle of Fifths. In essence, the Circle of Fifths served as a translator of sorts, that helped equate variations in our language stimuli (words) with variations in our musical stimuli (chords). Overall, a 2 (Initial Stimulus: language vs. music) 3 (Interference Type: matched, mismatched, silence) factorial design was used. It was predicted that conditions in which silence was the intervening stimulus would result in the greatest accuracy, as no interfering information would occupy the phonological loop. However, conditions in which intervening stimuli were words or chords should result in impaired performance relative to the silence conditions. Additionally, it was predicted that phonological interfering information would inhibit processing of phonological information more than would musical interfering information, and vice versa. This pattern of results would Ó 2018 Hogrefe Publishing

R. P. Atherton et al., Language, Music, and Working Memory

Table 1. Stimuli for the language conditions of the present study Intervening words

Comparison words

Target word

Dissimilar – e

Dissimilar – i

Dissimilar – o

Dissimilar – u

Same

Similar – First

Similar – Last

Cap

Few

Tic

Joy

Hub

Cap

Cab

Gap

Mad

Gel

Rib

Mop

Dux

Mad

Man

Had

Sat

Hem

Dim

Fog

Sup

Sat

Sax

Bat

Cad

Vet

Mix

Don

Cur

Cad

Can

Fad

Mat

Web

Zip

Cox

Rum

Mat

*Mal

Hat

Sap

Beg

Win

Hot

*Hug

Sap

*Sac

Lap

Cat

Sew

Pit

Cow

Bus

Cat

*Cal

Rat

Map

Fed

Jig

God

Mud

Map

*Maw

Tap

Sad

Pen

His

Low

Nut

Sad

*San

Pad

Notes. Every word is three letters long. Each target word consists of the same middle letter (“a”). Also, each beginning letter (“c,” “m,” and “s”) and ending letter (“p,” “d,” and “t”) is used three times without repeating a word. Comparison words can be the same or similar by either the first letter or last letter of the target word. “*” Refers to a word added by the author. While not from the list of Coltheart (1993) they are phonologically similar and should not present any added ease or difficulty in a discrimination task.

provide additional evidence in favor of shared processing mechanisms for phonological and musical information in the phonological loop of the multicomponent working memory model.

Method Participants Ninety-nine psychology students (35 males and 64 females) from a midsized Midwestern university participated in the study for course credit. The participants’ average age was 20.3 years, and self-identified as 80% Caucasian, 6% African-American, 12% Asian, and 2% Hispanic.

Materials Language stimuli consisted of monosyllabic words recorded and edited using the AudacityÓ software program (Audacity Team, 2012) on a mid-2007 2000 AppleÓ iMac computer. Spoken words were matched in pitch corresponding to the music conditions to control for any acoustical confounds. This was accomplished by specifying the desired frequency in AudacityÓ, which then resulted in a verbal output. Music stimuli were recorded on an iMac computer using chords generated from a Casio CTK-5000 keyboard. The amplitude and duration of the stimuli were then edited using the AudacityÓ software program. Language Stimuli Language stimuli were selected primarily from a list of phonologically similar and dissimilar words used by Coltheart (1993). Initially presented words (i.e., the target

Ó 2018 Hogrefe Publishing

words) consisted of nine monosyllabic, three-letter target words. Intervening language stimuli consisted of four randomly presented monosyllabic, three-letter phonologically dissimilar words. Comparison language stimuli were also randomly presented, and consisted of either the same word as the target word or one of two monosyllabic, three-letter phonologically similar words. These phonologically similar lure words differed from the target word in that either the first letter or last letter of the initial word was changed. For example, if the target word was “cap,” the lure word was either “gap” or “cab.” A list of all target and intervening words is provided in Table 1. Music Stimuli As discussed previously, musical information, versus tonal information, must reference specific musical notes, chords, etc. (vs. generic noise frequencies). This was accomplished by utilizing the organization presented in the Circle of Fifths. Music stimuli consisted of nine randomly presented major chords located on the circle of fifths. Intervening chords (four per trial) were displayed in a randomized order. These chords were chosen based on their distance from the target chord on the Circle of Fifths. Specifically, the intervening chords were the following number of turns from the target chord on the Circle of Fifths: four turns clockwise, five turns clockwise, four turns counterclockwise, and five turns counterclockwise (all distal). In this way, we created intervening chords that were dissimilar from the initially presented chord. Comparison chords were the same chord or a lure chord that was different from the target chord by either two turns clockwise or two turns counterclockwise (both proximal) from the target chord on the Circle of Fifths. In other words, the lure chords sounded similar to the initially presented chord in much the same way as the lure language stimuli sounded similar to

Experimental Psychology (2018), 65(1), 40–48

the initially presented language stimuli. A list of all target and intervening chords is provided in Table 2. Questionnaires Demographic information was collected using a write-in questionnaire that asked participants to identify their sex, age, race, and whether English was the participant’s primary spoken language. In addition, an adapted music background questionnaire originally developed by Dunleavy (2000) was administered to determine participants’ level of musical expertise.

Procedure Overview Participants were seated at separate tables in a campus computer laboratory. On the desk was a folder containing, in order, two consent forms, instructions, answer sheets, demographics questionnaire, music background questionnaire, and a debriefing script. Participants were instructed not to open the folder until asked to do so. They were then told to read only the top sheet, which was the informed consent form. After completing the consent form, the experimenter then read aloud the instructions as participants silently read along. Participants were told to ignore any intervening sounds or silence and the primary goal was to simply judge whether or not an initially presented word or musical chord (i.e., target stimulus) was the same as or different from a referenced word or musical chord (i.e., comparison stimulus). Participants then completed 12 practice trials followed by the 162 experimental trials. Upon completion participants filled out both the demographic and music background questionnaires and were then debriefed. Practice Trials Twelve practice trials were completed in an effort to familiarize participants with the experimental procedure. Of the twelve practice trials, two trials for each of the six different experimental conditions were included. Only data from participants who scored 7 or more correct were included in the final analyses, this resulted in the exclusion of 10 participants’ data (see Footnote 1). Thus, the practice trials served as a filter to eliminate participants who may have struggled with the task for one reason or another (e.g., inability to understand task instructions). As the practice trials were essentially identical to the experimental trials, a complete description of the task is provided in the next section. Experimental Trials Participants were tested in groups of 24–29. LogitechÓ Z523 (Logitech, Apples, Switzerland) speakers were placed at one end of the laboratory facing all of the participants. Experimental Psychology (2018), 65(1), 40–48

R. P. Atherton et al., Language, Music, and Working Memory

Table 2. Stimuli for the music conditions of the present study Intervening chords Presented chord

Comparison chords Same

Different+

Different

Notes. The “Different+” column refers to a clockwise rotation on circle of fifths and the “Different ” column refers to a counterclockwise rotation on circle of fifths. Additionally, “Different” refers to two turns on the circle of fifths in the indicated direction.

The speakers were connected to an 80 GB AppleÓ iPod Classic preloaded with the auditory stimuli. The volume was set at 70% of maximum volume and no participant indicated during debriefing that he or she could not clearly hear the stimuli. An initial tonal beep, signifying the start of each trial, was played for 750 ms, followed by 250 ms of silence. Next, the target stimulus was played for 750 ms, also followed by 250 ms of silence. The intervening stimulus phase of the trial lasted 6 s total and began and ended with 1 s of silence. Between the moments of silence was the onset of the intervening stimuli. Four separate chords or words (depending upon the condition) were each presented for 750 ms, followed by 250 ms of silence before the onset of the next stimulus. The intervening phase of the control condition was 6 s of silence. Following the last second of silence the comparison phase began. A comparison word or chord was played for 750 ms, followed by 250 ms of silence, as in the target stimulus presentation phase. Finally, 5 s of silence were provided for participants to circle their response (“same” or “different”) on their answer sheet. Each participant completed 162 trials (27 trials per condition). Every participant encountered each target, intervening, and comparison stimulus. Thus, each language and music target stimulus was at one point paired with an intervening stimulus of language, music, or silence and each comparison stimulus. To avoid fatigue, participants were given a 5-min break after 80 experimental trials.

Results In total, data from 78 participants were included for analysis (a link to the raw data is provided in the Electronic Ó 2018 Hogrefe Publishing

R. P. Atherton et al., Language, Music, and Working Memory

Supplementary Materials, ESM 1). These participants had minimal musical training, with the vast majority (N = 62) having less than 4 years of training and a sizeable proportion (N = 34) having no training whatsoever. Eleven participants who were determined to be “musicians” were excluded from analysis, as previous research has suggested that cognitive differences may exist between them and nonmusicians (e.g., Schlaug, 2003). A participant was coded as a musician if the following criteria were met: (1) 10+ years of formal (e.g., private) music lessons and (2) 15+ hours of practice on an instrument per week over the past month (these criteria are in accordance with the recommendations of Cohen, Evans, Horowitz, & Wolfe, 2011). Overall results are depicted in Figure 2. An omnibus General Linear Model Repeated-Measures Analysis of Variance (ANOVA) was conducted with the to-be-remembered word or chord loaded as factor 1 and the nature of overlap between stimulus modalities of the to-be-remembered stimulus and the interfering stimulus (matched, mismatched, or silence) loaded as factor 2. The proportion of correct responses per condition was used as the dependent variable. Results revealed a main effect of the to-beremembered stimulus, with accuracy being lower when individuals attempted to remember a chord (M = 78.90, SE = 1.09) as opposed to a word (M = 87.91, SE = 0.88), F(1, 77) = 76.06, p < .01, ηp2 = .50. Additionally, there was a main effect of the nature of overlap, F(2, 154) = 109.27, p < .01, ηp2 = .59. Post hoc analyses indicated that accuracy was lowest when the to-be-remembered stimulus and the interfering stimuli were matched (M = 77.26, SE = 0.96; e.g., language – language), followed by when the to-be-remembered stimulus and the interfering stimuli were mismatched (M = 84.02, SE = 0.99; e.g., language – music), and greatest accuracy was observed when the intervening stimulus was silence (M = 88.94, SE = 0.94; e.g., language – silence), ps < .01. Finally, there was a significant interaction between the to-be-remembered stimuli (i.e., word or chord) and the nature of overlap, F(2, 154) = 5.86, p < .01, ηp2 = .07, indicating that memory for the initial stimulus depended on the type of interference.

Memory for Chords If music and language share completely overlapping processing resources in working memory, it would be expected that both types of stimuli would interfere with the retention of musical stimuli relative to a silence control condition to a similar degree. However, if this overlap were only partial, it would be expected that musical interference would produce greater interference than language interference. The results appear to support this latter interpretation. Specifically, compared to the condition in which the Ó 2018 Hogrefe Publishing

Figure 2. Accuracy at Identifying the Target Stimulus. The first letter of the pairs on the x-axis denotes the presented and comparison stimulus type and the second letter represents the intervening stimulus type, with the accuracy percentage represented on the y-axis. LS = language-silence; LL = language-language; LM = language-music; MS = music-silence; MM = music-music; ML = music-language.

intervening stimulus was a moment of silence (M = 85.71, SD = 10.42), accuracy was lower when the intervening stimuli were both chords (M = 72.70, SD = 11.95), t(77) = 9.31, p < .01, and words (M = 78.30, SD = 12.33), t(77) = 6.45, p < .01. Importantly, accuracy when the intervening stimuli were chords (M = 72.70, SD = 11.95) was significantly lower than when the intervening stimuli were words (M = 78.30, SD = 12.33), t(77) = 4.55, p < .01. Thus, while both linguistic and musical stimuli interfered with the retention of musical stimuli, this deficit was greater when the interference and the to-be-remembered information matched (i.e., music – music).

Memory for Words Fitting with our initial hypotheses, a similar pattern emerged in terms of participants’ memory for the original language stimuli. Again, as compared to the condition in which the intervening stimulus was a moment of silence (M = 92.17, SD = 8.97), accuracy was significantly lower when the intervening stimuli were both words (M = 81.81, SD = 9.55), t(77) = 10.71, p < .01, and chords (M = 89.74, SD = 8.20), t(77) = 3.96, p < .01. In addition, accuracy when the intervening stimuli were words (M = 81.81, SD = 9.55) was significantly lower than when the intervening stimuli were chords (M = 89.74, SD = 8.20), t(77) = 8.54, p < .01. Once again, while both music and language interfered with language retention, the impairment was greater when the initial stimulus matched the interference (i.e., language – language). Overall, we interpret this pattern of data as evidence in favor of the notion that language and music share some degree of processing resources in working memory. Experimental Psychology (2018), 65(1), 40–48

Discussion The primary aim of the current study was to explore whether linguistic and musical information share storage in working memory. This issue was examined by exploring the impact of intervening information on participants’ ability to retain linguistic and musical information in immediate memory. Importantly, the current investigation incorporated the use of nonlanguage stimuli organized in a musical context (i.e., chords on the Circle of Fifths) into a cross-modal interference paradigm. The use of these stimuli represents an advantage over previous studies, which tended to rely on simple tones that lacked the complexity of music (e.g., Deutsch, 1970). Overall, greatest accuracy on the forced-choice discrimination task was observed when the intervening stimulus was silence, followed by the “mismatched” interference conditions (i.e., language – music and music – language) and then the “matched” interference conditions (i.e., language – language and music – music). It is important to note that this pattern of results was observed when the to-be-remembered stimuli were both words and chords. This means that each type of stimulus not only interferes with retention within its modality (e.g., language – language), but also across modalities (e.g., language – music). These findings extend previous research (e.g., Williamson et al., 2010) by employing a paradigm that more directly allowed for the assessment of the degree of interference produced when the nature of the test and interfering stimuli were not only matched, but mismatched as well – suggesting a likelihood of shared processing resources for both linguistic and musical stimuli. In addition to the similar pattern of performance observed across the silent, matched, and mismatched conditions for to-be-remembered chords and words, a significant interaction was also revealed. This interaction was mainly driven by a large relative difference in accuracy for the two mismatched conditions, with words disrupting retention of a to-be-remembered chord to a much greater degree than chords disrupted retention of a word. One way in which to interpret this result is through an expertise account (e.g., Gauthier, Skudlarski, Gore, & Anderson, 2000). It is possible that because all participants are experts in processing language, as opposed to music (as musicians were excluded from analysis), that words may be more likely to be automatically processed by the phonological loop, and thus compete with other information that is being actively processed within this structure. Music, on the other hand, may not have direct access to the phonological loop, which may also partially explain why performance for to-be-remembered chords is poorer across all three conditions (silence, matched, and mismatched). Based upon this expertise account, it would be predicted that musicians, who are experts at processing both music and language, Experimental Psychology (2018), 65(1), 40–48

R. P. Atherton et al., Language, Music, and Working Memory

would show similar levels of interference for both chord and word to-be-remembered stimuli for the three different conditions, as music stimuli would have more direct access to the phonological loop. Future research, perhaps incorporating articulatory suppression manipulations, would be needed to determine how musical stimuli access the phonological loop. At a broad level, the finding that musical information appears to be processed by a system seemingly so important for linguistic information may be analogous to what is observed within the face processing literature, wherein cortical resources that are reliably involved in the processing of faces can also be co-opted for processing classes of stimuli that the perceiver possesses expertise within. For example, it has been demonstrated that processing stimuli that one is an expert in (e.g., birds, cars) can activate regions that are also involved in face processing (Gauthier et al., 2000). Further, an increased degree of expertise in processing novel stimuli results in greater cortical activation in regions responsible for face processing (Gauthier, Tarr, Anderson, Skudlarski, & Gore, 1999). Thus, it could be that while the phonological loop is always involved in processing linguistic stimuli, it can be recruited for other classes of stimuli as well. Interestingly, this analogy may extend further than just a topical similarity. As mentioned in the Introduction, if two different stimuli share at least partial overlap in processing, it raises important questions about what core attribute is shared by those stimuli. The complex, multidimensional code associated with both language and music seems in many ways akin to the type of configural processing associated with face recognition – wherein multiple individual components are combined in order to create a broader, supra-level representation. Indeed, there is evidence of expertise playing a role in the processing of linguistic stimuli, especially with respect to its presentation through the visual modality (for a review, see McCandliss, Cohen, & Dehaene, 2003). However, more research is needed to fully understand how expertise can affect the cognitive and neural components involved in processing linguistic and musical information within the context of working memory. Regardless of the aforementioned interaction, the broad pattern of results reported above is also congruent with the predictions made by other theoretical accounts of immediate memory processing. For example, the perceptual-gestural account (e.g., Jones, Hughes, & Macken, 2006; Jones, Macken, & Nicholls, 2004) argues for a system that is functionally different than that proposed by Baddeley and colleagues. According to this framework, less emphasis is placed on specific stores such as the phonological loop. Instead, the model focuses on streams of information and maintains that the processing of language-based stimuli is purely articulatory, rather than phonological. Ó 2018 Hogrefe Publishing

R. P. Atherton et al., Language, Music, and Working Memory

Interestingly, this framework appears to be able to explain a critical finding that has been used to argue for the phonological nature of immediate memory processing. Specifically, the phonological similarity effect (Conrad & Hull, 1964) appears to disappear under articulatory suppression when verbal stimuli are presented visually as opposed to auditorily. It has been argued that articulatory suppression prevents visually presented information from being recoded and thus having access to the phonological loop. Because auditorily presented information has direct access to the phonological loop, the phonological similarity effect will continue to occur, even under conditions of suppression. However, Jones and colleagues (e.g., Jones et al., 2004, 2006) have argued that the persistence of the phonological similarity effect for auditorily presented verbal stimuli under suppression does not necessarily imply that any immediate memory store is phonological in nature. In a series of studies, they were able to eliminate the phonological similarity effect by presenting suffixes at the end of a series of to-be-remembered letters. The authors argue that the presence of the suffix disrupts the normal auditory grouping that occurs; particularly for items at the end of the list. Interestingly, the phonological similarity effect can be reconstituted if the added suffix is acoustically different than the to-be-remembered letters (i.e., spoken in a different voice). In essence, the authors argue that boundaries (either at the beginning or end of an auditory stream) are cues that help to group information together – resulting in increased interference (i.e., the phonological similarity effect). According to this account then, the phonological similarity effect can be explained by principles of auditory grouping, as opposed to postulating a phonologically based processing store. The results of the current investigation appear consistent with this theoretical account of processing in immediate memory. Specifically, consider the main finding of the current investigation: that modality matched stimuli (e.g., language – language) resulted in the greatest level of interference. According to the perceptualgestural account, when the comparison and intervening stimuli are the same, they are more likely to be perceptually grouped together, thus resulting in a greater amount of interference. It is worth noting, however, that the current study was not designed to distinguish between different models of immediate memory processing and thus future research would be needed to draw broader theoretical conclusions. From our perspective, the critical conclusion is that whatever the underlying nature of the processing system, it appears that it processes both musical and linguistic stimuli to some degree. The results of the current study are also in accord with evidence presented from neuroimaging studies that have investigated whether the active processing of verbal and tonal information relies on overlapping neural substrates Ó 2018 Hogrefe Publishing

(Hickok, Buchsbaum, Humphries, & Muftuler, 2003; Koelsch et al., 2009; Schulze, Zysset, Mueller, Friederici, & Koelsch, 2011). Results from these studies suggest a large degree of overlap in terms of the cortical resources involved in processes related to working memory for both classes of stimuli. For example, in tasks believed to rely on working memory, both tonal and verbal information in nonmusicians activated premotor cortex in addition to portions of Broca’s area in all three studies, while additional overlap was observed in the inferior parietal lobe and area Spt/planum temporale in two of the three (Hickok et al., 2003; Koelsch et al., 2009; Schulze et al., 2011; for a review, see Schulze & Koelsch, 2012). The observed cortical overlap involved in processing the two different classes of information suggests that similar cortical resources may underlie the active processing of both verbal and tonal information. Ultimately, however, the research to date has only begun to scratch the surface in terms of understanding the complex interaction between the processing of linguistic and musical information and how that may relate to the theoretical structures associated with processing in immediate memory. We believe that the experimental approach taken in the current investigation represents an important step forward. Using a behavioral paradigm, we were able to demonstrate cross-domain interference between linguistic and musical stimuli that were equated on a number of important dimensions. Future research will be needed to understand the degree of neural overlap that occurs, the theoretical structures involved, and the specific commonalities that are shared between linguistic and musical stimuli that result in the recruitment of similar brain regions. Acknowledgment The research reported in this article is based on a Master’s Thesis conducted by Ryan P. Atherton in partial fulfillment of the requirements for the MS degree. Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1618-3169/a000388 ESM 1. Data (.sav) Raw data of the study.

References Audacity Team. (2012). Audacity (Version 2.0) [Computer program]. Retrieved from https://www.audacityteam.org/ Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory (Vol. 8, pp. 47–89). New York, NY: Academic Press.

Experimental Psychology (2018), 65(1), 40–48

Baddeley, A., Lewis, V., & Vallar, G. (1984). Exploring the articulatory loop. The Quarterly Journal of Experimental Psychology, 36, 233–252. https://doi.org/10.1080/14640748408402157 Baddeley, A. D., Thomson, N., & Buchanan, M. (1975). Word length and the structure of short-term memory. Journal of Verbal Learning and Verbal Behavior, 14, 575–589. https://doi.org/10/ 1016/S0022-5371(75)80045-4 Brown, S., Martinez, M. J., & Parsons, L. M. (2006). Music and language side by side in the brain: A PET study of the generation of melodies and sentences. European Journal of Neuroscience, 23, 2791–2803. https://doi.org/10.1111/j.14609568.2006.04785.x Cohen, M. A., Evans, K. K., Horowitz, T. S., & Wolfe, J. M. (2011). Auditory and visual memory in musicians and nonmusicians. Psychonomic Bulletin & Review, 18, 586–591. https://doi.org/ 10.3758/s13423-011-0074-0 Coltheart, V. (1993). Effects of phonological similarity and concurrent irrelevant articulation on short-term memory recall of repeated and novel word lists. Memory and Cognition, 21, 539–545. https://doi.org/10.3758/BF03197185 Conrad, R., & Hull, A. J. (1964). Information, acoustic confusion, and memory span. British Journal of Psychology, 55, 429–432. https://doi.org/10.1111/j.2044-8295.1964.tb00928.x Deutsch, D. (1970). Tones and numbers: Specificity of interference in immediate memory. Science, 168, 1604–1605. https://doi. org/10.1126/science.168.3939.1604 Dunleavy, D. H. (2000). The effects of music exposure on spatialtemporal reasoning: A comparison of musicians and nonmusicians. Unpublished undergraduate honor’s thesis, Middlebury College, Middlebury, Vermont. Gauthier, I., Tarr, M. J., Anderson, A. W., Skudlarski, P., & Gore, J. C. (1999). Activation of the middle fusiform “face area” increases with expertise in recognizing novel objects. Nature Neuroscience, 2, 568–573. https://doi.org/10.1038/9224 Gauthier, I., Skudlarski, P., Gore, J. C., & Anderson, A. W. (2000). Expertise for cars and birds recruits brain areas involved in face recognition. Nature Neuroscience, 3, 191–197. https://doi. org/10.1038/72140 Hickok, G., Buchsbaum, B., Humphries, C., & Muftuler, T. (2003). Auditory-Motor interaction revealed by fMRI: Speech, music, and working memory in area Spt. Journal of Cognitive Neuroscience, 15, 673–682. https://doi.org/10.1162/jocn.2003.15.5.673 Jones, D. M., Hughes, R. W., & Macken, W. J. (2006). Perceptual organization masquerading as phonological storage: Further support for a perceptual-gestural view of short-term memory. Journal of Memory and Language, 54, 265–281. https://doi.org/ 10.1016/j.jml.2005.10.006 Jones, D. M., Macken, W. J., & Nicholls, A. P. (2004). The phonological store of working memory: Is it phonological and is it a store? Journal of Experimental Psychology: Learning, Memory, and Cognition, 30, 656–674. https://doi.org/10.1037/02787393.30.3.656 Koelsch, S., Gunter, T. C., von Cramon, D. Y., Zysset, S., Lohmann, G., & Friederici, A. D. (2002). Bach speaks: A cortical

Experimental Psychology (2018), 65(1), 40–48

R. P. Atherton et al., Language, Music, and Working Memory

“language-network” serves the processing of music. NeuroImage, 17, 956–966. https://doi.org/10.1006/S1053-8119(02) 91154-7 Koelsch, S., Schulze, K., Sammler, D., Fritz, T., Muller, K., & Gruber, O. (2009). Functional architecture of verbal and tonal working memory: An fMRI study. Human Brain Mapping, 30, 859–873. https://doi.org/10.1002/hbm.20550 McCandliss, B. D., Cohen, L., & Dehaene, S. (2003). The visual word form area: Expertise for reading in the fusiform gyrus. Trends in Cognitive Neuroscience, 7, 293–299. https://doi.org/ 10.1016/S1364-6613(03)00134-7 Posner, M. I., & Snyder, C. R. (2004). Attention and cognitive control. In D. Balota & E. Marsh (Eds.), Cognitive psychology: Key readings (pp. 205–223). Hove, UK: Psychology Press, Taylor & Francis Books. Schlaug, G. (2003). The brain of musicians. In P. Peretz & R. Zatorre (Eds.), The cognitive neuroscience of music (pp. 366–381). Oxford, UK: Oxford University Press. Schulze, K., & Koelsch, S. (2012). Working memory for speech and music. Annals of the New York Academy of Sciences, 1252, 229–236. https://doi.org/10.1111/j.1749-6632.2012.06447.x Schulze, K., Zysset, S., Mueller, K., Friederici, A. D., & Koelsch, S. (2011). Neuroarchitecture of verbal and tonal working memory in nonmusicians and musicians. Human Brain Mapping, 32, 771–783. https://doi.org/10.1002/hbm.21060 Semal, C., Demany, L., Ueda, K., & Halle, P. A. (1996). Speech versus nonspeech in pitch memory. Journal of the Acoustical Society of America, 100, 1132–1140. https://doi.org/10.1121/ 1.416298 Williamson, V., Baddeley, A., & Hitch, G. (2010). Musicians’ and nonmusicians’ short-term memory for verbal and musical sequences: Comparing phonological similarity and pitch proximity. Memory and Cognition, 38, 163–175. https://doi.org/ 10.3758/MC.38.2.163 Wilson, M., & Emmorey, K. (1998). A “word length effect” for sign language: Further evidence for the role of language in structuring working memory. Memory and Cognition, 26, 584–590. https://doi.org/10.3758/BF03201164

Received January 6, 2017 Revision received September 2, 2017 Accepted September 20, 2017 Published online February 8, 2018

Quin M. Chrobak Department of Psychology University of Wisconsin-Oshkosh 800 Algoma Blvd Oshkosh, WI 54901 USA chrobakq@uwosh.edu

Ó 2018 Hogrefe Publishing

Social Psychology

ISSN-Print 1864-9335 ISSN-Online 2151-2590 ISSN-L 1864-9335 6 issues per annum (= 1 volume)

Subscription rates (2018) Libraries / Institutions US $478.00 / € 374.00 Individuals US $223.00 / € 159.00 Postage / Handling US $24.00 / € 18.00

www.hogrefe.com

nline free o issue le samp

Editor-in-Chief Kai Epstude University of Groningen, The Netherlands

Editorial Office Wim Meerholz University of Groningen, The Netherlands

Associate Editors Anna Baumert, Germany Marco Brambilla, Italy Adam Fetterman, USA Ilka Gleibs, United Kingdom Michael Häfner, Germany

Hans J. IJzerman, The Netherlands Ulrich Kühnen, Germany Toon Kuppens, The Netherlands Ruth Mayo, Israel Kim Peters, Australia

About the Journal 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 was published until volume 38 (2007) as the Zeitschrift für Sozialpsychologie (ISSN 0044-3514). Drawing on over 30 years of experience and tradition in publishing high-quality, innovative science as the Zeitschrift für Sozialpsychologie, Social Psychology has an internationally renowned team of editors and consulting editors from all areas of basic and applied social psychology, thus ensuring that the highest international standards are maintained.

Manuscript Submissions All manuscripts should be submitted online at www.editorialmanager.com/sopsy, where full instructions to authors are also available. Electronic Full Text The full text of the journal – current and past issues (from 1999 onward) – is available online at econtent.hogrefe.com/loi/zsp (included in subscription price). A free sample issue is also available there. Abstracting Services The journal is abstracted / indexed in Current Contents / Social and Behavioral Sciences (CC / S&BS), Social Sciences Citation Index (SSCI), PsycINFO, PSYNDEX, ERIH, Scopus, and EMCare. Impact Factor (Journal Citation Reports®, Clarivate Analytics): 2016 = 2.602

European Journal of Health Psychology

in Now sh Engli

Editor-in-Chief Claus Vögele, University of Luxembourg, Luxembourg Editorial Assistant Nicole Knoblauch, Luxembourg

ISSN-Print 2512-8442 ISSN-Online 2512-8450 ISSN-L 2512-8442 4 issues per annum (= 1 volume)

Subscription rates (2018) Libraries / Institutions US $343.00 / € 268.00 Individuals US $120.00 / € 94.00 Postage / Handling US $16.00 / € 12.00

www.hogrefe.com

Associate Editors Verena Klusmann, Bremen, Germany Arnold Lohaus, Bielefeld, Germany Britta Renner, Konstanz, Germany Christel Salewski, Hagen, Germany Silke Schmidt, Greifswald, Germany Heike Spaderna, Trier, Germany

About the Journal The European Journal of Health Psychology was founded to provide a platform for research in health psychology, and for its application in a wide range of contexts. Health psychology is the scientific discipline within psychology that aims to promote and preserve health, to prevent diseases and contribute to their treatment by identifying disease-relevant aetiological processes, and to improve health provision.

Call for Papers The European Journal of Health Psychology invites you and/or your working group to submit papers on all aspects of the field!

The European Journal of Health Psychology strives to promote theory and practice in the analysis of psychological approaches to health and disease. Its aim is, therefore, to publish high quality empirical or experimental research as well as sound practice-oriented articles, current methodological developments, and comprehensive critical reviews of the scientific literature.

Electronic Full Text The full text of the journal – current and past issues (from 1999 onward) – is available online at econtent.hogrefe.com/loi/zgp (included in subscription price). A free sample issue is also available there.

The journal has been publishing high-quality, innovative research since 1993 (until 2016 as Zeitschrift für Gesundheitspsychologie, ISSN 0943-8149).

Manuscript Submissions All manuscripts, including Electronic Supplementary Material (ESM), should be submitted online at www.editorialmanager.com/zgp, where full instructions to authors are also available.

Abstracting Services The journal is abstracted / indexed in Social Sciences Citation Index (SSCI), Social Scisearch, Journal Citation Report/Social Sciences Edition, PsycInfo, PsycLit, PsyJOURNALS, PSYNDEX, Scopus, IBZ, IBR, and European Reference List for the Humanities (ERIH). Impact Factor (Journal Citation Reports®, Clarivate Analytics): 2016 = 0.909

Short Research Article

Contagion via Magical Thinking and via Mere Proximity Differences as a Function of Target Type lu, and Brian E. Rabinovitz Lennea R. Bower, Zehra F. Peynirciog Department of Psychology, American University, Washington, DC, USA

Abstract: People show an irrational dislike for objects that were once contaminated or had come into contact with an undesirable person, even if they are currently indistinguishable from other similar objects. To date, such negative contagion within the magical thinking literature has been shown only with inanimate objects. We addressed a boundary condition to see if it also extended to animate targets (dogs and children) while teasing out mere-proximity effects that would predict a similar contagion in the case of children. We used two different types of contagion, one based on proximity and one based on self-information. We found that magical thinking did extend to dogs but not to children when not confounded by mere-proximity effects. Also, contagion was less strong in the case of animate targets, but pity was not related to either this reduction or to the disappearance of the effect with children. Keywords: magical thinking, mere proximity, contagion, pity, empathy, decision making

People often rely on heuristics to make complicated decisions, which may sometimes lead to irrational decisions (e.g., Garcia-Retamero & Dhami, 2009). Magical thinking is a subset of heuristics based on resemblances or associations between objects (e.g., Rozin, Millman, & Nemeroff, 1986). One of the most-studied such heuristics is contagion, which can be summarized as “once in contact, always in contact” (Rozin & Nemeroff, 2002). This heuristic simplifies decisions by basing them not on the object in question, but rather on its past associations or connections to liked or disliked (safe or dangerous) people or things. Although it can be useful (e.g., it can help people avoid foodstuffs that have come in contact with known poisons), it can also be applied when the information is logically irrelevant but emotionally salient (e.g., a now perfectly cleaned foodstuff that had once touched a disgusting entity). Such thinking has been shown using a variety of sources of contagion and even when participants acknowledge the logical inconsistencies in their decisions (Hejmadi, Rozin, & Siegal, 2004; Rozin, Markwith, & Ross, 1990; Rozin & Nemeroff, 2002). Further, negative contagion, in which the target is rated less favorably because of its past contacts, has been shown to be more potent than positive contagion (e.g., Rozin & Royzman, 2001). Thus, in the present paper we focus on only negative contagion.

Ó 2018 Hogrefe Publishing

To date, contagion in magical thinking studies has involved only objects as targets, such as hairbrushes or sweaters. A conceptually similar idea within social psychology is the mere-proximity effect (e.g., Hebl & Mannix, 2003; Mehta & Farina, 1988; Neuberg, Smith, Hoffman, & Russell, 1994) and involves people as targets. Being related to or otherwise associated with others who are thought of as undesirable can lead to negative stereotypical assumptions about the target, as well, and this can happen even when there is no logical connection. For instance, Hebl and Mannix have shown that people were evaluated negatively when they were simply sitting next to another person with an undesirable characteristic, in this case, obesity. In some sense then, the mere-proximity effect in which negative information about associated others taint decisions about human targets can be viewed as a subset of magical thinking (cf. Lindeman & Svedholm, 2012). The reach of magical thinking is broader, however, in that the negative information does not need to be associations to others. Any negative association with any event, thing, or experience can also lead to contagion because the idea is that the essence of the undesirable occurrence is assumed to have been transmitted to the target. Thus, magical thinking studies, albeit to date confined to object targets, use both

Experimental Psychology (2018), 65(1), 49–60 https://doi.org/10.1027/1618-3169/a000389

“proximity” association (connections to undesirable people) and what we will call “self” association (having once experienced events or characteristics no longer present, as in the now perfectly cleaned foodstuff that had once touched a disgusting entity), and do not distinguish between the two types as long as there is negative contagion or what is generally thought of as an irrational aversion based on irrelevant information. In this study, we explore a boundary condition of magical thinking, whether contagion would extend to people (in our case, children) or indeed also to other animate targets, such as dogs, while teasing out mere-proximity contagion that occurs with people as a confounding variable. One could speculate that in magical thinking, with animate targets, such as children or dogs, the very same piece of information that leads to negative contagion might also lead to something akin to pity which could counter contagion, just as “money” has been shown to alleviate contagion from object targets (e.g., Rozin, Grant, Parker, & Weinberg, 2007). Targets that evoke pity garner empathy (Cikara & Fiske, 2011b) and can trigger altruistic behaviors (Batson & O’Quin, 1983; Carrera et al., 2012). Altruistic behaviors can also be initiated by biological imperatives that trigger responding to distressed others, especially the young or vulnerable (Preston, 2013). As such, they may lessen the negative feelings caused by contagion. The aim of Experiment 1 was to replicate traditional magical thinking study procedures to see if the contagion effect would be confined to inanimate targets or at least be alleviated with animate targets. Because we suspected pity might play a role in any observed modulation, we picked shelter dogs and children as our animate targets, since pity has been shown to be felt most strongly for targets that are high in warmth and low in competence (e.g., Fiske, Cuddy, & Glick, 2007). Participants made liking and preference decisions about inanimate targets (e.g., skateboard) as well as animate targets comprising children and dogs. The contagion manipulation was whether one piece of information in the entire description, the profession of a parent of a child or owner of a dog or object, was usually seen as undesirable or not (e.g., used-car salesman vs. police officer). The purpose of asking for two different decisions, one pertaining to liking and the other to spending a weekend with that target, was to see if a difference would emerge between just a momentary decision and a decision about being involved in some hypothetical action. Such a difference is immaterial in traditional magical thinking studies that have usually asked for preference decisions, since the targets have invariably been inanimate objects. However, with children and dogs, the critical information might have colored the preference to interact with them, but not affected a feeling of liking in the abstract. Contagion was evident regardless of decision or target type, Experimental Psychology (2018), 65(1), 49–60

L. R. Bower et al., Contagion

however, although the effect in liking was somewhat weaker with animate targets than that with objects. The aim of Experiment 2 was twofold. First, we wanted to control for the effects of mere proximity. In Experiment 1, to mimic the typical magical thinking studies, we used undesirable previous-owner information (and since children could not have “owners,” parent information). This specific manipulation was confounded by mere proximity, however, and, although this confound was irrelevant for object targets used in the magical thinking studies to date, it may have influenced the results with animate targets, especially with children. That is, we may have obtained the contagion with animate targets not because of the illusory assumption that the essence of the parent/owner had passed onto the target but because of stereotyping that was triggered by the proximal negative information (cf. Mehta & Farina, 1988). Thus, we tested whether the effect seen in Experiment 1 might vary, at least with children, as a function of the type of manipulation that led to the contagion – that is, whether the negative information was supplied through proximity association or self-association. To this end, we had two types of negative and positive associative information for both animate and inanimate targets, one type replicating those in Experiment 1 and pertaining to parent or previous owner (e.g., used-car salesman vs. police officer) and another type pertaining to the target itself (e.g., the child/dog fell in the sewer and had to get a tetanus shot, the object was dropped in the sewer and had to be sterilized, or the child/dog played on the carpet with his toys, the object was played with on the carpet). One conjecture was that, to the extent that pity plays a role in countering contagion, it might be more likely to do so in animate targets when the cause of contagion is a self-association rather than a proximity association. That is, whereas inanimate targets should not garner pity and thus the two types of contagion should be equivalent and indistinguishable in objects, animate targets might garner more pity when the negative information is about themselves rather than about those they are proximal to. Also, in Experiment 1, pity was an “assumed” factor. The second aim of Experiment 2 was to introduce pity as a measure. We explicitly asked for pity decisions, although they were removed from the initial liking decisions so as not to contaminate the liking decisions. In this way, we could see whether pity was indeed related to any alleviation of the contagion effects with animate targets, and whether the relationship would be different as a function of type of contagion (proximity association vs. self-association). Further, given that preference decisions in Experiment 1 could be construed to be not entirely illogical for children and dogs, and indeed did not show any differences as a function of target type, we asked only for liking decisions. Ó 2018 Hogrefe Publishing

L. R. Bower et al., Contagion

Experiment 1 Method Participants A total of 120 adults (28 males and 92 females), American University students (N = 14) or college-educated community members (N = 106), participated in the study for extra credit toward psychology courses or for fun. We placed participants into two age categories, with 45 acting as the dividing age between them because it corresponded most closely to the empathy drop-off age found by Schieman and Gundy (2000). There were thus 69 participants (20 males and 49 females) who were between the ages of 18 and 44, and 51 participants (8 males and 43 females) who were 45 or older. We took note of gender and age because although no differences have been shown in magical thinking, females and younger adults typically report more empathy than males and older adults, respectively (e.g., Eisenberg & Lennon, 1983; Schieman & Gundy, 2000). Materials, Design, and Procedure This study was conducted as an online survey. There were 18 pairs of photographs of children, dogs, and objects, all with accompanying texts. Photographs of the children and objects were selected from the free items available on the stock photograph site Stockvault (http://www.stockvault. com), and photographs of the dogs were selected from Petfinder (http://www.petfinder.com). All photographs were resized to be equivalent. All targets were presented in “like” pairs (e.g., child-child) side by side, and designated as Targets A and B. Target pairs were matched for basic physical features (e.g., gender, age, coloring/breed) or object type. For example, a female Caucasian child was presented with another female Caucasian child, a retriever dog was presented with another retriever dog, and a toy was presented with another toy. The text for each target included four pieces of information, of which one was the critical contagion information: owner’s or parent’s profession. The professions chosen for positive and negative information were selected from published lists of most admired and most distrusted professions (http://943thepoint.com, http://www.gallup.com, http:// www.pewforum.org, http://www.onlinedegreeprograms. com, http://www.scientificmarketingandadvertising.com). The other pieces of information were inherently about the target itself and were always positive or neutral. They comprised a subset of several categories and were all selected to be consistent with the photograph and similar to the information for the other target in the pair. Examples of materials from both experiments are presented in Appendix A and the lists of professions and other information in Appendix B. The order in which the four pieces of information were Ó 2018 Hogrefe Publishing

presented was randomized across targets, but kept the same for each target. In order to minimize how evident the contagion information manipulation might be to participants, half of all pairs for each target type were red-herring pairs that were similar in every way except that the descriptions did not contain any contagion information. Full counterbalancing measures are presented in Appendix C. Target pairs were transformed into images and uploaded into SurveyMonkey (http://www.surveymonkey.com). Each pair was presented in a different random order for each participant. The survey began with a consent page followed by an instruction/example page with cartoon figures to introduce participants to the procedure. The 18 target pairs followed. Each target pair was followed by the same questions regarding (a) which one they would prefer to care for or be in charge of for a weekend (with five responses ranging from strongly preferring A to strongly preferring B) and (b) how they felt about each target (with seven responses ranging from negatively-not liking at all to positively-liking a lot). It is important to note that the middle option in both cases was “neutral” in order to minimize demand characteristics by giving participants the ability to not have to make a choice and not be forced to fall into a magical thinking mode. Thus, any contagion that emerged would do so under more stringent conditions than in traditional magical thinking experiments. Participants responded by marking the radio-button next to their choice. The final page of the survey gathered demographic information including age and gender.

Results The results are summarized in Table 1 (see Electronic Supplementary Materials, ESM 1 for raw data). Likeability Each target was given a separate likeability rating, and these ratings were analyzed in a 2 (Contagion Information) 3 (Target Type) 2 (Age) 2 (Gender) mixed-design analysis of variance (ANOVA). Contagion information and target type were within-subject variables and age and gender were between-subject variables. All post hoc t-tests involved Tukey corrections. There was a main effect of contagion information, F(1, 115) = 22.54, MSE = 9.00, p < .001, ηp2 = .16. Indeed, a negative contagion effect was present in each of the target types separately, as well, ts(119) = 4.35 (p < .001), 3.18 (p = .002), and 5.19 (p < .001) for children, dogs, and objects, respectively. There was also a main effect of target type, F(2, 230) = 26.16, MSE = 31.87, p < .001, ηp2 = .19. Likeability for objects was lower than that for dogs or children, ts(119) = 8.69 and 6.87, respectively, ps < .001; interestingly, likeability for dogs was higher than that for Experimental Psychology (2018), 65(1), 49–60

L. R. Bower et al., Contagion

Table 1. Percentage of selection and mean likeability ratings (1–7; with standard deviation in parentheses) for children, dogs, and objects as a function of contagion valence (positive or negative) Target type

Preference

No preference

Positive info

Negative info

Child

28.6

11.6

Dog

27.7

14.4

Object

45.0

22.2

children, t(119) = 2.37, p = .02. More importantly, though a very small effect, there emerged an interaction between contagion information and target type, F(2, 230) = 4.27, MSE = 1.31, p = .036, ηp2 = .04. Contagion was greater for objects than for either dogs or children, ts(119) = 3.47 (p < .001) and 3.02 (p = .003), respectively, and there was no difference between dogs and children (p > 0.10). There were no effects of age or gender (ps > .10). Preference Unlike likeability, where each target was given a separate rating, preference measure involved choosing one of the target pairs over the other or indicating no preference on a 5-point scale. Each preference was thus quantified as a single number ranging from 2 (negative-information target strongly preferred) to 2 (positive-information target strongly preferred). Preference for the positive-information target, indicated by a positive preference score, is thus consistent with a contagion effect. The mean scores were 0.24 for children, 0.18 for dogs, and 0.36 for objects. Although small, each of these differences was significantly different from 0, which would have indicated no preference, ts(119) = 5.06, 4.26, and 5.46, respectively, all ps < .001. Unlike with likeability ratings, there were no differences between preference scores for animate and inanimate targets, with children and dogs compared to objects separately or as collapsed into a single animate category (all ps > .10). There were also no gender or age differences, and no difference in the magnitudes of the contagion effect between inanimate and animate targets (all ps > .10).

Experiment 2 The aims of Experiment 2 were to replicate Experiment 1 with the liking decisions and also add a different category of contagion, one that was not based on proximity to another person. Because we eliminated preference decisions, we presented each target on its own rather than as part of a pair, and thus each judgment was made independently of any comparison target. Further, we also asked for pity ratings after the initial liking ratings were Experimental Psychology (2018), 65(1), 49–60

Likeability Positive info

Negative info

No info

59.7

5.58 (1.17)

5.36 (1.21)

5.71 (1.12)

57.7

5.81 (1.13)

5.65 (1.19)

5.89 (1.16)

32.7

5.14 (1.04)

4.58 (1.14)

5.26 (0.97)

given. Because there were no age or gender effects in Experiment 1, we did not take those factors into consideration.

Method Participants A total of 72 American University students participated in the experiment for extra credit toward psychology courses. Materials, Design, and Procedure The photographs were the same as in Experiment 1. However, each target was presented singly with its accompanying text. For any given participant, for a third of the targets of each type, the text contained the same type of contagion information as in Experiment 1 (indeed a subset of exactly the same professions), half negative and half positive – the proximity association condition. For another third, the text contained another type of contagion information (about the target itself) that would not lead to mere-proximity effects, again half negative and half positive – the selfassociation condition. For the final third, the text contained no contagion information – as in Experiment 1, the purpose was to minimize how evident the contagion information manipulation might be to participants. All targets were presented randomly with respect to association type, the presence (or valence) of contagion information, and target type. The experiment used a 3 (Target Type) 2 (Association Type) 2 (Contagion Information) and entirely within-participant design, requiring 12 counterbalancing groups for systematic rotation of descriptions through the pictures. Thus, all photographs were shown in the same randomly selected order to all participants, but across the counterbalancing groups, each description appeared for each target type equally often and each target appeared in the positive and negative contagion as well as association type conditions equally often. Participants were tested individually or in group settings. They were first given a booklet, each page showing 4–6 target items (photographs with their accompanying texts and a liking rating scale below each item). The first page of the booklet contained the instructions and a practice item with a cartoon character and a neutral description Ó 2018 Hogrefe Publishing

L. R. Bower et al., Contagion

to familiarize the participants with the task. Next, the participants went through the booklet at their own pace and rated each item for “liking” on a scale from 1 (= not at all) to 10 (= very much). These booklets were then collected and another set of booklets distributed. This second set of booklets was exactly the same as before but contained three targets on each page, with four questions following each target to be rated on a scale of 1–10: (1) How typical does this child (dog/object) look? (2) How intelligent does this child (dog) look – or how expensive does this object look? (3) Do you have any feelings of pity for this child (dog/ object)? (4) How comfortable would you feel if you had to interact with this child (dog/object)? The critical question was the third question regarding pity. The first page was reserved for instructions and the practice item, and the participants went through the booklet at their own pace.

Results The results of main interest are summarized in Table 2 (see ESM 2 for raw data). Because the two dependent variables of interest, liking and pity ratings, were not correlated (r = .06, p > .10), we conducted a 2 (Contagion Information) 2 (Association Type) 3 (Target Type) repeatedmeasures multivariate analysis of variance (MANOVA). Again, all post hoc t-tests involved Tukey corrections. There was a main effect contagion information on liking, F (1, 71) = 62.04, MSE = 1.24, p < .001, ηp2 = .47. In fact, a negative contagion effect was present in each of the target types separately, as well, ts(71) = 2.34, 4.10, and 15.77 for children, dogs, and objects, respectively, with p = .02 for children and p < .001 for dogs and objects. Also, as in Experiment 1, contagion was greater for objects than for either dogs or children, ts(71) = 11.57 and 12.90, respectively, ps < .001, and there was no difference between dogs and children (p = .05). There was also a main effect of contagion information on pity ratings, F(1, 71) = 161.59, MSE = 3.75, p < .001, ηp2 = .70, and this was true for each target type separately, as well, ts(71) = 10.69, 11.99, and 4.79, for children, dogs, and (even) objects, respectively, all ps < .001. The differences in pity ratings between the targets mirrored the differences in liking ratings, where negative information resulted in greater differences between objects and either dogs or children, ts(71) = 4.73 and 5.26, respectively, ps < .001, and again there was no difference between children and dogs (p > .10). Further, there was a main effect of association type on both liking ratings, F(1, 71) = 6.77, MSE = 1.01, p = .01, Ó 2018 Hogrefe Publishing

Table 2. Mean likeability and pity ratings (1–10; with standard deviations in parentheses) for children, dogs, and objects as a function of contagion type (parent/owner professions or self-information) and contagion information valence (negative or positive) in Experiment 2 Parent/owner profession

Self-information

Positive

Negative

Positive

Negative

Liking

7.78 (1.66)

7.15 (1.87)

7.71 (1.70)

7.88 (1.89)

Pity

2.62 (2.08)

3.05 (2.02)

1.81 (1.36)

5.76 (2.70)

Liking

8.57 (1.51)

7.90 (1.81)

8.43 (1.62)

8.02 (1.85)

Pity

2.75 (2.20)

3.29 (2.43)

2.17 (1.82)

5.54 (2.31)

Liking

6.57 (1.75)

5.28 (1.93)

6.51 (1.76)

5.77 (2.11)

Pity

1.76 (1.43)

2.12 (1.65)

1.50 (1.22)

2.90 (2.04)

Child

Dog

Object

ηp2 = .09, and pity ratings, F(1, 71) = 44.10, MSE = 2.30, p < .001, ηp2 = .38. Parent/owner profession contagion led to overall lower liking ratings as well as overall lower pity ratings compared to self-information contagion. However, with liking judgments, the effect was present only for children, t(71) = 2.48, p = .02. There were no differences for dogs or objects (ps > .10). With pity judgments, the effect was present for children and dogs, ts(71) = 4.87, 4.12 (ps < .001) but not for objects (p = .06). Finally, there was an interaction between association type and contagion information for both liking ratings, F(1, 71) = 12.07, MSE = 1.29, p = .001, ηp2 = .15, and pity ratings, F(1, 71) = 154.44, MSE = 2.12, p < .001, ηp2 = .69, suggesting that the main effect of association type was driven primarily by the negative information. For liking, the decreases resulting from negative information were greater in the parent/ profession-information than in the self-information condition. More importantly, this was true only for children, ts(71) = 3.98 p < .001, and not for dogs or objects (p > .10 for dogs, p = .07 for objects). Similarly, the increases in pity ratings with negative information were greater in the self-information condition than in the parent/owner profession condition. However, in this case, this was true for children and dogs, ts(71) = 4.87 and 4.12, p < .001, but not for objects (p = .06). Because the effects appeared to be modulated by target type, we conducted more detailed analyses looking at each target type separately. In terms of liking, ANOVAs showed an interaction between contagion information and association type only with child targets, F(1, 71) = 8.12, MSE = 2.01, p = .006, ηp2 = .21, and not for dog or object targets, Fs(1, 71) = 1.57 and 2.58, MSEs = 0.80 and 2.05, both ps > .10. Children appeared to be special: there was no contagion in the self-information condition, t(71) = 1.28, p > .10 [despite the replicated robust contagion in the parent information condition, t(71) = 4.43, p < .001]. In fact, this was true even Experimental Psychology (2018), 65(1), 49–60

though likeability for dogs was again higher than that for children, t(71) = 2.43, p = .018. We had predicted pity would be related to the observed mitigation of the contagion effect with animate targets. This prediction was not supported. Even though, as with liking decisions, pity decisions also showed a main effect of contagion information as well as association type and an interaction between the two variables, the two decisions were not correlated. Correlations were not expected in the noncontagion cases, and none were obtained (all ps > .10). However, the correlations in contagion cases were also not significant, not even in the self-information condition where pity ratings were much higher for child and dog targets. With proximity associations, the correlations between liking and pity were r = .04, .09, and .01, and with self-associations they were .21, .01, and .10, for children, dogs, and objects, respectively, all ps > .10. The lack of significant correlations was also not due to a truncated range, as both liking and pity ratings spanned the entire scale in all but a few cases. Interestingly, although it was not high and did not reach significance, the correlation between pity and liking in the case of children in the self-information contagion condition stood out among the others. This result may hint at some relationship that is specific to children, but it may also be due to a third variable that increased both liking and pity ratings. Finally, none of the other questions in the second half of the experiment correlated with liking decisions, either.

Discussion To date, research on contagion in magical thinking due to past associations with undesirable others/things/events had focused on contamination of inanimate targets, such as food and clothing (e.g., Rozin et al., 1986). We explored whether decisions about animate targets might be different, at least partly because animate targets with any negative association not in their control could also evoke pity which could counter the effects of contagion (cf. Batson & O’Quin, 1983; Carrera et al., 2012). Further, magical thinking studies had not needed to distinguish between different types of association that create contagion (association with others vs. with things/ events), since with inanimate targets both types were assumed to have the same underlying mechanism with respect to how they affected the target. We explored whether the type or source of negative association would make a difference with animate targets. We used children and dogs as our animate targets because vulnerable and nonthreatening targets as well as those not responsible for their negative situation tend to be more likely to be Experimental Psychology (2018), 65(1), 49–60

L. R. Bower et al., Contagion

labeled as objects of pity (Cikara & Fiske, 2011a, 2011b; Lishner, Batson, & Huss, 2011). In Experiment 1, following standard procedures, magical thinking was set up by introducing a contact with a negative entity. A robust contagion effect emerged for animate targets even when there was an option for “no choice,” or “feel neutral,” although the effect with liking ratings was somewhat reduced compared to that with inanimate targets. There were two caveats to Experiment 1, however. First, we did not ask for pity judgments. We simply assumed that pity should arise for animate targets with negative information that was not in their control and was irrelevant to their own merits. Second, we used only contagion via “connection to other people.” With this type of association we might have also tapped into the mereproximity effects with animate targets, at least with children, and thus the source of the observed contagion might have been different (stereotyping) from that observed with object targets (essence transfer). If this were the case, we could not conclude that magical thinking per se extended to animate targets. Indeed, contagion due to mere proximity would render the decisions to be somewhat less irrational for child targets than is typically assumed in magical thinking studies for inanimate targets, and perhaps also reduce the likelihood of pity in such situations. In Experiment 2, we replicated Experiment 1 results, and also teased out mere-proximity effects that were likely to have contaminated the results with children. When the negative information pertained to the items themselves and was unconfounded from proximity to or associations with others, there was indeed no negative contagion observed with children. The contagion effect was still strong with dogs and objects, however. Thus, magical thinking, outside of the effects of mere proximity, did not extend to all animate targets. Children appeared to be immune to contagion when proximity-association was replaced by selfassociation information. Dogs, on the other hand, despite being “liked” more than children and pitied as much, were still prone to the effects of magical thinking and thus treated more like the inanimate targets. These results suggest that the two types of contagion association tested in this experiment might not be dissociable for dogs and objects, perhaps not surprisingly since contagion due to mere proximity is likely not meaningful with nonhumans and is subsumed by the general negative associations tested in traditional magical thinking studies. However, with children, the results implicated a dissociation between the two sources of contagion and were consistent with the idea that the contagion resulting from negative information about parent profession observed in both experiments was at least to some degree due to mereproximity effects. When negative information was about Ó 2018 Hogrefe Publishing

L. R. Bower et al., Contagion

the self, contagion effects, although still seen when the targets were dogs and objects, disappeared when the targets were children. Also, although Experiment 1 suggested that contagion effects were reduced for animate targets (including dogs) compared to inanimate targets, and Experiment 2 showed this to be the case, our prediction about pity being the likely source for the reduction was not supported. There was no relationship between pity and liking, regardless of whether the negative information was supplied through proximity or self-association. Pity ratings were higher in the negative information conditions, especially in the self-association cases, thus supporting our rationale that the same information that would trigger contempt or distancing behaviors via magical thinking would also trigger pity. Nevertheless, it appeared that pity did not influence liking and that a different construct was likely responsible for the increases in both variables. A future direction might be to identify this different construct, which could overlap with pity in some respects or subsume it but could also be responsible for countering contagion in at least child targets without relying on the pity component. One possible candidate might be empathy, which can subsume pity but is not limited to it. Suffering and negative situations can trigger empathy and personal distress at the same time (Batson & O’Quin, 1983; Carrera et al., 2012), and helping behaviors, which could counter contagion effects, can be shaped by the relative strengths of these two feelings. When empathy is stronger than personal distress, behaviors tend toward altruism whereas the reverse balance is linked to more egoistic behaviors that focus on lessening the person’s own distress. If the latter dominates, individuals might be in a mindset that would make them more motivated to ignore the irrelevancy of the negative information and give it more weight than warranted, enabling the contagion effect to surface. If the former dominates, however, not only might feelings of pity increase, but so might, independently, the feelings of liking, which would in turn counter the effect of contagion. In fact, one prediction from this would be that for dog lovers, we might expect the same results as with children when the source of contagion is self-information and so magical thinking to disappear with these targets, as well. In sum, our findings underscored the importance of the type of contagion leading to irrational decisions, dissociating between contagion elicited by mere proximity and that elicited by other associations for children. A similar dissociation did not occur for either dogs or objects. We suggest that the mere-proximity effect and magical thinking are both at least partially rooted in automatic spreading activation at the emotional level which cannot be ignored through reason (cf. Berenbaum, Boden, & Baker, 2009). However, in the case of people, including children, there emerge Ó 2018 Hogrefe Publishing

different mechanisms by which the negative information is evaluated or weighted, which lead to the differences in the ability to counter the effects of contagion. In the case of non-people, including dogs, the mere-proximity effect is not applicable and thus the type of contagion association becomes irrelevant, and in such cases, all irrational decisions can be attributed to only magical thinking. Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/10.1027/ 1618-3169/a000389 ESM 1. Data (.xls) Raw data of Experiment 1. ESM 2. Data (.xls) Raw data of Experiment 2.

References Batson, C., & O’Quin, K. (1983). Influence of self-reported distress and empathy on egoistic versus altruistic motivation to help. Journal of Personality and Social Psychology, 45, 706–718. https://doi.org/10.1037/0022-3514.45.3.706 Berenbaum, H., Boden, M., & Baker, J. (2009). Emotional salience, emotional awareness, peculiar beliefs, and magical thinking. Emotion, 9, 197–205. https://doi.org/10.1037/a0015395 Carrera, P., Oceja, L., Caballero, A., Muñoz, D., López-Pérez, B., & Ambrona, T. (2012). I feel so sorry! Tapping the joint influence of empathy and personal distress on helping behavior. Motivation and Emotion, 37, 335–345. https://doi.org/10.1007/s11031012-9302-9 Cikara, M., & Fiske, S. T. (2011a). Bounded empathy: Neural responses to outgroup targets’ (mis)fortunes. Journal of Cognitive Neuroscience, 23, 3791–3803. https://doi.org/10.1162/ jocn_a_00069 Cikara, M., & Fiske, S. T. (2011b). Stereotypes and Schadenfreude: Affective and physiological markers of pleasure at outgroup misfortunes. Social Psychological and Personality Science, 3, 63–71. https://doi.org/10.1177/1948550611409245 Eisenberg, N., & Lennon, R. (1983). Sex differences in empathy and related capacities. Psychological Bulletin, 94, 100–131. https://doi.org/10.1037/0033-2909.94.1.100 Fiske, S. T., Cuddy, A. J. C., & Glick, P. (2007). Universal dimensions of social cognition: Warmth and competence. Trends in Cognitive Sciences, 11, 77–83. https://doi.org/10.1016/j. tics.2006.11.005 Garcia-Retamero, R., & Dhami, M. K. (2009). Take-the-best in expert-novice decision strategies for residential burglary. Psychonomic Bulletin & Review, 16, 163–169. https://doi.org/ 10.3758/PBR.16.1.163 Hebl, M. R., & Mannix, L. M. (2003). The weight of obesity in evaluating others: A mere proximity effect. Personality and Social Psychology Bulletin, 29, 28–38. https://doi.org/10.1177/ 0146167202238369 Hejmadi, A., Rozin, P., & Siegal, M. (2004). Once in contact, always in contact: Contagious essence and conceptions of purification in American and Hindu Indian children. Developmental Psychology, 40, 467–476. https://doi.org/10.1037/0012-1649. 40.4.467

Experimental Psychology (2018), 65(1), 49–60

Lindeman, M., & Svedholm, A. (2012). What’s in a term? Paranormal, superstitious, magical and supernatural beliefs by any other name would mean the same. Review of General Psychology, 16, 241–255. https://doi.org/10.1037/a0027158 Lishner, D. A., Batson, C. D., & Huss, E. (2011). Tenderness and sympathy: Distinct empathic emotions elicited by different forms of need. Personality and Social Psychological Bulletin, 37, 614–625. https://doi.org/10.1177/0146167211403157 Mehta, S. I., & Farina, A. (1988). Associative stigma: Perceptions of the difficulties of college-aged children of stigmatized fathers. Journal of Social & Clinical Psychology, 75, 70–81. https://doi. org/10.1521/jscp.1988.7.2-3.192 Neuberg, S. L., Smith, D. M., Hoffman, J. C., & Russell, F. J. (1994). When we observe stigmatized and “normal” individuals interacting: Stigma by association. Personality and Social Psychology Bulletin, 20, 196–209. https://doi.org/10.1177/ 0146167294202007 Preston, S. D. (2013). The origins of altruism in offspring care. Psychological Bulletin, 139, 1305–1341. https://doi.org/ 10.1037/a0031755 Rozin, P., Grant, H., Parker, S., & Weinberg, S. (2007). “Head versus heart”: Effect of monetary frames on expression of sympathetic magical concerns. Judgment and Decision Making, 2, 217–224. Rozin, P., Markwith, M., & Ross, B. (1990). The sympathetic magical law of similarity, nominal realism and neglect of negatives in response to negative labels. Psychological Science, 1, 383–384. https://doi.org/10.1111/j.1467-9280.1990.tb00246.x

Experimental Psychology (2018), 65(1), 49–60

L. R. Bower et al., Contagion

Rozin, P., Millman, L., & Nemeroff, C. (1986). Operation of the laws of sympathetic magic in disgust and other domains. Journal of Personality and Social Psychology, 50, 703–712. https://doi. org/10.1037/0022-3514.50.4.703 Rozin, P., & Nemeroff, C. (2002). Sympathetic magical thinking: The contagion and similarity “heuristics”. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and Biases (pp. 201– 216). New York, NY: Cambridge University Press. Rozin, P., & Royzman, E. (2001). Negativity bias, negativity dominance, and contagion. Personality and Social Psychology Review, 5, 296–320. https://doi.org/10.1207/S15327957PSPR0504_2 Schieman, S., & Gundy, K. V. (2000). The personal and social links between age and self-reported empathy. Social Psychology Quarterly, 63, 152–174. https://doi.org/10.2307/269588 Received August 10, 2015 Revision received September 6, 2017 Accepted September 20, 2017 Published online February 8, 2018 lu Zehra F. Peynirciog Department of Psychology American University Washington, DC, 20016 USA peynir@american.edu

Ó 2018 Hogrefe Publishing

L. R. Bower et al., Contagion

Appendix A Example Materials From Experiment 1 They illustrate the differences in target type (child, dog, or object) and contagion information (positive or negative). Counterbalancing measures are presented in Appendix C.

Ă&#x201C; 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 49â&#x20AC;&#x201C;60

L. R. Bower et al., Contagion

Example Materials From Experiment 2 They illustrate the differences in object type (child, dog, or object), contagion type (parent/owner profession or self-information), and contagion information (positive or negative).

Appendix B Professions Used in Experiment 1 Negative: disbarred lawyer, drug dealer, peddler, politician, sleazy mechanic, stockbroker, telemarketer, TV evangelist, usedcar salesman. Positive: chef, decorated military officer, doctor, engineer, firefighter, hot-line volunteer, paramedic, police officer, teacher.

Professions Used in Experiment 2 Negative: politician, sleazy mechanic, stockbroker, telemarketer, TV evangelist, used-car salesman. Positive: chef, decorated military officer, firefighter, hot-line volunteer, paramedic, teacher.

Experimental Psychology (2018), 65(1), 49â&#x20AC;&#x201C;60

Ă&#x201C; 2018 Hogrefe Publishing

L. R. Bower et al., Contagion

Self-Information Used in Experiment 2 The pieces of information are either the same or analogous for each target type, and the positive and negative pieces are conceptually similar. Any given participant saw each such information only once. For instance, a participant who saw “puppy parvo” did not see “infant leukemia” or “damaged in the factory”; neither did he/she see “healthy metabolism” or “manufactured using the best materials.”

Negative Children: On the way to school passes by a cosmetics research facility targeted by the press. Survived infant leukemia but completely cured now. Found as a baby on the front steps of a hospital and adopted. Fell in the open sewer last week while playing and had to get a tetanus shot. Went on a mandatory field trip to a slaughterhouse last year. Lives in a house that was initially built to house orphan children. Dogs: While being walked, passes by a cosmetics research facility targeted by the press. Survived parvo as a puppy. Rescued from a shelter the day before euthanasia. Fell in the open sewer last week while playing and had to get a tetanus shot. Went to a flea-infested construction site last year and had many flea-baths before returning to normal. Lives in a house that was initially built to house orphan children. Objects: Belongs to someone who passes by a cosmetics research facility targeted by the press. Damaged in the factory but was fixed completely before being sold. Accidentally thrown in the garbage but rescued the next day and cleaned thoroughly. Dropped in an open sewer last week and had to be professionally sterilized. Kept in the dank attic until last year but was not damaged. Is in a house that was initially built to house orphan children.

Positive Children: On the way to school passes by a bookstore that also sells ice cream. Born with an incredibly healthy metabolism. Born in a hospital overlooking the beach. Played on the carpet last week with toys. Went to a botanical garden on a field trip. Lives in a house that was initially built as a cheerful combination toy/bookstore. Dogs: While being walked, passes by a bookstore that also sells ice cream. Born with an incredibly healthy metabolism. Picked out from a litter of 12 cute puppies. Played on the carpet last week with toys. Went to a huge dog-park with tropical trees. Lives in a house that was initially built as a cheerful combination toy/bookstore. Object: Belongs to someone who passes by a bookstore that also sells ice cream. Manufactured using the best materials available. Bought from among three others that looked exactly the same. Was played with on the carpet last week. Was taken to summer camp last year. Is in a house that was initially built as a cheerful combination toy/bookstore.

Other Pieces of Information in Both Experiments These pieces of information were specific to the picture and include Age: (Children) first grader, 3 years old, 4 years old; (Dogs) 3 years old, 2 years old, a year old, year and a half; (Objects) lightly used, new, 6 months old. Coloring: (Children) brown hair and brown eyes, red hair and blue eyes, black hair and dark eyes, fair skinned with blue eyes, dark haired, blond haired; (Dogs) brown with a black muzzle, white with brown patches, white with brown markings, mostly white with a brown face, tan and brown, brown with hints of white; (Objects) black with red wheels, green, brown with a white face, blue with white stripes, brown wood grain finish, brown. Hobby/Use: (Children) enjoys jump rope and hopscotch, playing with dinosaurs, plays outside, play in sandbox, plays with stuffed animals, enjoys bike riding, love to play outside, likes to collect flowers; (Dogs) enjoys playing with toys, enjoys working for food, enjoys chasing balls and sticks, enjoys being petted, is energetic, loves his training routine, loves to explore, loves to go for walks; (Objects) perfect for tricks, perfect for hugging, gets used regularly, has a full QWERTY keyboard, perfect for racing games, well-tuned guitar. Size: (Children) slight, lean build, tall, sturdy, average sized, small for her age; (Dogs) medium sized, small, large breed, small sized; (Objects) small, medium sized, standard sized.

Ó 2018 Hogrefe Publishing

Experimental Psychology (2018), 65(1), 49–60

L. R. Bower et al., Contagion

Appendix C The Counterbalancing Measures in Experiment 1 Two Groups: Exchange the negative and positive critical information between pairs so that each member of a given pair appears in the contagion and noncontagion information conditions equally often. For instance, for one group, Target A’s parent/owner was a police officer and Target B’s parent/owner was a used-car salesman, for the other group, Target A’s parent/owner was a used-car salesman and Target B’s parent/owner was a police officer. Two Subgroups: Exchange the position of each member of the pair so that each member of a given pair also appears on the right or the left side of the page equally often. For instance, for one subgroup, Target A appeared on the right and Target B appeared on the left, and for another subgroup, Target B appeared on the right and Target A appeared on the left. Three Further Subgroups: Rotate the critical information across target types such that each target is presented with each parent/owner information equally often. For instance, for one subgroup, the profession police officer would be associated with a child, for another subgroup, it would be associated with a dog, and for the third subgroup, it would be associated with an object. Two Final Subgroups: Exchange the pictures in the critical and noncritical-information text conditions such that each item (picture) serves in the target and red-herring conditions equally often. Half of the items did not have any profession information associated with them in order to draw attention away from the contagion manipulation. For instance, for one subgroup, Item A would contain parent/owner profession information and Item B would not, and for the other subgroup, Item B would contain parent/owner profession information and Item A would not. Hence there were 24 counterbalancing groups, with 5 people in each.

Experimental Psychology (2018), 65(1), 49–60

Ó 2018 Hogrefe Publishing

Instructions to Authors Experimental Psychology publishes innovative, original, high-quality experimental research. The scope of the journal is defined by experimental methodology and thus papers based on experiments from all areas of psychology are welcome. To name just a few fields and domains of research, Experimental Psychology considers manuscripts reporting experimental work on human learning, memory, perception, action, language, thinking, problem-solving, judgment and decision making, social cognition, and neuropsychological aspects of these topics. Apart from the use of experimental methodology, a primary criterion for publication is that research papers make a substantial contribution to theoretical research questions. For experimental papers that have a mainly applied focus, Experimental Psychology is not the appropriate outlet. A major goal of Experimental Psychology is to provide a particularly fast outlet for such research. Authors usually receive an editorial decision within 6 weeks of manuscript submission. Experimental Psychology publishes the following types of article: Research Articles, Short Research Articles, Theoretical Articles, and Registered Reports. Replication studies should be submitted as a Registered Report. Manuscript Submission. All manuscripts should in the first instance be submitted electronically at http://www.editorial manager.com/exppsy. Detailed instructions to authors are provided at http://www.hogrefe.com/j/exppsy Copyright Agreement. By submitting an article, the author confirms and guarantees on behalf of him-/herself and any coauthors that he or she holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches, tables, raw data, and other electronic supplementary material, and that the article and its contents does not infringe in any way on the rights of third parties. ESM and raw data files 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 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, and other form, in the journal to which it has been submitted and in other independent publications, with no limits on the number of copies or on the form or the extent of the distribution. These rights are transferred for the duration of copyright as defined by international law.

Ó 2018 Hogrefe Publishing

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 analogue, 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 Hogrefe will send the corresponding author of each accepted paper free of charge an e-offprint (PDF) of the published version of the paper when it is first released online. This e-offprint is provided exclusively for the author’s personal use, including for sharing with coauthors. Other uses of the e-offprint/ published version of record, including but not limited to the following, are not permitted except with the express written permission of the publisher: posting the e-offprint/published version of record to a personal or institutional website or to an institutional or disciplinary repository; changing or modifying the digital file; reproducing, distributing, or licensing the article in whole or in part for commercial use. If you wish to post the article to your personal or institutional website or to archive it in an institutional or disciplinary repository, please use either a pre-print or a post-print of your manuscript in accordance with the publication release for your article and the document ‘‘Guidelines on sharing and use of articles in Hogrefe journals’’ on the journal’s web page at www.hogrefe. com/j/exppsy. December 2016

Experimental Psychology (2018), 65(1)

How to assess the social atmosphere in forensic hospitals and identify ways of improving it “All clinicians and researchers who want to help make forensic treatment environments safe and effective should buy this book.” Mary McMurran, PhD, Professor of Personality Disorder Research, Institute of Mental Health, University of Nottingham, UK

Norbert Schalast / Matthew Tonkin (Editors)

The Essen Climate Evaluation Schema – EssenCES A Manual and More

2016, x + 108 pp. US $49.00 / € 34.95 ISBN 978-0-88937-481-2 Also available as eBook The Essen Climate Evaluation Schema (EssenCES) described here is a short, well-validated questionnaire that measures three essential facets of an institution’s social atmosphere. An overview of the EssenCES is followed by detailed advice on how to administer

www.hogrefe.com

and score it and how to interpret findings, as well as reference norms from various countries and types of institutions. The EssenCES “manual and more” is thus a highly useful tool for researchers, clinicians, and service managers working in forensic settings.

Test development and construction: Current practices and advances “This book is indispensable for all who want an up-to-date resource about constructing valid tests.” Prof. Dr. Johnny R. J. Fontaine, President of the European Association of Psychological Assessment, Faculty of Psychology and Educational Sciences, Ghent University, Belgium

Karl Schweizer / Christine DiStefano (Editors)

Principles and Methods of Test Construction Standards and Recent Advances

(Series: Psychological Assessment – Science and Practice – Vol. 3) 2016, vi + 336 pp. US $69.00 / € 49.95 ISBN 978-0-88937-449-2 Also available as eBook This latest volume in the series Psychological Assessment – Science and Practice describes the current state-of-the-art in test development and construction. The past 10–20 years have seen substantial advances in the methods used to develop and administer tests. In this volume many of the world’s leading authorities collate these advances and provide information about current practices, thus equipping researchers and students to successfully construct new tests using the best modern standards and

www.hogrefe.com

techniques. The first section explains the benefits of considering the underlying theory when designing tests, such as factor analysis and item response theory. The second section looks at item format and test presentation. The third discusses model testing and selection, while the fourth goes into statistical methods that can find group-specific bias. The final section discusses topics of special relevance, such as multitraitmultimethod analyses and development of screening instruments.

A valuable tool for assessing and documenting psychopathology in Now sh! Engli

Matthew R. Broome / Ronald Bottlender / Michael Rösler / Rolf-Dieter Stieglitz (Editors)

The AMDP System

Manual for Assessment and Documentation of Psychopathology in Psychiatry 9th edition 2018, xii + 198 pp. US $34.80 / € 27.95 ISBN 978-0-88937-542-0 Also available as eBook Now in its 9th edition, the AMDP System is a widely used tool for documenting psychiatric symptoms in clinical and research projects. It is not only an essential part of many internal clinical documentation systems, but it is also a valuable instrument for training the identification of psychopathological symptoms. This new edition, now in English, involved the complete revision of the psychopathological and somatic symp-toms, with particular attention to the elimination of ambiguities and inconsistencies, to the precision of definitions, to the exact differentiation of the self and other ratings, as well as to the user friendliness of the system.

www.hogrefe.com

Taking account of developments in recent years in the description and rating of psychopathological symptoms, this latest edition of the AMDP System has been extended by an additional eleven psychopathological and three somatic symptoms and a new section on syndrome formation in the appendix. This clearly structured manual enables the standardized application of the system, making it an invaluable tool in the training of medicine and psychology students and an essential reference volume in the psychiatric field. The AMDP System documentation forms are also available separately.