European Journal of Health Psychology by Hogrefe

Volume 27 / Number 2 / 2020

European Journal of

Health Psychology

European Journal of Health Psychology

Editor-in-Chief Heike Spaderna Associate Editors Heike Eschenbeck Verena Klusmann Christel Salewski Silke Schmidt Andreas Schwerdtfeger Anja Tausch Claus Vögele

A concise guide to the assessment and treatment of insomnia for busy professionals “This book provides an excellent concise introduction to insomnia and its treatment with cognitive behavioral therapy. A great addition to any therapist’s bookshelf!” Philip Gehrman, PhD, CBSM, Associate Professor, Department of Psychiatry, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA

William K. Wohlgemuth / Ana Imia Fins

Insomnia (Series: Advances in Psychotherapy – Evidence-Based Practice – Volume 42) 2019, viii + 94 pp. US $29.80 / € 24.95 ISBN 978-0-88937-415-7 Also available as eBook About 40% of the population experiences difficulty falling or staying asleep at some time in a given year, while 10% of people suffer chronic insomnia. This concise reference written by leading experts for busy clinicians provides practical and upto-date advice on current approaches to assessment, diagnosis, and treatment of insomnia. Professionals and students learn to correctly identify and diagnose insomnia and gain hands-on information on how to

www.hogrefe.com

carry out treatment with the best evidence base: cognitive behavioral therapy for insomnia (CBT-I). The American Academy of Sleep Medicine (AASM) and the American College of Physicians (ACP) both recognize CBT-I as the first-line treatment approach to insomnia. Appendices include useful resources for the assessment and treatment of insomnia, which readers can copy and use in their clinical practice.

European Journal of

Health Psychology Volume 27 / Number 2 / 2020

Editor-in-Chief Editorial Ofﬁce

Heike Spaderna, PhD, Health Psychology, Department of Nursing Science, Trier University, Am Wissenschaftspark 25-27, 54296 Trier, Germany, E-mail spaderna@uni-trier.de Lisa-Marie Maukel, Health Psychology, Department of Nursing Science, Trier University, Am Wissenschaftspark 25-27, 54296 Trier, Germany, E-mail ejhp@uni-trier.de

Associate Editors

Heike Eschenbeck, Schwäbisch-Gmünd, Germany Verena Klusmann, Hamburg, Germany Christel Salewski, Hagen, Gemany Silke Schmidt, Greifswald, Germany

Andreas Schwerdtfeger, Graz, Austria Anja Tausch, Riedlingen, Germany Claus Vögele, Luxembourg, Luxembourg

Editorial Board

Urs Baumann, Salzburg, Austria Elmar Brähler, Leipzig, Germany Birte Dohnke, Schwäbisch Gmünd, Germany Michael Eid, Berlin, Germany Toni Faltermaier, Flensburg, Germany Dieter Frey, München, Germany Edgar Geissner, Prien am Chiemsee, Germany Nina Knoll, Berlin, Germany Carl-Walter Kohlmann, Schwäbisch Gmünd, Germany Thomas Kubiak, Mainz, Germany

Arnold Lohaus, Bielefeld, Germany Friedrich Lösel, Cambridge, UK Mike Martin, Zürich, Switzerland Britta Renner, Konstanz, Germany Wolfgang Schlicht, Stuttgart, Germany Urte Scholz, Zürich, Switzerland Ralf Schwarzer, Berlin, Germany Monika Sieverding, Heidelberg, Germany Wolfgang Stroebe, Utrecht, The Netherlands Petra Warschburger, Potsdam, Germany Jürgen Wegge, Dresden, Germany

Publisher

Hogrefe Publishing, Merkelstraße 3, 37085 Göttingen, Germany, Tel. 0551 99950-0, Fax 0551 99950-111, E-mail publishing@hogrefe.com North America: Hogrefe Publishing, 361 Newbury Street, 5th Floor, Boston, MA 02115, USA, Tel. (857) 880-2002, E-mail publishing@hogrefe.com Regina Pinks-Freybott, Hogrefe Publishing, Merkelstraße 3, 37085 Göttingen, Germany, Tel. +49551 99950-0, Fax 0551 99950-111, E-mail production@hogrefe.com Hogrefe Publishing, Herbert-Quandt-Straße 4, D-37081 Göttingen, Germany, Tel. +49551 99950-900, Fax 0551 99950-998, E-mail zeitschriftenvertrieb@hogrefe.com Melanie Beck, Hogrefe Publishing, Merkelstraße 3, 37085 Göttingen, Germany, Tel. +49551 99950-423, Fax 0551 99950-111, E-mail marketing@hogrefe.com ISSN-L 2512-8442, ISSN-Print 2512-8442, ISSN-Online 2512-8450

Production Subscriptions Advertising/Inserts ISSN Copyright Information

Publication

Subscription Prices

Payment

Electronic Full Text Abstracting Services

Ó 2020 Hogrefe Publishing. This journal as well as the individual contributions and illustrations contained within it are protected under international copyright law. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microﬁlming, recording or otherwise, without prior written permission from the publisher. All rights, including translation rights, reserved. Published in 4 issues per annual volume. European Journal of Health Psychology is the continuation of Zeitschrift für Gesundheitspsychologie (ISSN 0943-8149), the last annual volume of which (Volume 24) was published in 2016. Calendar year subscriptions only. Rates for 2020: Institutions – from US $326.00/€249.00 (print only; pricing for online access can be found in the journals catalog at hgf.io/journalscatalog); Individuals – US $128.00/€98.00 (print & online). Postage and handling – US $16.00/€12.00. Single copies – US $86.00/€67.00 + postage and handling. Payment may be made by check, international money order, or credit card, to Hogrefe Publishing, Merkelstr. 3, D-37085 Göttingen, Germany. US and Canadian subscriptions can also be ordered from Hogrefe Publishing, 361 Newbury Street, 5th Floor, Boston, MA, 02115, USA. The full text of European Journal of Health Psychology is available online at http://econtent.hogrefe.com and in PsycARTICLES. Abstracted/Indexed in Social Sciences Citation Index (SSCI), Social Scisearch, Journal Citation Report/Social Sciences Edition, PsycINFO, PsycARTICLES, PSYNDEX, Scopus, IBZ, and IBR. 2019 Impact Factor 0.692, Journal Citation Reports (Clarivate Analytics, 2020)

European Journal of Health Psychology (2020), 27(2)

Ó 2020 Hogrefe Publishing

Contents Original Articles

Does Perceived Stress Affect the Relationship Between Personality and Sports Enjoyment? Freya Dunker, Philipp A. Freund, and Eliane S. Engels

Short Form of the State-Trait Anger Expression Inventory-2

Ana N. Tibubos, Karin Schermelleh-Engel, and Sonja Rohrmann

News and Announcements

Ó 2020 Hogrefe Publishing

Affect Improvements and Measurement Concordance Between a Subjective and an Accelerometric Estimate of Physical Activity Björn Pannicke, Julia Reichenberger, Dana Schultchen, Olga Pollatos, and Jens Blechert

Meeting Calendar

European Journal of Health Psychology (2020), 27(2)

Original Article

Does Perceived Stress Affect the Relationship Between Personality and Sports Enjoyment? Freya Dunker , Philipp A. Freund, and Eliane S. Engels Institute of Psychology, Leuphana University of Lüneburg, Germany

Abstract: Background: Sports enjoyment is a prerequisite for continued engagement in sports (e.g., Mullen et al., 2011; Woods, Tannehill, & Walsh, 2012). To date, it is understudied whether perceived stress and personality relate to the experience of sports enjoyment. Aim: The aim of the present study was to investigate the relationships between perceived stress, personality (the Big Five), and sports enjoyment. Method: Data were collected online from a group of N = 195 adults. There were two points of measurements with an interval of 4 weeks in between. Personality was assessed at t1, while perceived stress and sports enjoyment were measured at t2. Results: The results indicate that perceived stress is significantly related to sports enjoyment: A high amount of perceived stress was associated with less enjoyment of sports. Neuroticism, extraversion, conscientiousness, and agreeableness were related to sports enjoyment, but openness was not. Models testing the mediating effect of perceived stress for personality on sports enjoyment showed significant direct effects for extraversion and conscientiousness and significant indirect effects for neuroticism, extraversion, and conscientiousness. Limitations: Limitations concerning the sample characteristics and some poor fit statistics for the models including openness and extraversion are discussed. Conclusion: Overall, our findings suggest that perceived stress influences the affective experience in physical exercise. Keywords: sports enjoyment, perceived stress, personality, Big Five, health

Regular physical activity is crucial for an individual’s physical and psychological health (e.g., Janssen, 2007; Physical Activity Guidelines Advisory Committee, 2008). Still, about 40% of adults in Europe do not reach the WHO recommendation of at least 150 min of moderate physical activity weekly (Marques, Sarmento, Martins, & Nunes, 2015). Why do some people not engage in sports regularly despite being aware of its beneficial effects? In the literature, sports enjoyment1 has been found to be an important predictor of long-term engagement in sports (e.g., Mullen et al., 2011; Woods, Tannehill, & Walsh, 2012). Therefore, people who engage in physical activity regularly differ from less physically active people in the perception of sports enjoyment. It is important to understand influencing factors of sports enjoyment in order to promote a long-term oriented physically active lifestyle in individuals (Engels & Freund, 2019). Previous research has consistently shown the benefits of regular physical activity during stressful times (e.g., Rimmele et al., 2009; Schmid, Schröder, Eschenbeck, & Kohlmann, 2016) and at the same time provided evidence that perceived stress is negatively related to sports frequency (Schmid et al., 2016; Steptoe, Wardle, Pollard, Canaan, & Davies, 1996). However, no study has yet 1

systematically looked at the relationship between perceived everyday stress and sports enjoyment despite consensus on the importance of sports enjoyment for continued engagement in sports (e.g., Mullen et al., 2011; Woods et al., 2012). The Big Five factors neuroticism, extraversion, openness for experience, and conscientiousness have been reported as predictors of stress perception (e.g., Britton, Kavanagh, & Polman, 2017; Kaiseler, Polman, & Nicholls, 2012), yet research on the relationship between these factors and sports enjoyment is also scarce. Courneya and Hellsten (1998) provided first evidence of the influence of neuroticism, extraversion, and openness for experience on the subjective value of the enjoyment of sports, while further studies imply the importance of conscientiousness in the initiation and maintenance of sports related behavior (e.g., Rhodes & Smith, 2006; Wilson & Dishman, 2015). Further systematical research on the relationship between personality (i.e., the Big Five) and sports enjoyment is missing to this point. The aim of the present study was to investigate the relationships of perceived stress and the Big Five with sports enjoyment. An empirical approach to investigating the effect of perceived stress and personality on the enjoyment

Sports enjoyment will be used synonymously here with the term enjoyment of sports.

Ó 2020 Hogrefe Publishing

European Journal of Health Psychology (2020), 27(2), 45–54 https://doi.org/10.1027/2512-8442/a000048

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

of sports appears useful to help understanding basic mechanisms of why some people do not like to engage in sports as well as possibly preventing physically inactive lifestyles, while also fostering long-term oriented engagement in sports – especially during stressful times.

may be due to physical improvement (Stein, Fisher, Berkey, & Colditz, 2007), mastery of training goals (Papaioannou, Bebetsos, Theodorakis, Christodoulidis, & Kouli, 2006), or social factors, such as social relatedness (Engels & Freund, 2018). Therefore, someone who participates in sports only irregularly is likely to perceive less sports enjoyment and, hence, is likely to engage in physical activity less often. Therefore, sports enjoyment is also both an antecedent and a consequence of physical activity (e.g., Mullen et al., 2011; Scanlan & Simons, 1992; Williams et al., 2006). Physical activity, in turn, modulates an individuals’ response to stress (Rimmele et al., 2009), meaning that regular physical activity shows protective effects to stress, such as a higher self-esteem, social integration, or a decrease in stress hormones (Rimmele et al., 2009; Sipos, Jeges, & Tóth, 2015). Sports enjoyment is thus beneficial to the ability to cope well with stress by positively affecting physical activity. However, very little is currently known about predictors of sports enjoyment in adults and the effect of perceived stress on the experience of sports enjoyment despite the demonstrated importance of sports enjoyment on our health during stressful times.

Sports Enjoyment Sports enjoyment refers to the relatively stable yet malleable affective component of the experience in sports (Engels & Freund, 2019). Following Engels and Freund (2019), the construct consists of the three facets pleasure (positive affective response to performing a sport), flow (perceived optimal physical demand and ideal process during the sport), and recovery (feeling balanced and relaxed after sport). The extant work of Engels and Freund (2018, 2019) has focused on the measurement and formation of sports enjoyment in an educational context (i.e., specifics of physical education leading to the enjoyment of physical activity). It has been shown that sports enjoyment consists of rather stable (pleasure) but also influenceable components (flow and recovery).2 This implies that sports enjoyment may vary depending on the situation and is therefore subject to change: For instance, the type of sport – and how it is taught/practiced in physical education class – can impact the enjoyment someone gets out of it (e.g., Ball, Bice, & Parry, 2014; Engels & Freund, 2018, 2019; Frederick & Ryan, 1993; McCarthy, Jones, & ClarkCarter, 2008). Also, a sport context that fulfills the basic needs of Self-Determination-Theory (need for autonomy, need for competence, need for social relatedness; Deci & Ryan, 1985) is associated with an increase in sports enjoyment in physical education (Mouratidis, Vansteenkiste, Sideridis, & Lens, 2011; Sanchez-Oliva, Sanchez-Miguel, Leo, Kinnafick, & García-Calvo, 2014). Deci and Ryan (1985) argue that both intrinsic and extrinsic factors are relevant for enjoying an activity, hence it is reasonable to assume that sports enjoyment is both a motivation to participate and a varying state (or outcome) of sport participation (Ashford, Biddle, & Goudas, 1993). Sports enjoyment is indispensable for maintaining a physically active lifestyle because it is a vital predictor of regular physical activity (Mullen et al., 2011; Woods et al., 2012). Sports enjoyment can help predicting whether someone exercises regularly, as the likelihood of exercising increases in relation to the anticipation of positive emotions. According to Rovniak, Anderson, Winett, and Stephens (2002), sports enjoyment can also be seen as a consequence of regular physical activity. The more often someone exercises, the more she or he will enjoy doing sports. This effect

Perceived Stress While a variety of understandings of the term stress have been suggested, this paper will refer to the definition of perceived stress by Lazarus (2006), who describes stress as a consequence of a subjective cognitive appraisal of a possibly stressful situation or stimulus. Accordingly, stress is not necessarily the outcome of an event but rather an individuals’ interpretation of a stressor and an evaluation of existing coping resources (Lazarus, 2006). Therefore, it differs between individuals (Cohen, Kamarck, & Mermelstein, 1983). Personality, for instance, has been shown to be related to stress reactivity and perceived stress in a way that individuals high on neuroticism and low on extraversion, openness, and conscientiousness perceive more stress (Britton et al., 2017; Roohafza et al., 2016). Physical activity plays an important role in regulating reactions to stress by reducing the number of stress hormones which could lead to high blood pressure if levels of the stress hormone cortisol remain high (Rimmele et al., 2009). Sports reduces stress reactivity, the “tendency to respond to a stressor” (Schlotz, 2013, para. 4), when it takes place after the stressful event. Additionally, physical activity helps to build stress-protective resources, such as selfesteem (Sipos et al., 2015) or social integration (Wankel & Berger, 1990), which prevent individuals from forming a high stress reactivity. Accordingly, the positive effects of

These findings are in part supported by the work of Csikszentmihalyi (2013) who defines flow-experience as a state-component.

European Journal of Health Psychology (2020), 27(2), 45–54

Ó 2020 Hogrefe Publishing

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

sports – preventing and buffering – on the experience of stress are well established while less is known about the effects of personal stress on physical activity itself (Reiner, Niermann, Krapf, & Woll, 2013). Research shows that high levels of perceived everyday stress relate to a decrease in physical activity (e.g., Burg et al., 2017; Schmid et al., 2016). Accordingly, the more stress someone perceives, the less often this person will perform a sport. This may be explained with (perceived) lack of time and/or exhaustion (Schmid et al., 2016). In contrast, Gerber (2008) points out that adults might perform a sport more frequently if they experience more stress because it can be used as a way of coping and reducing stress. Additionally, Ullrich-French and Smith (2009) provided evidence that perceived stress during a sport correlates negatively with perceived sports competence, motivation, and sports enjoyment in adolescents. Additionally, daily stress is associated with more negative affect (Lippold, Davis, McHale, Buxton, & Almeida, 2016), increased fatigue (Zohar, 1999), lack of energy (Schmid et al., 2016), anxiety, and depression (Charles, Piazza, Mogle, Sliwinski, & Almeida, 2013; Spada, Nikčević, Moneta, & Wells, 2008). With regard to the negative effect of perceived daily stress on physical activity and affective experiences in general, a negative relation between perceived everyday stress and sports enjoyment is expected. However, and to the best of our knowledge, no previous study has specifically examined this relationship.

and some research (Sutin et al., 2016) indicates a positive relation between agreeableness and physical activity. These findings indicate the relevance of the Big Five factors for the affective experience in sports. While sports enjoyment has been found to be highly relevant for a long-term oriented sports engagement (e.g., Mullen et al., 2011), research on the relationship between personality and sports enjoyment is scarce to this point. Courneya and Hellsten (1998) provided evidence that neuroticism correlated negatively while extraversion and openness correlated positively with sports enjoyment measured as a motive for regular exercise. Participants rated the importance of enjoyment for exercise on a scale ranging from 1 (= minor importance) through 3 (= moderate importance) to 5 (= major importance). Their findings indicate that sports enjoyment is less important for individuals with high values in neuroticism and/or low values in extraversion and openness. Although these findings indicate the relevance of neuroticism, extraversion, and openness for sports enjoyment, it needs to be pointed out that Courneya and Hellsten (1998) merely asked participants how important they rated enjoyment in sports. Thus, there is uncertainty whether these findings extend to the affective experience of sports enjoyment in adults. Hence, previous research (Courneya & Hellsten, 1998) has only measured the perceived value of sports enjoyment as a means to facilitate sport engagement, but the actual experience of sports enjoyment has not yet been studied. Research on the relations between the Big Five and stress is well established. For individuals scoring high on neuroticism, there are consistent findings reporting higher stress reactivity (e.g., Britton et al., 2017), higher stress intensity (e.g., Kaiseler et al., 2012), and lower perceived control over stressors (e.g., Kaiseler et al., 2012). In general, neuroticism is associated with anxiety, worrying, and a general negative affect (Robinson, Ode, Moeller, & Goetz, 2007). Low stress reactivity is related to extraversion and openness (Britton et al., 2017). Conscientiousness is associated with low exposure to stress and threat appraisals (Carver & ConnorSmith, 2010), while agreeableness is connected to low social conflicts (Carver & Connor-Smith, 2010). In general, extraversion, openness, conscientiousness, and agreeableness are related to more adaptive and effective coping strategies and thus lower stress levels (Kaiseler et al., 2012).

Personality Up to now, only little attention has been paid to the impact of personality on the experience of sports enjoyment in adults. For sports in general, neuroticism has been found to be negatively related to motivation, energy, and longterm engagement in sports (Courneya & Hellsten, 1998), as well as negatively related to physical activity in general (Rhodes & Smith, 2006; Sutin et al., 2016; Wilson & Dishman, 2015). Furthermore, athletes with high values in neuroticism report being frightened of humiliation during physical activity (Courneya & Hellsten, 1998). In contrast, extraversion has been found to predict long-term oriented sport activity (Rhodes & Smith, 2006; Sutin et al., 2016; Wilson & Dishman, 2015), and people with high values in extraversion are less likely to report lacking energy when performing a sport (Courneya & Hellsten, 1998). Openness has been found to predict an overall physically active lifestyle (Rhodes & Smith, 2006; Sutin et al., 2016; Wilson & Dishman, 2015) and athletes score higher on openness compared to individuals less likely to engage in a sport (Allen, Greenlees, & Jones, 2013). Finally, conscientiousness is positively related to physical activity (Rhodes & Smith, 2006; Sutin et al., 2016; Wilson & Dishman, 2015) Ó 2020 Hogrefe Publishing

The Present Study This study aimed to find out if perceived stress and personality affect the enjoyment of sports in adults. Specifically, the aim was to shed light on the effect of perceived everyday stress and the Big Five factors on sports enjoyment. Sports enjoyment is a principal predictor of regular physical activity (e.g., Mullen et al., 2011; Scanlan & Simons, European Journal of Health Psychology (2020), 27(2), 45–54

1992), which in turn is essential to cope well with stress (Rimmele et al., 2009; Sipos et al., 2015). At the same time, perceived everyday stress appears to be negatively related to sport activity (e.g., Burg et al., 2017; Schmid et al., 2016) and overall associated with negative affect (Lippold et al., 2016), increased fatigue (Zohar, 1999), and lack of energy (Schmid et al., 2016). This indicates a need to investigate the relationship between the perception of stress and sports enjoyment. Moreover, we tested if perceived stress mediates the relationship between personality and sports enjoyment. On the grounds of the theoretical and empirical review above, the following hypotheses are being proposed: Hypothesis 1 (H1): Perceived stress is negatively related to sports enjoyment. Hypothesis 2a (H2a): Neuroticism is positively related to perceived stress. Hypothesis 2b (H2b): Extraversion is negatively related to perceived stress. Hypothesis 2c (H2c): Openness is negatively related to perceived stress. Hypothesis 2d (H2d): Conscientiousness is negatively related to perceived stress. Hypothesis 2e (H2e): Agreeableness is negatively related to perceived stress. Hypothesis 3a (H3a): Neuroticism is negatively related to sports enjoyment. Hypothesis 3b (H3b): Extraversion is positively related to sports enjoyment. Hypothesis 3c (H3c): Openness is positively related to sports enjoyment. Hypothesis 3d (H3d): Conscientiousness is positively related to sports enjoyment. Hypothesis 3e (H3e): Agreeableness is positively related to sports enjoyment. Hypothesis 4a (H4a): Perceived stress mediates the effect of neuroticism on sports enjoyment.

3 4

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

Hypothesis 4b (H4b): Perceived stress mediates the effect of extraversion on sports enjoyment. Hypothesis 4c (H4c): Perceived stress mediates the effect of openness on sports enjoyment. Hypothesis 4d (H4d): Perceived stress mediates the effect of conscientiousness on sports enjoyment. Hypothesis 4e (H4e): Perceived stress mediates the effect of agreeableness on sports enjoyment.

Material and Methods Procedure/Study Design The present study3 was conducted as an online survey. Participants were recruited from a German university with the incentive of course credit. Further participants were contacted via social media and e-mail lists of sports clubs with the incentive of winning one out of ten 10€ gift vouchers. There were two points of measurements with an interval of 4 weeks in between. Personality was assessed at t1 and general information on exercise behavior and demographic information of participants were also collected at this point of measurement. Stress and sports enjoyment were assessed at t1 and t2, but for the present study, only the t2 measures were used.

Sample Characteristics The initial sample4 at t1 consisted of N = 245 participants aged 16–61 (Mage = 28.37, SDage = 11.01, 61.3% female). All participants received an invitation to fill out the second questionnaire 4 weeks after t1 via e-mail. The sample size at t2 was reduced to N = 195 (Mage = 28.07, SDage = 10.63, 62.9% female) after eliminating 38 respondents who failed to participate at both times of measurement and 12 t2 data sets that could not be matched with any data from t1. Selection criterion prior to data assessment was regular engagement in sports. All participants fulfilled this criterion. All ensuing analyses are based on this final sample of N = 195. More than half of the participants were students (62.8%) but other professions were represented as well, making the sample diverse with regard to age and occupations.

This study is part of a complex multi-study research design. Study variables not relevant for this work are omitted from this report. The required sample size was estimated prior to data collection by using guidelines on sample size estimations for structural equation models (Wolf, Harrington, Clark, & Miller, 2013). For instance, under the assumption of three factors (i.e., one personality domain, perceived stress, and sports enjoyment) per model and standardized factor loadings of .50, a sample size of 150 was anticipated. However, this estimate was interpreted as a rather general guideline instead of a precise power analysis. We ultimately decided to keep the online survey open until participation dwindled significantly, also anticipating potential drop out after t1.

European Journal of Health Psychology (2020), 27(2), 45–54

Ó 2020 Hogrefe Publishing

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

The sample was diverse in terms of sports behavior. Participants had to indicate what type of sport they referred to while filling out the questionnaire. They were asked to choose the sport they performed most frequently. The most common sport types in female participants were running (20%) and fitness classes (26%), followed by weight training (11%), volleyball (8%), and dancing (7%). Most men referred to their sports experience in either football (soccer; 22%), running (12%), or fitness classes (12%), followed by weight training (11%) and cycling (9%). 30.7% of participants reported taking part in sports competitions regularly. Most participants engaged in sports at least once a week for more than 45 min. 49% of participants were members of a gym and 46.3% reported to be a member of a sports club.

Measures Sports Enjoyment Sports enjoyment was assessed using a modified version of the Questionnaire for the Assessment of Enjoyment in Physical Education (QUAEPE; Engels & Freund, 2019). The QUAEPE measures three facets of sports enjoyment, namely, pleasure (e.g., “This type of sport is a lot of fun”), flow (e.g., “While doing this type of sport I feel optimally strained”), and recovery (e.g., “After performing this type of sport I feel recovered from my everyday life”). Pleasure relates to the affective experience in sports (Engels & Freund, 2019). Flow is based on the flow experience (Csikszentmihalyi, 2013) and describes the perceived optimal physical demand and ideal process. Recovery describes a state of feeling balanced and relaxed by sports (Engels & Freund, 2019). Each facet was measured with three items. Answer options included “never,” “sometimes,” “often,” and “always.” For subsequent analyses, we used a global sports enjoyment score integrating all three facets (cf. Engels & Freund, 2019). The original QUAEPE was developed and validated to assess sports enjoyment in physical education classes. Therefore, the item content was altered slightly to assess the enjoyment of sports in a general sports context.5 A g-factor model fit the data well (see Table 1)6 and the data showed good internal consistency for the global sports enjoyment score (α = .83). Perceived Stress Perceived Stress was measured using the German version (Reis, Lehr, Heber, & Ebert, 2017) of the Perceived Stress Scale (PSS; Cohen et al., 1983). The scale (10 items) measures how stressful (e.g., “In the last month, how often have you felt nervous and “stressed”?”) and uncontrollable 5 6

Table 1. Models for CFAs and hypothesis testing w2

CFI

RMSEA

Sports Enjoyment

82.25

.95

.10

Perceived Stress

112.10

.91

.11

Neuroticism

100.26

.94

.08

Extraversion

95.79

.92

.08

Factor CFA models

Openness

88.98

.88

.09

Conscientiousness

86.84

.94

.07

Agreeableness

78.98

.89

.08

Hypothesis 1

205.86

151

.95

.04

Hypothesis 2a (N)

347.61

198

.92

.06

Hypothesis 2b (E)

374.94

198

.89

.07

Hypothesis 2c (O)

353.36

181

.87

.07

Hypothesis 2d (C)

315.25

196

.92

.06

Hypothesis 2e (A)

304.88

191

.91

.06

Hypothesis 3a (N)

234.73

178

.95

.04

Hypothesis 3b (E)

346.17

178

.85

.07

Hypothesis 3c (O)

268.06

162

.90

.06

Hypothesis 3d (C)

222.14

176

.96

.04

Hypothesis 3e (A)

231.05

171

.95

.04

Hypothesis 4a (N)

573.13

450

.92

.04

Hypothesis 4b (E)

637.76

450

.87

.05

Hypothesis 4d (C)

485.71

448

.97

.02

Note. N = 195. N = Neuroticism; E = Extraversion; O = Openness; C = Conscientiousness; A = Agreeableness.

(e.g., “In the last month, how often have you found that you could not cope with all the things that you had to do?”) someone has perceived their life over the course of the previous 4 weeks. The rating scale made use of five answering options (never, almost never, sometimes, fairly often, and very often). A one-factor model exhibited acceptable fit (see Table 1) and internal consistency was good (α = .89). Personality The German version (Borkenau & Ostendorf, 2008) of Costa and McCrae’s NEO Five-Factor Inventory was used to assess the Big Five personality dimensions neuroticism (e.g., “I often feel tense and nervous”), extraversion (e.g., “I like being the center of attention”), openness for experience (e.g., “I often try new and foreign dishes”), agreeableness (e.g., “I always try to act considerately and sensitively”), and conscientiousness (e.g., “I try to accomplish all of my tasks very conscientiously”). Each scale consists of 12 items and answers can be given on a 5-point rating scale, ranging from strongly disagree over disagree, neutral, and agree to strongly agree. As recommended by Marsh et al. (2010), we included correlated uniquenesses

The modified questionnaire can be obtained from the corresponding author via e-mail. All measurement and structural models were estimated using Mplus 8 (Muthén & Muthén, 1998–2017).

Ó 2020 Hogrefe Publishing

European Journal of Health Psychology (2020), 27(2), 45–54

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

(CUs) between items of the same personality facet within each factor when specifying the respective measurement models. With CUs included, model fit for neuroticism, extraversion, and conscientiousness was relatively good but for openness and agreeableness, it was not optimal (see Table 1). For openness, there was one item with a factor loading close to zero (item 38, “I believe that we should look to our religious authorities for decisions on moral issues”), so for all subsequent analyses, this item was dropped. Apparently, using the NEO-FFI, the measurement of openness and agreeableness did not function as well as expected in the present study. Internal consistencies in the present study were α = .89 (neuroticism), α = .80 (extraversion), α = .77 (openness for experience), α = .85 (conscientiousness), and α = .77 (agreeableness). Table 1 details information on model fit for all scales. Note that because of the rather small sample size (N = 195), we only conducted separate CFAs and refrained from testing the complete measurement model of all seven constructs simultaneously.

Hypotheses 4a–4e

Results Hypothesis 1 A model specifying the two factors sports enjoyment and perceived stress showed good fit to the data (Table 1). The latent correlation between sports enjoyment and perceived stress was r = .36 (p < .01), indicating support for Hypothesis 1.

Hypotheses 2a–2e The fit for all models investigating the relationships between perceived stress and the Big Five is detailed in Table 1. The CFI estimates for the models featuring extraversion and openness were lower than .90, so that the results from these models need to be interpreted with caution. The latent correlations between perceived stress and the Big Five were r = .75 (p < .01; neuroticism, H2a), r = .17 (p < .05; extraversion, H2b), r = .10 (ns; openness, H2c), r = .20 (p < .01; conscientiousness, H2d), and r = .01 (ns; agreeableness, H2e).

Hypotheses 3a–3e Again, the fit for the models including extraversion and openness was rather poor (see Table 1). The latent correlations between sports enjoyment and the Big Five were r = .24 (p < .01; neuroticism, H3a), r = .31 (p < .01; extraversion, H3b), r = .13 (ns; openness, H3c), and r = .23 (p < .01; conscientiousness, H3d). Between agreeableness and sports enjoyment, the latent correlation was r = .19 (p < .01; H3e). European Journal of Health Psychology (2020), 27(2), 45–54

We set up three mediation models, separately testing if perceived stress mediated the influence of neuroticism, extraversion, and conscientiousness on sports enjoyment. Since openness was not correlated with either perceived stress or sports enjoyment (see Results section above for Hypotheses 2c and 3c), and agreeableness was not related to perceived stress (see Results section for Hypothesis 2e), we did not set up corresponding mediation models for these two Big Five dimensions. Age was controlled for as a covariate because it correlated significantly with both perceived stress and sports enjoyment (first order correlations: r = .20 with perceived stress and r = .27 with sports enjoyment, both p < .01). We used one-tailed tests of significance because of the specific hypotheses. Figure 1 shows the standardized path coefficients (omitting the effects from the control variable age). Once more, the model for extraversion did not show a good fit to the data. All reported effects are standardized. For neuroticism, the total effect from N to sports enjoyment was estimated as .18, with a specific direct effect of .12 and an indirect effect of .30. Both the total and the indirect effects were significant at p < .01, whereas the direct effect was not significant, indicating a complete mediation. For extraversion, the total effect was .29, the direct effect was .19, and the indirect effect was .10. Both the total and the indirect effect were significant at p < .01 and the direct effect was significant at p < .05. For conscientiousness, the total effect was .20, the direct effect was .15, and the indirect effect was .05. Here, the direct and the indirect effect were significant at p < .05 and the total effect was significant at p < .01. Thus, there was evidence for partial mediation in the models including extraversion and conscientiousness.

Discussion The aim of the present study was to investigate if perceived stress and the Big Five affect the enjoyment of sports in adults. An interesting finding was that high amounts of perceived stress indicated less enjoyment of sports. This relationship had not been reported before and our findings highlight the impact of everyday stress on someone’s affective sports experience. It emphasizes the need for further research on ways to keep sports enjoyment high in stressed individuals in order to foster regular physical exercise. High values in neuroticism were associated with more perceived stress, while high values in extraversion and conscientiousness were related to lower levels of perceived stress. Our results show a negative relation between neuroticism and sports enjoyment, as well as significant positive Ó 2020 Hogrefe Publishing

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

Figure 1. Mediation models for neuroticism, extraversion, and conscientiousness (latent factors, standardized coefficients).

associations of extraversion and conscientiousness with sports enjoyment. The hypotheses concerning significant associations for openness with perceived stress (H2c) and sports enjoyment (H3c) had to be entirely dismissed. These findings were contrary to previous research (e.g., Britton et al., 2017; Wilson & Dishman, 2015), however, they might be due to the sample characteristics, which may explain the rather poor fit statistics for these two models: The mean for openness in the present study was higher than the mean of a representative quota sample for the German population reported in the manual for the NEO-FFI (Borkenau & Ostendorf, 2008). Higher average levels of openness might be related to the selectivity of the present sample. University students score higher on openness measures compared to non-students (Borkenau & Ostendorf, 2008). More than 50% of the present sample consisted of students. Furthermore, in the current study we required participants to be regularly physically active. Extant research reports higher levels of openness in athletes compared to non-athletes (Allen et al., Ó 2020 Hogrefe Publishing

2013). Therefore, a higher overall mean for openness in the present study might have provided difficulties in detecting a relation between this Big Five factor and perceived stress and sports enjoyment. The poor fits for the models including extraversion might be explained by the sample characteristics as well, with students and athletes scoring overall higher in extraversion as compared to the norm sample(s) (Allen et al., 2013; Borkenau & Ostendorf, 2008). Increasing the percentage of non-students in the sample in future research would provide an opportunity to investigate this assumption more thoroughly. Results on indirect effects of the Big Five on sports enjoyment through perceived stress were significant for neuroticism, extraversion, and conscientiousness. Partial mediations were found for extraversion and conscientiousness, while a full mediation of neuroticism on sports enjoyment through perceived stress was found, indicating that the effect of neuroticism on sports enjoyment is completely mediated through the perceived stress level. These findings highlight the importance of the stress level for the European Journal of Health Psychology (2020), 27(2), 45–54

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

relationship between these three Big Five factors and sports enjoyment. A high degree of neuroticism is associated with a high level of perceived stress which prohibits individuals from enjoying sports. On the other hand, individuals scoring high in extraversion and/or conscientiousness report lower levels of perceived stress which is associated with more enjoyment in sports. Age (included as a control variable) was negatively related to perceived stress and positively related to sports enjoyment, implying that the enjoyment of sports increases with age while older participants reported less perceived stress compared to the younger individuals. A negative relation between age and perceived stress is consistent with previous findings (e.g., Scott, Sliwinski, & Blanchard-Fields, 2013; Xu et al., 2015). Higher sports enjoyment in older participants is also in accordance with other research (e.g., McCarthy et al., 2008). Besides some poor model fits, another limitation of the present study was the drop-out rate between t1 and t2. The rate of non-responders at t2 was 15%, which is not uncommonly high, but 12 data sets from t2 could not be matched with data from t1, resulting in a total drop-out rate of 20%. Here, personalized URLs would have eliminated the problem of unmatched data-sets because participants’ answers at t2 can be automatically matched with the corresponding data-set of t1. A second study with more measurement points, ideally daily or at least weekly in form of a diary or using app-based assessment methods, would be beneficial to assess changes in the perception of stress and sports enjoyment in order to analyze the dynamic relationship between these variables more precisely. This would also provide an opportunity to investigate the stability of the construct sports enjoyment, i.e., the question as to how strongly sports enjoyment can be considered a trait versus a state-like construct. In addition, a multi-method approach to assess perceived stress could be aimed for. For example, cortisol levels (Jobin, Wrosch, & Scheier, 2014) or stressful live events (Cohen et al., 1983) could be assessed to increase objectivity and reduce external factors that might occur during online questionnaires. Overall, our findings highlight the relevance of perceived stress levels for the enjoyment of sports in individuals with different personality (Big Five) characteristics. These findings on the effect of perceived stress and personality on sports enjoyment may help understanding why some people do not like to engage in sports: We have provided empirical support that neuroticism is negatively associated with sports enjoyment via the mediating effect of higher perceived stress levels, while extraversion and conscientiousness are indirectly (through perceived stress) and directly related to more sports enjoyment. Further research is needed to identify ways to keep sports enjoyment high in individuals who experience stress, but usually report an average to high level of sports enjoyment. For example,

an intervention could help individuals identify what they enjoy in a sport and teach them to focus on these enjoyable aspects of sports (e.g., social interaction, distraction, using their muscles, clearing the mind, etc.). Furthermore, lowthreshold individual daily activities, such as yoga practices or mindfulness enhancing strategies, are known for their stress-reducing effect (Baer, Carmody, & Hunsinger, 2012; West, Otte, Geher, Johnson, & Mohr, 2004), and could therefore be considered as a basis for the development of a possible intervention program. Based on the current findings, the Big Five could be taken into account here as a fundamental description of an individual’s personality. Considering that the distinct Big Five factors can be seen as either facilitating OR impeding antecedents of perceived stress – but have been found to be rather stable and thus not easily malleable traits (e.g., Borkenau & Ostendorf, 2008) – it may be a promising avenue to tailor stress intervention programs to different personalities. For example, individuals with high levels of neuroticism report less enjoyment of team sports compared to emotionally more stable individuals (Allen et al., 2013) and are more worried about embarrassment in sports (Courneya & Hellsten, 1998). Highly introverted individuals, for example, report less enjoyment of competition in sports and participate less often in risky sports (Allen et al., 2013). Customization of an intervention program to different personalities might then help individuals maintain a stable level of sports enjoyment, even when daily stress increases, and ultimately foster long-term engagement in sports. The creation and validation of such an intervention could be an ensuing project to the research presented in this paper.

European Journal of Health Psychology (2020), 27(2), 45–54

Conclusion The present study contributes to our understanding of influencing factors of sports enjoyment. Perceived stress was found to negatively predict sports enjoyment. This is an important finding, considering that sports enjoyment predicts regular physical exercise which in turn is crucial for physical and mental health, especially during stressful times. Therefore, it is important for future research to analyze why perceived stress has negative effects on sports enjoyment and how to buffer this effect. Another finding of this study was that neuroticism was negatively, and extraversion, conscientiousness, and agreeableness were positively related to sports enjoyment. The relationship between the Big Five factors neuroticism, extraversion, and conscientiousness and sports enjoyment was mediated by perceived stress, indicating the relevance of the stress level on the experience of sports enjoyment. The findings of this study help to understand the experience in sports during stressful times, aiming at fostering long-term Ó 2020 Hogrefe Publishing

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

engagement in sports and buffering the negative effects of stress on our physical and mental health.

Questionnaire for measuring enjoyment in school sport in adolescence]. Diagnostica, 65(3), 166–178. https://doi.org/ 10.1026/0012–1924/a000222 Frederick, C. M., & Ryan, R. M. (1993). Differences in motivation for sport and exercise and their relations with participation and mental health. Journal of Sport Behavior, 16, 124. Gerber, M. (2008). Sport, Stress und Gesundheit bei Jugendlichen [Sport, stress, and health in adolescents]. Schorndorf, Germany: Hofmann. Janssen, I. (2007). Physical activity guidelines for children and youth. Applied Physiology, Nutrition, and Metabolism, 32, S109– S121. https://doi.org/10.1139/H07-109 Jobin, J., Wrosch, C., & Scheier, M. F. (2014). Associations between dispositional optimism and diurnal cortisol in a community sample: When stress is perceived as higher than normal. Health Psychology, 33, 382–391. https://doi.org/ 10.1037/a0032736 Kaiseler, M., Polman, R. C., & Nicholls, A. R. (2012). Effects of the Big Five personality dimensions on appraisal coping, and coping effectiveness in sport. European Journal of Sport Science, 12, 62–72. https://doi.org/10.1080/17461391.2010.551410. Lazarus, R. S. (2006). Stress and emotion: A new synthesis. New York, NY: Springer Publishing Company. Lippold, M. A., Davis, K. D., McHale, S. M., Buxton, O. M., & Almeida, D. M. (2016). Daily stressor reactivity during adolescence: The buffering role of parental warmth. Health Psychology, 35, 1027. https://doi.org/10.1037/hea0000352 Marsh, H. W., Lüdtke, O., Muthén, B., Asparouhov, T., Morin, A. J., Trautwein, U., & Nagengast, B. (2010). A new look at the Big Five factor structure through exploratory structural equation modeling. Psychological Assessment, 22, 471–491. https://doi. org/10.1037/a0019227 Marques, A., Sarmento, H., Martins, J., & Nunes, L. S. (2015). Prevalence of physical activity in European adults – Compliance with the World Health Organization’s physical activity guidelines. Preventive Medicine, 81, 333–338. https://doi.org/ 10.1016/j.ypmed.2015.09.018 McCarthy, P. J., Jones, M. V., & Clark-Carter, D. (2008). Understanding enjoyment in youth sport: A developmental perspective. Psychology of Sport and Exercise, 9, 142–156. https://doi. org/10.1016/j.psychsport.2007.01.005 Mouratidis, A. A., Vansteenkiste, M., Sideridis, G., & Lens, W. (2011). Vitality and interest–enjoyment as a function of classto-class variation in need-supportive teaching and pupils’ autonomous motivation. Journal of Educational Psychology, 103, 353–366. http://psycnet.apa.org/doi/10.1037/a0022773 Mullen, S. P., Olson, E. A., Phillips, S. M., Szabo, A. N., Wójcicki, T. R., Mailey, E. L., . . . McAuley, E. (2011). Measuring enjoyment of physical activity in older adults: Invariance of the physical activity enjoyment scale (paces) across groups and time. International Journal of Behavioral Nutrition and Physical Activity, 8, 1–9. https://doi.org/10.1186/1479-5868-8-103 Muthén, L. K., & Muthén, B. O. (1998–2017). Mplus User’s Guide (8th ed.). Los Angeles, CA: Muthén & Muthén. Papaioannou, A., Bebetsos, E., Theodorakis, Y., Christodoulidis, T., & Kouli, O. (2006). Causal relationships of sport and exercise involvement with goal orientations, per-ceived competence and intrinsic motivation in physical education: A longitudinal study. Journal of Sports Sciences, 24, 367–382. https://doi.org/ 10.1080/02640410400022060 Physical Activity Guidelines Advisory Committee. (2008). Physical activity guidelines advisory committee report (A1-H14). Washington, DC: US Department of Health and Human Services. Reiner, M., Niermann, C., Krapf, F., & Woll, A. (2013). Stress, Sport und Beschwerdewahrnehmung. Sportwissenschaft, 43, 264– 275. https://doi.org.10.1007/s12662-013-0294-8

References Allen, M. S., Greenlees, I., & Jones, M. V. (2013). Personality in sport: A comprehensive review. International Review of Sport and Exercise Psychology, 6, 184–208. https://doi.org/10.1080/ 1750984X.2013.769614 Ashford, B., Biddle, S., & Goudas, M. (1993). Participation in community sports centres: Motives and predictors of enjoyment. Journal of Sports Sciences, 11, 249–256. https://doi.org/ 10.1080/02640419308729992 Baer, R. A., Carmody, J., & Hunsinger, M. (2012). Weekly change in mindfulness and perceived stress in a mindfulness-based stress reduction program. Journal of Clinical Psychology, 68, 755–765. https://doi.org/10.1002/jclp.21865 Ball, J. W., Bice, M. R., & Parry, T. (2014). Adults’ motivation for physical activity: Differentiating motives for exercise, sport, and recreation. Recreational Sports Journal, 38, 130–142. https:// doi.org/10.1123/rsj.2014-0048 Borkenau, P., & Ostendorf, F. (2008). NEO-FFI: NEO-Fünf-Faktoren Inventar nach Costa und McCrae, Manual (2nd ed.) [NEO-FFI: NEO-Five-Factor Inventory by Costa and McCrae, Manual (2nd ed.)]. Göttingen, Germany: Hogrefe. Britton, D., Kavanagh, E., & Polman, R. (2017). The Perceived Stress Reactivity Scale for adolescent athletes. Personality and Individual Differences, 116, 301–308. https://doi.org/10.1016/j. paid.2017.05.008 Burg, M. M., Schwartz, J. E., Kronish, I. M., Diaz, K. M., Alcantara, C., Duer-Hefele, J., & Davidson, K. W. (2017). Does stress result in you exercising less? Or Does Exercising Result in You Being Less Stressed? Or Is It Both? Testing the Bi-directional StressExercise Association at the Group and Person (N of 1) Level. Annals of Behavioral Medicine, 51, 799–809. https://doi.org/ 10.1007/s12160-017-9902-4 Carver, C. S., & Connor-Smith, J. (2010). Personality and coping. Annual Review of Psychology, 61, 679–704. https://doi.org/ 10.1146/annurev.psych.093008.100352 Charles, S. T., Piazza, J. R., Mogle, J., Sliwinski, M. J., & Almeida, D. M. (2013). The wear and tear of daily stressors on mental health. Psychological Science, 24, 733–741. https://doi.org/ 10.1177/0956797612462222 Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385–396. Courneya, K. S., & Hellsten, L. A. M. (1998). Personality correlates of exercise behavior, motives, barriers and preferences: An application of the five-factor model. Personality and Individual Differences, 24, 625–633. https://doi.org/10.1016/S0191-8869 (97)00231-6 Csikszentmihalyi, M. (2013). Flow: The psychology of happiness. Munich, Germany: Random House e-books. Google Scholar. Deci, E. L., & Ryan, R. M. (1985). Intrinsic Motivation and Selfdetermination of Human Behaviour. New York, NY: Plenum Press. Engels, E. S., & Freund, P. A. (2018). Welche Faktoren beeinflussen das Erleben von Freude am Schulsport im Jugendalter? [Which factors influence the perception of sport enjoyment in physical education classes?] Zeitschrift für Sportpsychologie, 25, 68–78. https://doi.org/10.1026/1612-5010/a000230 Engels, E. S., & Freund, P. A. (2019). Entwicklung und Validierung eines Fragebogens zur Erfassung von Freude am Schulsport im Jugendalter (FEFS-J) [Development and Validation of a

Ó 2020 Hogrefe Publishing

European Journal of Health Psychology (2020), 27(2), 45–54

F. Dunker et al., Effects of Perceived Stress on Sports Enjoyment

Reis, D., Lehr, D., Heber, E., & Ebert, D. D. (2017). The German version of the Perceived Stress Scale (PSS-10): Evaluation of dimensionality, validity, and measurement in-variance with exploratory and confirmatory bifactor modeling. Assessment, 26(7), 1246–1259. https://doi.org/10.1177/1073191117715731 Rhodes, R. E., & Smith, N. E. I. (2006). Personality correlates of physical activity: A review and meta-analysis. British Journal of Sports Medicine, 40, 958–965. https://doi.org/10.1136/bjsm. 2006.028860 Rimmele, U., Seiler, R., Marti, B., Wirtz, P. H., Ehlert, U., & Heinrichs, M. (2009). The level of physical activity affects adrenal and cardiovascular reactivity to psycho-social stress. Psychoneuroendocrinology, 34, 190–198. https://doi.org/ 10.1016/j.psyneuen.2008.08.023 Robinson, M. D., Ode, S., Moeller, S. K., & Goetz, P. W. (2007). Neuroticism and affective priming: Evidence for a neuroticismlinked negative schema. Personality and Individual Differences, 42, 1221–1231. https://doi.org/10.1016/j.paid.2006.09.027 Roohafza, H., Feizi, A., Afshar, H., Mazaheri, M., Behnamfar, O., Hassanzadeh-Keshteli, A., & Adibi, P. (2016). Path analysis of relationship among personality, perceived stress, coping, social support, and psychological outcomes. World Journal of Psychiatry, 6, 248–256. https://doi.org/10.5498/wjp.v6.i2.248 Rovniak, L. S., Anderson, E. S., Winett, R. A., & Stephens, R. S. (2002). Social cognitive determinants of physical activity in young adults: A prospective structural equation analysis. Annals of Behavioral Medicine, 24, 149–156. https://doi.org/ 10.1207/S15324796ABM2402_12 Sanchez-Oliva, D., Sanchez-Miguel, P. A., Leo, F. M., Kinnafick, F. E., & García-Calvo, T. (2014). Physical education lessons and physical activity intentions within Spanish secondary schools: A self-determination perspective. Journal of Teaching in Physical Education, 33, 232–249. https://doi.org/10.1123/jtpe.2013-0043 Scanlan, T. K., & Simons, J. P. (1992). The construct of sport enjoyment. In G. C. Roberts (Ed.), Motivation in sport and exercise (pp. 199–215). Champaign, IL: Human Kinetics Books. Schlotz, W. (2013). Stress reactivity. In M. D. Gellman & J. R. Turner (Eds.), Encyclopedia of behavioral medicine. New York, NY: Springer. Schmid, S., Schröder, I., Eschenbeck, H., & Kohlmann, C. W. (2016). Stressbewältigung und körperlich-sportliche Aktivität bei Schülerinnen und Schülern [Students’ Coping with Stress and Physical Activity]. Sportunterricht, 65, 227–232. Scott, S. B., Sliwinski, M. J., & Blanchard-Fields, F. (2013). Age differences in emotional responses to daily stress: The role of timing, severity, and global perceived stress. Psychology and Aging, 28, 1076–1087. https://doi.org/10.1037/a0034000 Sipos, E., Jeges, S., & Tóth, Á. (2015). Sport, sense of coherence, and self-esteem among 16 and 17-year-olds. European Journal of Mental Health, 10, 62–78. https://doi.org/10.5708/EJMH.10. 2015.1.4 Spada, M. M., Nikčević, A. V., Moneta, G. B., & Wells, A. (2008). Metacognition, perceived stress, and negative emotion. Personality and Individual Differences, 44, 1172–1181. https://doi. org/10.1016/j.paid.2007.11.010 Stein, C., Fisher, L., Berkey, C., & Colditz, G. (2007). Adolescent physical activity and perceived competence: Does change in activity level impact self-perception? Journal of Adolescent Health, 40, 462-e1–462-e8. https://doi.org/10.1016/j.jadohealth. 2006.11.147 Steptoe, A., Wardle, J., Pollard, T. M., Canaan, L., & Davies, G. J. (1996). Stress, social support and health-related behavior: A study of smoking, alcohol consumption and physical exercise. Journal of Psychosomatic Research, 41, 171–180. Sutin, A. R., Stephan, Y., Luchetti, M., Artese, A., Oshio, A., & Terracciano, A. (2016). The five-factor model of personality and

physical inactivity: A meta-analysis of 16 samples. Journal of Research in Personality, 63, 22–28. http://doi.org/10.1016/ j.jrp.2016.05.001 Ullrich-French, S., & Smith, A. L. (2009). Social and motivational predictors of continued youth sport participation. Psychology of Sport and Exercise, 10, 87–95. https://doi.org/10.1016/ j.psychsport.2008.06.007 Wankel, L. M., & Berger, B. G. (1990). The psychological and social benefits of sport and physical activity. Journal of Leisure Research, 22, 167–182. https://doi.org/10.1080/00222216.1990. 11969823 West, J., Otte, C., Geher, K., Johnson, J., & Mohr, D. C. (2004). Effects of Hatha yoga and African dance on perceived stress, affect, and salivary cortisol. Annals of Behavioral Medicine, 28, 114–118. https://doi.org/10.1207/s15324796abm2802_6 Williams, D. M., Papandonatos, G. D., Napolitano, M. A., Lewis, B. A., Whiteley, J. A., & Marcus, B. H. (2006). Perceived enjoyment moderates the efficacy of an individually tailored physical activity intervention. Journal of Sport and Exercise Psychology, 28, 300–309. https://doi.org/10.1123/jsep.28.3.300 Wilson, K. E., & Dishman, R. K. (2015). Personality and physical activity: A systematic review and meta-analysis. Personality and Individual Differences, 72, 230–242. https://doi.org/ 10.1016/j.paid.2014.08.023 Wolf, E. J., Harrington, K. M., Clark, S. L., & Miller, M. W. (2013). Sample size requirements for structural equation models: An evaluation of power, bias, and solution propriety. Educational and Psychological Measurement, 73, 913–934. https://doi.org/ 10.1177/0013164413495237 Woods, C. B., Tannehill, D., & Walsh, J. (2012). An examination of the relationship between enjoyment, physical education, physical activity and health in Irish adolescents. Irish Educational Studies, 31, 263–280. https://doi.org/10.1080/03323315. 2012.710068 Xu, X., Bao, H., Strait, K., Spertus, J. A., Lichtman, J. H., D’Onofrio, G., . . . Bueno, H. (2015). Sex differences in perceived stress and early recovery in young and middle-aged patients with acute myocardial infarction. Circulation, 131, 614–623. https://doi.org/10.1161/CIRCULATIONAHA.114.012826 Zohar, D. (1999). When things go wrong: The effect of daily work hassles on effort, exertion and negative mood. Journal of Occupational and Organizational Psychology, 72, 265–283. https://doi.org/10.1348/096317999166671

European Journal of Health Psychology (2020), 27(2), 45–54

History Received October 21, 2019 Revision received March 20, 2020 Accepted March 23, 2020 Published online July 1, 2020 Authorship Eliane S. Engels is now with the Department of Sport Science and Sport, University Erlangen-Nuremberg, Germany. ORCID Freya Dunker https://orcid.org/0000-0003-2827-3214 Freya Dunker Institute of Psychology Leuphana University of Lüneburg 21335 Lüneburg Germany freya.dunker@gmx.de

Ó 2020 Hogrefe Publishing

Original Article

Short Form of the State-Trait Anger Expression Inventory-2 Ana N. Tibubos1,2 , Karin Schermelleh-Engel2, and Sonja Rohrmann3 1

Department of Psychosomatic Medicine and Psychotherapy, University Medical Center, Johannes Gutenberg University Mainz, Germany

Department of Psychological Research Methods and Evaluation, Institute of Psychology, Goethe University, Frankfurt am Main, Germany

Department of Differential Psychology and Psychological Diagnostics, Goethe University, Frankfurt am Main, Germany

Abstract: The purpose of the present study was to develop a short form of the State-Trait Anger Expression Inventory-2 (STAXI-2) based on the German STAXI-2. Item selection was performed based on exploratory factor analyses (EFA) using descriptive statistical parameters and content-related considerations on calibration samples (N1 = 215, N2 = 310). The factorial structure of the final extracted scales was validated via confirmatory factor analyses (CFA) (N3 = 216, N4 = 310). Overall, results present an economic and reliable questionnaire with a total length of 24 items: State Anger short scales Feeling Angry, Verbal Anger Impulse, and Physical Anger Impulse (3 items each), that can be aggregated to a total State Anger score, as well as Trait Anger short scales Angry Reaction (3 items), Anger Expression-In, Anger Expression-Out, and Anger Control (4 items each). The structure of State Anger is identical to the German long version with improved internal consistency in the short form. Regarding the Trait scales, critique on the STAXI-2 has been taken into account resulting in the elimination of the subscale Trait Temperament due to redundancy with Trait Anger Expression-Out and for economic reasons. Other than that, the structure has remained the same. In addition, strict measurement invariance was established based on multi-group CFA for both the State and the Trait scales across gender and age groups, which has not been investigated for STAXI-2 versions to date. Keywords: anger, state, trait, anger tendency, anger expression

Anger is one of the basic human emotions (Averill, 1982; Hodapp & Schwenkmezger, 1993; Spielberger & Reheiser, 2010; Weber, 2002). It is an everyday emotional reaction most frequently experienced in interpersonal interactions caused by frustration, the experience of blocked satisfaction of needs, disturbed goal achievement, or provocation (Izard, 1991; Novaco, 1975, 1978). According to Spielberger (1979), fundamental emotional dimensions such as fear, anger, and curiosity are conceived as emotional states as well as personality dimensions. In his Trait-State Model Spielberger (1966; 1972) distinguishes on the one hand between the current state of anger (State Anger), on the other hand between the habitual tendency to experience situations as anger provocative and to feel anger frequently and intensively (Trait Anger) and additionally between different forms of habitual anger expressions (Spielberger, 1988). The tendency to experience anger and the way in which it is dealt with can have a strong negative impact on subjective well-being and the quality of social relationships (cf. Schmitt & Altstötter-Gleich, 2010). Furthermore, there is a connection between dispositional anger components and psychosomatic disorders, which is – however – moderated by coping skills and psychosocial resources (Weber, 2002). Various researches suggest that anger and anger expression are important psychosocial factors in the Ó 2020 Hogrefe Publishing

etiology and course of hypertension and coronary heart disease. It has been proposed that negative health outcomes, like hypertension, are associated with the inhibition of angry feelings, while positive health outcomes seem to be related to anger control (for a summary see Potegal, Stemmler, & Spielberger, 2010). In particular, the pattern of anger, hostility, and aggression is said to be a risk factor for coronary heart disease (CHD) (Chesney & Rosenman, 1985; Spielberger & Reheiser, 2010). As a significant risk factor for CHD in modern industrial society, emotional anger is of special importance. For this reason, anger has been explored intensely, and there is a high interest in recording anger and individual differences in the tendency and expression of it. In 1988, Spielberger published the State-Trait Anger Expression Inventory (STAXI) for recording State Anger, Trait Anger, and three forms of Anger Expression (Anger-Out, Anger-In, and Anger Control), which gained worldwide recognition. In 1999, Spielberger published a revised and expanded version of the STAXI, the STAXI-2, which developed into the most widely used self-assessment tool for anger (cf. Fernandez, 2013). In the STAXI-2 version proposed by Spielberger, new features were the Anger Control subscales: Anger Control-In and Anger ControlOut. In order to make the procedure accessible to German-speaking countries, a German-language version of European Journal of Health Psychology (2020), 27(2), 55–65 https://doi.org/10.1027/2512-8442/a000049

the STAXI-2 (Rohrmann et al., 2013) was developed, which is based on both the American STAXI-2 and the Germanlanguage STAXI version by Schwenkmezger, Hodapp, and Spielberger (1992). The German adaptation of the STAXI-2 can be regarded as an objective and reliable procedure whose validity could be sufficiently confirmed. It contains 51 items. State Anger is assessed with 15 items measuring the intensity of the subjective state anger at a given time or in a defined situation. The state items can be assigned to three subscales: Feeling Angry, Verbal Anger Impulse, and Physical Anger Impulse. The Trait Anger scale consists of 10 items and measures the tendency to react in an anger-provoking situation by increasing the level of anger. It can be divided into two subscales: Angry Temperament and Angry Reaction. Another 26 items assess the habitual tendency to express or control the experienced anger in a certain way. The items can be assigned to two Anger Expression scales (Anger ExpressionOut, Anger Expression-In) and an Anger Control scale. The Anger Expression-Out scale measures the frequency with which an individual directs anger at other people or objects in their environment while the Anger Expression-In scale captures the frequency with which an individual suppresses anger or does not react outwardly. The Anger Control scale is an indicator of the frequency of attempts to control anger or to not let anger arise. This scale can be divided into two subscales assessing the control of anger by calming or relaxing (Anger Control-In) or the control of anger expression (Anger Control-Out), but is not recommended. A main difference from the English version refers to the lack of empirical evidence of the divided Anger Control scale into -In and -Out. Thus, Rohrmann et al. (2013) recommend the use of the aggregated Anger Control scale for the German version. Furthermore, as already stated in the first STAXI version (Schwenkmezger et al., 1992) and criticized by Krohne and Hock (2007, p. 297), a very high intercorrelation (r .70) between the Trait Anger scale and the habitual Anger Expression scale Anger Expression-Out was found by Rohrmann et al. (2013).

The Present Study With 51 items the German STAXI-2 is quite extensive and time-consuming. Thus, the aim of the present study was to develop an economic short form reduced by about half of the items of the original questionnaire without major losses in reliability and validity. Therefore, in a first step, we developed short forms of the State and Trait scales of the STAXI-2 in calibrations samples. Second, we confirmed the new short form structure in validation samples. Additionally, we provided thorough psychometric properties of the STAXI-2 short form demonstrating its quality as a European Journal of Health Psychology (2020), 27(2), 55–65

A. N. Tibubos et al., STAXI-2 Short Form

reliable and valid screener for anger experience, anger expression, and control. For state anger an experimental validation including an anger-inducing situation was performed. Finally, we tested measurement invariance across gender and age groups, which has not been investigated to date for the German STAXI or STAXI-2 version, respectively.

Method STAXI-2 The German version of the State scale (S-Ang) of the STAXI-2 comprises the three subscales: Feeling Angry (S-Ang/F), Verbal Anger Impulse (S-Ang/V), and Physical Anger Impulse (S-Ang/P), each measured by five items. The Trait scale consists of 36 items, divided into four main and four subscales: Trait Anger (T-Ang) as well as the Anger Expression scales Anger Expression-Out (AX-O), Anger Expression-In (AX-I), and Anger Control (AC), whereby Trait Anger is divided into Angry Temperament (T-Ang/T) and Angry Reaction (T-Ang/R), and Anger Control is divided into Anger Control-Out (AC-O) and Anger Control-In (AC-I). All item responses are recorded on a 4-point rating scale. Responses to State Anger items represent ratings of intensity (1 = not at all, 2 = a little, 3 = rather, 4 = very), responses to Trait Anger items ratings of frequency (1 = almost never, 2 = sometimes, 3 = often, 4 = almost always). Internal consistency coefficients (Cronbach’s alpha (α); Cronbach, 1951; Guttman, 1945) of the Trait Anger, Anger Expression, and Anger Control scales (including subscales) range between α = .79 and α = .91. For the State scales (including subscales), values are between α = .74 and α = .87. Retest reliability (Brown, 1910; Guttman, 1945) was estimated using three different time intervals 6 weeks apart. For the State Anger scale, estimates vary between rtt = .14 and rtt = .29 for the individual subscales depending on the time interval. For the Trait Anger scales estimates range between rtt = .67 and rtt = .78 and between rtt = .63 and rtt = .81 for habitual anger expression and anger control. The validity of the German STAXI-2 scales has been proven in various studies (Etzler, Rohrmann, & Brandt, 2014; Rohrmann et al., 2013; Tibubos, Schnell, & Rohrmann, 2013; Tibubos, Pott, Schnell, & Rohrmann, 2014).

Samples and Procedure First, two equivalent samples were created for the State and Trait scales, a calibration and a validation sample each. The calibration samples were used as a basis for the decision to shorten the STAXI-2 scales (see Electronic Supplementary Materials, ESM 1, Table E1 for detailed descriptive statistics of the calibration samples). The validation samples were Ó 2020 Hogrefe Publishing

A. N. Tibubos et al., STAXI-2 Short Form

used for the cross-validation of the factorial structure and the determination of the psychometric characteristics of the STAXI-2 short form. In order to provide a first experimental validation of the State Anger short form, an anger-inducing situation was investigated including pre- and post-measurement of state anger. Participants were asked to put themselves in an anger-inducing situation in which they were subjected to an unfair examination evaluation. Samples for the Development of the STAXI-2 State Short Scales The short form of the State scale was developed as part of an experimental online study in which 435 students of a German Distance University (346 women) aged 32.2 ± 10.7 years participated. Since variance in state anger is very low in neutral situations, we needed post-treatment data of an anger-inducing situation. Since gathering a large sample of data based on an anger-induction experiment is difficult in a non-student sample, the state anger data is based on a student sample. At the beginning of the study, the subjects of the online study worked on the Trait and State versions of the STAXI-2, in the post-measurement only state anger was assessed. Before compilation of the calibration and validation sample of the State scales, participants which had not given any information in the post-measurement (N = 4) were excluded from the analysis sample. The sample was then divided into two random halves, means and standard deviations of both samples did not differ significantly (t429 = 1.38, p = .17). This resulted in a calibration sample of N = 215 and a validation sample of N = 216 persons based on the post-manipulation. Samples for the Development of the STAXI-2 Trait Short Scales With the extensive Trait scales, the goal was to shorten them to a final length of a maximum of four items per scale. For the development of the Trait Anger scales, three existing datasets of the German-language STAXI-2 (Rohrmann et al., 2013) of a total of N = 2,014 individuals were used: a representative sample of the general population, a patient and a prisoner sample. The representative sample consisted of 1,889 subjects (age: 49.7 ± 18.2 years). Furthermore, 68 patients (age: 71.3 ± 11.1 years) with coronary heart disease associated with elevated anger levels (cf. Chesney & Rosenman, 1985; Spielberger & Reheiser, 2010) were used from the validation sample for data analysis. Finally, data from 57 male prisoners (age: 37 ± 9.2 years) from another validation sample were also used, as it was demonstrated that delinquency is associated with higher values in Trait

Ó 2020 Hogrefe Publishing

Anger and Anger Expression-Out and lower values in Anger Control (Cherek, Lane, Dougherty, Moeller, & White, 2000). Since aggressive behavior which is the behavioral component of anger appears in approximately 25% in a community sample (Lewis & Rudolph, 2014), a ratio of approximately 4:1 of the representative sample to the extreme samples (patients and prisoners) was established to increase variance. Thus, a calibration sample of N = 310 and a validation sample of N = 310 were available for the analyses of the Trait Anger scales, each consisting of 250 individuals from the calibration sample, 27 individuals from the prisoner sample, and 33 individuals from the clinical sample. Selection of Items for the Short Form The selection of items for the short scales was based on content-related aspects as well as on statistical characteristics, that is, factor loadings and item easiness (Kelava & Moosbrugger, 2020; see also Hecht, Hardt, Driver, & Voelkle, 2019; Moosbrugger, Gäde, Schermelleh-Engel, & Rauch, 2020). Items with cross-loadings on different factors were eliminated as well as items with low factor loadings. Additionally, items with lower or higher easiness compared to the other items of the scale were also eliminated (see ESM 2 for a detailed explanation of how to calculate item easiness). The aim of the content analysis was, on the one hand, to reduce redundancies by retaining only one of several very similarly formulated items. Furthermore, care was taken to ensure that the meaningfulness of the individual scales remained as broad as possible. The STAXI-2 short form was developed – analogous to the STAXI-2 – based on classical test theory. We used congeneric measurement models for all scales. These models assume uni-dimensionality for all items supposed to measure the same scale and uncorrelated error variables, and all parameters (factor loadings, intercepts, error variances) are freely estimated (see, e.g., Raykov & Marcoulides, 2011). Exploratory factor analysis (EFA) was used to reduce the item pool and to generate the scale structure of the STAXI-2 short form, while confirmatory factor analysis (CFA) was used for model validation. In addition, the measurement invariance was examined on sub-samples of group (calibration, validation), gender, and age (see ESM 2 for Mplus syntax of the EFA and measurement invariance testing using CFA). In order to reduce the number of items, a separate EFA with oblique rotation due to expected correlated factors was initially performed for both the State and the Trait variables using the calibration samples. The items for the short forms were then selected on the basis of factor loadings and theoretical considerations. In addition, item easiness and item variances were also taken into account. Further details are described below.

European Journal of Health Psychology (2020), 27(2), 55–65

Statistical Analyses For analyses on the level of manifest variables, the item scores were summed up to scale scores. Factor loadings were estimated both for the exploratory factor analyses (EFA) and for the confirmatory factor analyses (CFA) using the robust maximum likelihood method (MLR) of the Mplus program, version 7.4 (Muthén & Muthén, 1998–2015). This robust estimation method is suitable for items with four response categories if the variables do not deviate too much from the normal distribution (see Rhemtulla, BrosseauLiard, & Savalei, 2012), which was the case here. MLR makes corrections to the w2-value and the standard error estimates compared to the commonly used ML estimator. We used full information maximum likelihood (FIML) within MLR taking missing values into account when estimating the parameters. In all datasets used, there were isolated cases in which individual values of the STAXI-2 items were missing. However, these formed only a small proportion of the data: for the Trait scales 99.06% and for the State scales even 99.53% of all values were available. Based on the EFA results of the calibration samples, the following CFA models were conducted in the validation samples: For the State items, a correlated three-factor model, and for the Trait items, a correlated four-factor model was performed.

Model Fit The Yuan-Bentler (YB)-corrected w2-test (Yuan & Bentler, 2000) for non-normally distributed variables and various descriptive fit criteria were used to assess the quality of the EFA and CFA models. A good model fit is indicated by a nonsignificant YB-w2 (p > .05) as well as RMSEA .05, CFI .97, and SRMR < .08, while an acceptable model fit is indicated by a w2/df 3, RMSEA .08, CFI .95, and SRMR .10 (Hu & Bentler, 1999; Schermelleh-Engel, Moosbrugger, & Müller, 2003). The scaled w2-difference test was used to assess the model comparisons (cf. Muthén & Muthén, 1998–2015, p. 489; Satorra & Bentler, 2001). Invariance Tests To check the measurement invariance using multi-group analyses, the calibration and validation samples were recombined and divided into partial samples by gender (male/female) and age ( or > 40 years) (Schieman, 1999; Stone, Schwartz, Broderick, & Deaton, 2010). Only if measurement invariance is given, it is possible to attribute mean value differences and differences in the variances of the items (and thus also of the scales) to differences in the latent anger constructs. The meaning of each invariance assumption and their consequences are explained in the following paragraph. European Journal of Health Psychology (2020), 27(2), 55–65

A. N. Tibubos et al., STAXI-2 Short Form

Each examination was carried out using four hierarchically nested models: (1) configural invariance, (2) metric invariance, (3) scalar invariance, and (4) strict invariance (cf. Gregorich, 2006; Kline, 2016; Meredith, 1993). When assuming configural invariance the factorial structure across groups is equal. Only if the factor model in the groups to be compared has the same structure with the same pattern of factor loadings are the estimated parameters comparable across the groups. For metric invariance, the model requires corresponding factor loadings to be equal across groups. If this condition is met, it is guaranteed that the factors have the same meaning across the groups which means that identical relationships between the factors and relevant criterion variables across groups exist (see, e.g., Gregorich, 2006). Scalar invariance additionally assumes the item intercepts across the groups to be equal. If the intercepts are the same, the differences in the mean values of the items are only due to the differences in the mean values of the latent variables. This means no systematically higher or lower valued item responses occur in one of the groups compared with the other. If the error variances across the groups are also the same (strict invariance), differences in the variances of the items across the groups can only be attributed to differences in the variances of the latent variables. A nonsignificant w2-difference (p .01) indicates measurement invariance (MI) among the nested models. As the w2-statistic is sensitive to sample size, we additionally focus on the differences ΔCFI and ΔRMSEA, with difference scores .01 indicating model invariance (Cheung & Rensvold, 2002; Putnick & Bornstein, 2016). If one of the invariance tests does not apply to individual items, partial invariance can be tested by removing a parameter restriction for such an item (Byrne, Shavelson, & Muthén, 1989; Steenkamp & Baumgartner, 1998). At least this must be guaranteed, so that group differences can be interpreted meaningfully.

Results Development of the State Anger Short Scales The EFA with robust maximum likelihood estimation (MLR) and oblimin rotation over the 15 State items did not yield a clear result. The descriptive criteria were more in favor of two factors: two eigenvalues were greater than 1, the scree plot suggested the extraction of two factors and parallel analysis also supported the extraction of two factors (empirical eigenvalues: 7.119, 3.298, 0.892; eigenvalues of parallel analysis: 1.480, 1.369, 1.287). The w2-test of the EFA showed an unacceptable model fit for two factors (YB-w2 = 257.34, df = 76, p < .01, RMSEA = .105, Ó 2020 Hogrefe Publishing

A. N. Tibubos et al., STAXI-2 Short Form

Table 1. Descriptive statistics, item easiness, and factor loadings of the State Anger items based on the confirmatory factor analysis of the validation sample (N = 216) Scale Feeling angry

Verbal anger impulse

Physical anger impulse

Item S-Ang/F-1

Item content Upset

3.38

0.71

.79

0.82

S-Ang/F-6

Angry

3.49

0.70

.83

0.82

S-Ang/F-11

Sulkily

3.47

0.80

.82

0.78

S-Ang/V-2

Shout at

2.38

1.06

.46

0.84

S-Ang/V-5

Swear

2.86

1.04

.61

0.72

S-Ang/V-15

Yell

2.25

1.14

.41

0.84

S-Ang/P-7

Kick

1.45

0.84

.15

0.90

S-Ang/P-10

Hit

1.35

0.76

.11

0.94

S-Ang/P-12

Hurt

1.38

0.78

.12

0.94

Note. M = item mean; SD = standard deviation; p = item easiness; λ = factor loading. The item abbreviations correspond to the nomenclature in the test manual of the State-Trait Anger Expression Inventory (STAXI-2).

CFI = .912, SRMR = .043), but a satisfactory fit for three factors (YB-w2 = 121.85, df = 63, p < .01, RMSEA = .066, CFI = .971, SRMR = .021). The loading pattern of the three-factorial solution corresponded to the expectations regarding the assignment of the items to the factors Feeling Angry, Verbal Anger Impulse, and Physical Anger Impulse (see ESM 1, Table E2). According to the item selection criteria described above, six items were eliminated from the original state item set, two per subscale. In all three scales, some items had cross-loadings on different (non-target) factors or had lower factor loadings compared to the other items on the scale. Additionally, two items of Feeling Angry were eliminated (“furious,” “angry”) and two items of Verbal Anger Impulse (“swearing,” “scolding”) because of similar item wordings. For reasons of content validity, the core item “I am angry” was retained, although it showed a high easiness (> .80). The final short form of the State scales finally contained nine items (see Table 1). Following item selection, a CFA was used to check whether the nine selected items are based on the three State Anger scales. The CFA with the calibration sample showed a satisfactory model fit with YB-w2 = 47.96, df = 24, p < .01, RMSEA = .068, CFI = .976 and SRMR = .039. The results of the CFA of the validation sample are given in Table 1 and indicate, as expected, that the easiness

coefficients of items that involve more socially undesirable behavior are relatively low. This concerns the items on the Physical Anger Impulse scale, while Feeling Angry as a more socially desirable behavior contains relatively easy items. The factor loadings of the items are high with coefficients ranging between λ = .72 and λ = .94. In order to judge total score interpretability, omega hierarchical (ωH) is often used to estimate the percentage of variance in observed scores due to variance on a single common latent variable, while omega total (ωT) is an estimate of the total true-score variance of a general factor and all specific or group factors compared to the observed variance (McDonald, 1999; Reise, Bonifay, & Haviland, 2013; Zinbarg, Revelle, Yovel, & Li, 2005). The results show that both reliability coefficients are quite high with ωH = .805 and ωT = .972, meaning that .805/.972 = 83% of the reliable variance in state anger scores is due to a general factor. Therefore, the three subscales can be added up to a total score that corresponds to a global measure of the anger state (State Anger) (Schermelleh-Engel & Gäde, 2020). In line with the long version, highest mean scores were observed for angry feelings, followed by verbal and physical anger impulse. The reliability coefficients of the State short scales were high with ω .84 and even higher than the internal consistency of the original version (α .74). Also, the correlation pattern of the short scales corresponded to

Table 2. Reliability (McDonald’s ω), descriptive statistics and intercorrelations of the State Anger short scales (lower triangular matrix) and the State Anger scales of the long version (upper triangular matrix), and correlations between the two versions (main diagonal), based on the validation sample (N = 216) ω

(1) Feeling angry

0.85

10.13

(2) Verbal anger impulse

0.84

7.27

(3) Physical anger impulse

0.95

4.19

2.34

State scale

(1)

(2)

(3)

2.03

0.96

0.72

0.32

2.84

0.61

0.97

0.62

0.21

0.49

0.97

Note. All correlations are significant (p < .01).

Ó 2020 Hogrefe Publishing

European Journal of Health Psychology (2020), 27(2), 55–65

A. N. Tibubos et al., STAXI-2 Short Form

that of the long version. The State short scales correlated very highly with the State scales of the long version of the STAXI-2 (r .96), which indicates that the nine items of the short scales represent the original state scales very well and reflect the essential contents of these three subscales sufficiently well (see Table 2).

Invariance Tests of the State Scales A two-group comparison of the validation sample with the calibration sample, in which all factor loadings, intercepts, and error variances were defined as invariant over the two groups, showed a good model fit with YB-w2 = 94.63, df = 69, p = .02, RMSEA = .042, CFI = .986 and SRMR = .062. This means that both samples are based on the same model with identical factor loadings, intercepts, and error variances, so that the two samples could be combined into a common sample (N = 431) (see Table 3). As the other results of the invariance tests show, there was strict invariance regarding the age groups. Partially strict invariance was established in the gender groups because the intercept of the item S-Ang/V-15 (“I could yell”) has a lower value in the female group than in the male group. Finally, initial evidence for the experimental validity of the State Anger scales was demonstrated. There was a significant increase in the State Anger value due to the anger manipulation (Mpre = 1.16 ± 0.40; Mpost = 2.43 ± 0.66; t(430) = 38.45, p < .01).

Development of the Trait Anger Short Scales An EFA with MLR estimation and oblimin rotation over the 36 Trait items showed that the data was based on either four (YB-w2 = 905.06, df = 492, p < .01) or five factors (YB-w2 = 778.23, df = 460, p < .01) (see ESM 1, Table E3). Seven eigenvalues were greater than one. The scree plot suggested the extraction of four or five factors, but the parallel analysis of only four factors (empirical eigenvalues: 9.722, 5.205, 2.085, 1.783, 1.285; eigenvalues of the parallel analysis: 1.716, 1.625, 1.557, 1.502, 1.449). Since the factor Angry Temperament correlated very highly with both Angry Reaction and Anger Expression-Out, this factor was completely eliminated. Furthermore, a subdivision of the Anger Control items into the two subscales Anger Control-In versus Anger Control-Out, as assumed in the US American original of the questionnaire, was just as little shown in the available data as in the original validation of the German-language STAXI-2 (Rohrmann et al., 2013). Therefore, only one Anger Control scale was used in the further analyses. This resulted in four scales: Angry Reaction, Anger Expression-Out, Anger Expression-In, and Anger Control. In total, the four scales of the questionnaire were reduced by 21 items. The selection of the items for the short scales was analogous to the selection of the State items. In addition to the elimination of the scale Angry Temperament (5 items), 13 other items were eliminated due to statistical characteristics: eight items due to double loadings, four of them of the Anger Expression-Out scale and four of them

Table 3. Invariance tests of the state variables over samples formed by chance, age, and gender ΔYB-w2

Δdf

p-value

8.22

.223

1.48

.961

6.12

.728

YB-w2

96.99

.000

0.069

0.973

M2 Metric

104.07

.0001

0.066

0.972

M2 vs. M1

M3 Scalar

104.82

.0003

0.059

0.975

M3 vs. M2

M4 Strict

94.69

.0220

0.042

0.986

M4 vs. M3

M1 Configural

87.82

.0004

0.062

M2 Metric

97.02

.0003

0.061

0.975

M2 vs. M1

9.80

.133

M3 Scalar

112.59

.0000

0.064

0.970

M3 vs. M2

15.56

.016

M4 Strict

111.10

.0010

0.053

0.976

M4 vs. M3

11.18

.264

Model

p-value

RMSEA

CFI

Model comparison

Random sample (Calib.: N = 215, Valid.: N = 216) M1 Configural

Age ( 40 years: N = 315, > 40 years: N = 116) 0.977

Gender (women: N = 344, men: N = 87) M1 Configural

99.76

.0000

0.071

0.972

M2 Metric

103.30

.0001

0.065

0.974

M2 vs. M1

5.55

.476

M3 Scalar

120.61

.0000

0.068

0.968

M3 vs. M2

17.15

.009

M3a Partial scalar

109.12

.0001

0.063

0.973

M3a vs. M2

5.90

.316

M4 Partial strict

115.93

.0003

0.057

0.979

M4 vs. M3a

12.06

.210

Note. YB-w2 = Yuan-Bentler-scaled w2-test statistic; CFI = comparative fit index; RMSEA = root mean-square error of approximation; ΔYB-w2 = scaled w2-difference test; Δdf = difference in degrees of freedom; Calib. = calibration sample; Valid. = validation sample.

European Journal of Health Psychology (2020), 27(2), 55–65

Ó 2020 Hogrefe Publishing

A. N. Tibubos et al., STAXI-2 Short Form

Table 4. Descriptive statistics, item easiness, and factor loadings of the Trait Anger items based on the confirmatory factor analysis of the validation sample (N = 310) Scale Trait angry reaction

Anger expression-out

Anger expression-in

Anger control

T-Ang/R-4

Furious

2.36

0.84

.45

0.62

T-Ang/R-6

Irate

2.04

0.72

.34

0.69

T-Ang/R-7

Anger

2.29

0.75

.43

0.73

AX-O-6

Nasty remarks

1.45

0.62

.14

0.66

AX-O-12

Quarrel

1.76

0.81

.25

0.55

AX-O-14

Say mean things

1.50

0.72

.16

0.67

AX-O-18

Attack

1.32

0.59

.10

0.71

AX-I-11

Not talk about it

2.05

0.86

.35

0.51

AX-I-15

Not perceptible

2.24

0.83

.41

0.86

AX-I-19

Dissemble

2.29

0.86

.42

0.77

AX-I-26

Flip inside

2.09

0.85

.36

0.51

AC-I-7

Alleviate anger

2.82

0.85

.60

0.65

AC-I-10

Calm down

3.06

0.80

.68

0.67

AC-I-17

Keep calm

2.66

0.90

.55

0.58

AC-O-23

Control oneself

2.96

0.89

.65

0.64

Item

Item content

of the Anger Control scale, and five items due to low factor loadings, including one item of the Angry Reaction scale, three items of the Anger Expression-In scale and one item of the Anger Control scale. In addition, three items were sorted out on the basis of similar item wordings: one item each in the scales Angry Reaction (“It upsets me”), Anger Expression-In (“could burst,” “freak out”), and Anger Control (“soothe”, “calm down”). The final short form of the Trait scales thus contains 15 items (see Table 4), with Trait Anger being recorded by three items and the three anger management scales Anger Expression-Out, Anger Expression-In, and Anger Control being measured by four items each. A CFA was performed on the validation sample to verify the factorial structure and the model fit. The CFA showed a satisfactory model fit with YB-w2 = 147.19, df = 84, p < .01, RMSEA = .049, CFI = .936, and SRMR = .053. The descriptive statistics as well as the factor loadings of the validation sample are shown in Table 4. As expected, easiness values of items that involve socially undesirable behavior were relatively low. This mainly concerns the items of Anger Expression-Out, while Anger Control as a socially more desirable behavior contains easier items.

The factor loadings are all sufficiently high with values ranging from λ = .51 to λ = .86. In line with the long version, highest mean scores were observed for Anger Control, followed by Anger ExpressionIn and -Out. The reliability coefficients of the Trait scales were satisfactory with values ranging between ω = .72 and ω = .77, and comparable to the long version with values .79. As with the State scales, the correlations of the Trait short scales were also comparable with those of the long version with regard to the sign and the size of the coefficients. The Trait short scales correlated very highly with the Trait scales of the long version of the STAXI-2 (r = .88 to r = .96), which indicates that the 15 items of the Trait short scales represent the original Trait scales very well and reflect the essential contents of these four subscales sufficiently well (see Table 5).

Invariance Tests of the Trait Scales A two-group comparison using the CFA with the validation and calibration samples, in which all factor loadings, intercepts, and error variances were defined as invariant over the two groups, showed a good model fit with

Table 5. Reliability (McDonalds Omega), descriptive statistics, and intercorrelations of the Trait Anger short scales (lower triangle matrix) and the Trait Anger scales of the long version (upper triangle matrix) and correlations between the two versions (main diagonal), based on the validation sample (N = 310) ω

(1) Trait angry reaction

0.72

6.71

2.03

(2) Anger expression-out

0.74

6.04

2.06

(3) Anger expression-in

0.77

8.66

(4) Anger control

0.73

11.46

Trait scale

(1)

(2)

(3)

(4)

0.96

0.49

0.31

0.23

0.38

0.88

0.06

0.46

2.61

0.32

0.08ns

0.93

0.22

2.59

0.18

0.35

0.13

0.93

Note. ω = McDonald’s Omega; ns = not significant. All correlations are significant at p < .01 (until noted otherwise).

Ó 2020 Hogrefe Publishing

European Journal of Health Psychology (2020), 27(2), 55–65

A. N. Tibubos et al., STAXI-2 Short Form

Table 6. Invariance tests of the trait variables over samples formed by chance, age, and gender Model

YB-w2

M1 Configural

225.59

168

.0020

0.033

0.971

M2 Metric

242.80

179

.0011

0.034

0.968

M3 Scalar

249.45

190

.0025

0.032

0.970

M4 Strict

269.15

205

.0018

0.032

0.968

p-value

RMSEA

CFI

ΔYB-w2

Δdf

p-value

M2 vs. M1

17.08

.106

M3 vs. M2

5.72

.891

M4 vs. M3

19.90

.176

Model comparison

Random sample (Calib.: N = 310 Valid.: N = 310)

Age ( 40 years: N = 173, > 40 years: N = 445) M1 Configural

243.21

168

.0001

0.038

0.961

M2 Metric

251.99

179

.0001

0.037

0.960

M2 vs. M1

10.01

.530

M3 Scalar

263.94

190

.0002

0.036

0.960

M3 vs. M2

11.29

.419

M4 Strict

270.81

205

.0010

0.033

0.965

M4 vs. M3

11.26

.734

Gender (women: N = 278, men: N = 342) M1 Configural

230.61

168

.0010

0.035

0.968

M2 Metric

241.51

179

.0013

0.033

0.968

M2 vs. M1

11.19

.428

M3 Scalar

254.46

190

.0013

0.033

0.967

M2 vs. M3

12.84

.304

M4 Strict

272.73

205

.0005

0.034

0.963

M4 vs. M3

18.87

.220

Note. YB-w = Yuan-Bentler-scaled w -test statistic; CFI = comparative fit index; RMSEA = root mean-square error of approximation; ΔYB-w = scaled w2difference test; Δdf = difference in degrees of freedom; Calib. = calibration sample; Valid. = validation sample. 2

YB-w2 = 269.15, df = 205, p = .002, RMSEA = .032, CFI = .968 and SRMR = .059. Thus, both samples are based on the same model with identical factor loadings, intercepts, and error variances, so that the two samples could be combined into a common sample (N = 620). As the results of the invariance tests show (see Table 6), strict invariance could be established both for age groups and gender groups: the measurement models of the subgroups do not differ in the factor loadings, intercepts, and error variances.

Intercorrelation of the State-Trait Scales of the STAXI-2 Short Form In accordance with the Trait-State model (Spielberger, 1988), the relationships between the State Anger scales of the STAXI-2 short form and the Trait Anger scales of the short form, and the original version were examined on the basis of the validation sample of the State scale. The correlations of the State scales of the short form with the scales assessing dispositional anger experience with both the short form and the original version were comparatively moderate or high. In particular, the correlations between the State Anger short scales with the original scale for Trait Angry Reaction and the Trait Anger scale of the short form, which is limited to the component of the dispositional anger reaction, were almost identically high (Table 7).

Discussion Within the scope of this study, a short form of the German-language STAXI-2 (Rohrmann et al., 2013) was European Journal of Health Psychology (2020), 27(2), 55–65

conceptualized. The State Anger scale was reduced from 15 to 9 items, the Trait Anger scale was reduced from 10 items to 3 items, and the habitual anger expression and anger control scales (Anger Expression-In, Anger Expression-Out, and Anger Control) from 26 to 12 items. This means that the number of items has been reduced by more than half. While the structure of State Anger is identical to the German long version with improved internal consistency in the short form, improvements have been made regarding the scales assessing traits. Critique on the STAXI-2 has been taken into account resulting in the elimination of the subscale Trait Temperament due to redundancy with Trait Anger Expression-Out and for economic reasons. Other than that, the structure has remained the same. Additionally, strict measurement invariance was established for both the State and the Trait scales across gender and age groups with the exception that for gender, only partially strict invariance was established for the State Anger subscale Verbal Angry Impulse. To the best of our knowledge, this study is the first investigating measurement invariance of the German version of STAXI or STAXI-2, respectively, across gender and age groups. Regarding the State Anger short scales, the CFA confirmed the three-factorial structure in the short form, which can also be found in the original version. As in the original version of the STAXI-2, the three subscales can be added up to a total value that corresponds to a global measure of anger (State Anger). In line with the long version, highest mean scores are observed for Angry Feelings, followed by Verbal and Physical Anger Impulse. The State short scales have a good internal consistency, with even higher coefficients compared to the original version. The very high correlations between the short scales and the corresponding Ó 2020 Hogrefe Publishing

A. N. Tibubos et al., STAXI-2 Short Form

Table 7. Intercorrelation of the State scales of the STAXI-2 short form with the Trait Anger scales of the short form and the original version based on the validation sample of the State scales (N = 216) Trait short form Trait angry reaction

Trait original version Trait anger

Trait angry temperament

Trait angry reaction

State short form State anger

.46

.51

.43

.49

Feeling angry

.40

.37

.25

.42

Verbal anger impulse

.44

.51

.44

.48

Physical anger impulse

.22

.29

.28

.24

Note. All correlations are significant (p < .01).

scales of the long version of the STAXI-2 (r .96) indicate that the short scales are able to grasp the essence of the State scales well. The comparison of the correlations of the State Anger short scales with the Trait Anger scales of the short form and the original version are also comparably moderate to high. Overall, the correlation patterns show initial evidence for the construct validity of the State Anger short scales. Also, first evidence for experimental validity of the State Anger scale was demonstrated by the increase of state anger in the anger-induction experiment. Despite the reduction in the number of items, the internal consistency (with coefficients between .72 and .77) of the four Trait scales still proved to be satisfactory. In line with the long version, highest mean scores are observed for Anger Control, followed by Anger Expression-In and Out. The reliability coefficients of the Trait scales are satisfactory with values comparable to the long version. To ensure the reliability of the Trait short scales, it would be necessary to additionally check the stability of the measurements over time for further analyses. The analysis of the intercorrelations of the Trait short scales showed the expected correlation pattern from the original version: Anger Expression-In and Anger Expression-Out scales also represent independent dimensions in the short version. Trait Anger correlates positively with both Anger Expression scales and negatively with Anger Control. However, a new feature of the STAXI-2 short form is that the very high intercorrelation (r .70) between the Trait Anger scale and the habitual scale Anger Expression-Out is no longer present. This high intercorrelation exists in both STAXI versions, STAXI (Schwenkmezger et al., 1992) and STAXI-2 (Rohrmann et al., 2013), which was criticized by Krohne and Hock (2007, p. 297). Especially items of the Trait scale Angry Temperament correlate very highly with items of the scale Anger Expression-Out. In the course of item reduction when developing the STAXI-2 short form, the items of the Angry Temperament scale were eliminated, which have so far led to the criticized high intercorrelation of both scales. Overall, the intercorrelations of the scales suggest that the

Ó 2020 Hogrefe Publishing

Trait Anger short scales reflect the scale structure of the STAXI-2 well (Rohrmann et al., 2013). With regard to the anger management scales, the theoretical division of Anger Control items into the two subscales Anger Control-In versus Anger Control-Out proposed in the US American STAXI-2 by Spielberger (1999) could – in accordance with the German adaptation of the STAXI-2 (Rohrmann et al., 2013, pp. 50 & 53) – not be found in the STAXI-2 short form. The theoretical subdivision of the Anger Control dimension made in the STAXI-2 by Spielberger (1999) was not checked via confirmatory analysis in any study of the US American manual. As already explained in Rohrmann et al. (2013), there appears to be no empirical evidence for a largely independent operationalization of these two anger management subscales of Anger Control. In order to maintain a parallelism between the US American and German STAXI-2 versions, however, Rohrmann et al. (2013) retained the subscale structure in the presentation of the test manual but pointed out (see, e.g., p. 57) that the subscales of Anger Control are rather explorative scales, so that the reported standard values should only be interpreted very carefully. Since the development of the short form of the STAXI-2 primarily aims at optimizing the procedure and a complete parallelism to the US American STAXI-2 – compared to the German adaptation by Rohrmann et al. (2013) – represents a subordinate priority, there is only one scale for assessing Anger Control in the STAXI-2 short form. Our study has some limitations that should be noted. First, further validation studies with external criteria, for instance, psychophysiological correlates of the State Anger scales short form and higher change sensitivity of these scales compared to the Trait Anger scales are needed. Second, sample sizes for invariance testing of state items were unbalanced for age groups and gender, and the male group was very small (N = 87). As Yoon and Lai (2018) pointed out, unequal sample sizes are more likely to reduce power to detect invariant items. Therefore, invariance testing should be repeated using larger and equal sample sizes. Finally, cross-cultural measurement invariance

European Journal of Health Psychology (2020), 27(2), 55–65

(see, e.g., Tibubos & Kröger, 2020) of the STAXI-2 short form needs to be tested in future studies since the questionnaire has been developed by using samples of the German population. To sum up, it was possible to develop an economic form of STAXI-2 by approximately half of the items without major restrictions in test quality in terms of reliability and validity. The STAXI-2 short form now consists of three State Anger scales and four Trait Anger scales. Essentially, there is solely one change between the short and the original version in the operationalization of anger. This concerns the elimination of the Angry Temperament dimension, which in previous versions showed a high correlation with the Anger Expression-Out scale. Due to the objective criterion of economics in the development of the STAXI-2 short form, this facet was eliminated after empirical and contextual considerations. The present study provides some evidence for the construct validity of the short form of the STAXI-2, while an investigation of the criterion validity of the STAXI-2 short form is therefore still pending. The strengths of the STAXI-2 short scales in comparison to the STAXI-2 reside particularly in the economy. Due to the reduced number of items, the short form of the STAXI-2 represents a great added value for use in practical questions, for example, in the context of clinical, educational and forensic psychology. The shorter State Anger scales can be better integrated into follow-up examinations. Similarly, the STAXI-2 short form scales are more likely to be used in large scale surveys in psychological, medical, pedagogical, and social science areas than the detailed original version. In particular, the short form is better usable as a component of extensive test batteries than the STAXI-2 (Rohrmann et al., 2013). The STAXI-2 short scales enable the economic recording of various facets of anger, anger tendency, and anger management within a few minutes.

Electronic Supplementary Materials The electronic supplementary material is available with the online version of the article at https://doi.org/ 10.1027/2512-8442/a000049 ESM 1. Descriptive statistics and EFA results for State and Trait Anger Items of the calibration samples ESM 2. Calculation of item easiness and Mplus syntax for EFA and invariance testing

References Averill, J. R. (1982). Anger and aggression. An essay on emotion. New York, NY: Springer. Brown, W. (1910). Some experimental results in the correlation of mental abilities. British Journal of Psychology, 3, 296–322.

European Journal of Health Psychology (2020), 27(2), 55–65

A. N. Tibubos et al., STAXI-2 Short Form

Byrne, B. M., Shavelson, R. J., & Muthén, B. (1989). Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychological Bulletin, 105, 456–466. https://doi.org/10.1037/0033-2909.105. 3.456 Cherek, D. R., Lane, S. D., Dougherty, D. M., Moeller, F. G., & White, S. (2000). Laboratory and questionnaire measures of aggression among female parolees with violent or nonviolent histories. Aggressive Behavior, 26, 291–307. Chesney, M. A. & Rosenman, R. H. (Eds.). (1985). Anger and hostility in cardiovascular and behavioral disorders. Washington, DC: Hemisphere. Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-offit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233–255. https://doi.org/10.1207/ S15328007SEM0902_5 Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297–334. Etzler, S. L., Rohrmann, S., & Brandt, H. (2014). Validation of the STAXI-2: A study with prison inmates. Psychological Test and Assessment Modeling, 56, 178–194. Fernandez E. (Ed.). (2013). Treatments for anger in specific populations. New York, NY: Oxford University Press. Gregorich, S. E. (2006). Do self-report instruments allow meaningful comparisons across diverse population groups? Testing measurement invariance using the confirmatory factor analysis framework. Medical Care, 44(Suppl. 3), 78–94. https://doi.org/ 10.1097/01.mlr.0000245454.12228.8f Guttman, L. (1945). A basis for analyzing test-retest reliability. Psychometrika, 10, 255–282. Hecht, M., Hardt, K., Driver, C. C., & Voelkle, M. C. (2019). Bayesian continuous-time Rasch models. Psychological Methods, 24, 516–537. https://doi.org/10.1037/met0000205 Hodapp, V. & Schwenkmezger, P. (Eds.). (1993). Ärger und Ärgerausdruck [Anger and Anger Expression]. Bern, Switzerland: Hans Huber. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. Izard, C. E. (1991). The psychology of emotions. New York, NY: Plenum Press. Kelava, A., & Moosbrugger, H. (2020). Deskriptiv-statistische Itemanalyse und Testwertbestimmung [Descriptive statistical item analysis and test score calculation]. In H. Moosbrugger & A. Kelava (Eds.), Testtheorie und Fragebogenkonstruktion [Test theory and scale construction] (3rd ed., pp. 75–102). Heidelberg, Germany: Springer. Kline, R. B. (2016). Principles and practice of structural equation modeling (4th ed.). New York, NY: The Guilford Press. Krohne, H. W., & Hock, M. (2007). Psychologische Diagnostik: Grundlagen und Anwendungsfelder [Psychological Diagnostics: Basics and fields of application]. Stuttgart, Germany: W. Kohlhammer. Lewis, M. & Rudolph, K. D. (Eds.). (2014). Handbook of developmental psychopathology (3rd ed.). New York, NY: Springer Science & Business Media. McDonald, R. P. (1999). Test theory: A unified treatment. Hillsdale, NJ: Erlbaum. Meredith, W. (1993). Measurement invariance, factor analysis and factorial invariance. Psychometrika, 58, 525–543. https://doi. org/10.1007/BF02294825 Moosbrugger, H., Gäde, J. C., Schermelleh-Engel, K., & Rauch, W. (2020). Klassische Testtheorie (KTT) [Classical test theory (CTT)]. In H. Moosbrugger & A. Kelava (Eds.), Testtheorie und Fragebogenkonstruktion [Test theory and scale construction] (3rd ed., pp. 103–117). Heidelberg, Germany: Springer.

Ó 2020 Hogrefe Publishing

A. N. Tibubos et al., STAXI-2 Short Form

Muthén, L. K., & Muthén, B. O. (1998–2015). Mplus user’s guide (7th ed.). Los Angeles, CA: Muthén & Muthén. Novaco, R. W. (1975). Anger control: The development and evaluation of an experimental treatment. Lexington, MA: Lexington Books. Novaco, R. W. (1978). Anger and coping with stress. In J. Foreyt & D. Rathjen (Eds.), Cognitive behavior therapy (pp. 135–173). New York, NY: Plenum Press. Potegal, M., Stemmler, G., & Spielberger, C. (Eds.). (2010). International handbook of anger: Constituent and concomitant biological, psychological, and social processes. New York, NY: Springer Science & Business Media. Putnick, D. L., & Bornstein, M. H. (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. https://doi.org/10.1016/J.DR.2016.06.004 Raykov, T., & Marcoulides, G. A. (2011). Psychometric theory. New York, NY: Routledge. Reise, S. P., Bonifay, W. E., & Haviland, M. G. (2013). Scoring and modeling psychological measures in the presence of multidimensionality. Journal of Personality Assessment, 95, 129–140. Rhemtulla, M., Brosseau-Liard, P. E., & Savalei, V. (2012). When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychological Methods, 17, 354–373. https://doi.org/10.1037/a0029315 Rohrmann, S., Hodapp, V., Schnell, K., Tibubos, A. N., Schwenkmezger, P., & Spielberger, C. D. (2013). Das State-TraitÄrgerausdrucks-Inventar-2 (STAXI-2). Deutschsprachige Adaptation des State-Trait Anger Expression Inventory-2 (STAXI-2) von Charles D. Spielberger [State-Trait-Anger-Expression-Inventory-2 (STAXI-2). German Adaptation of the State-Trait-AngerExpression-Inventory-2 (STAXI-2) of Charles D. Spielberger]. Bern, Switzerland: Hans Huber. Satorra, A., & Bentler, P. M. (2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66, 507–514. https://doi.org/10.1007/BF02296192 Schermelleh-Engel, K., & Gäde, J. C. (2020). Methoden der Reliabilitätsschätzung [Reliability estimation]. In H. Moosbrugger & A. Kelava (Eds.), Testtheorie und Fragebogenkonstruktion [Test theory and scale construction] (3rd ed., pp. 119–141). Heidelberg, Germany: Springer. Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2003). Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of Psychological Research Online, 8, 23–74. Retrieved from https://www. dgps.de/fachgruppen/methoden/mpr-online/issue20/ Schieman, S. (1999). Age and anger. Journal of Health and Social Behavior, 40, 273–289. Schmitt, M., & Altstötter-Gleich, C. (2010). Differentielle und Persönlichkeitspsychologie – kompakt [Differential and Personality Psychology – compact]. Weinheim, Germany: Beltz. Schwenkmezger, P., Hodapp, V., & Spielberger, C. D. (1992). Das (STAXI) [State-TraitState-Trait-Ärgerausdrucks-Inventar Anger-Expression-Inventory-2 (STAXI-2)]. Bern, Switzerland: Hans Huber. Spielberger, C. D., & Reheiser, E. C. (2010). The nature and measurement of anger. In M. Potegal, G. Stemmler, & C. D. Spielbeger (Eds.), International handbook of anger (pp. 403–412). New York, NY: Springer. Spielberger, C. D. (1966). The theory and research on anxiety. In C. D. Spielberger (Ed.), Anxiety and behaviour (pp. 3–20). New York, NY: Academic Press. Spielberger, C. D. (1972). Anxiety is an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–49). New York, NY: Academic Press.

Ó 2020 Hogrefe Publishing

Spielberger, C. D. (1979). Preliminary manual for the State-TraitPersonality Inventory (STPI) (Unpublished manuscript). Tampa, FL: University of South Florida. Spielberger, C. D. (1988). State-Trait Anger Expression Inventory (STAXI). Palo Alto, CA: Consulting Psychologist Press. Spielberger, C. D. (1999). State-Trait Anger Expression Inventory – 2 (STAXI-2) (Professional manual). Tampa, FL: Psychological Assessment Resources. Steenkamp, J.-B. E. M., & Baumgartner, H. (1998). Assessing measurement invariance in cross-national consumer research. Journal of Consumer Research, 25, 78–90. https://doi.org/ 10.1086/209528 Stone, A. A., Schwartz, J. E., Broderick, J. E., & Deaton, A. (2010). A snapshot of the age distribution of psychological well-being in the United States. Proceedings of the National Academy of Sciences, 107, 9985–9990. https://doi.org/10.1073/pnas. 1003744107 Tibubos, A. N., & Kröger, H. (2020). A Cross-cultural comparison of the ultrabrief mental health screeners PHQ-4 and SF-12 in Germany. Psychological Assessment. Advance online publication. https://doi.org/10.1037/pas0000814 Tibubos, A. N., Pott, E. K., Schnell, K., & Rohrmann, S. (2014). All anger is not created equal: A multi-method approach to understanding the link of anger responses with approach and avoidance motivation. Personality and Individual Differences, 70, 160–164. https://doi.org/10.1016/j.paid.2014.06.045 Tibubos, A. N., Schnell, K., & Rohrmann, S. (2013). Anger makes you feel stronger: The positive influence of trait anger in a reallife experiment. Polish Psychological Bulletin, 44, 147–156. https://doi.org/10.2478/ppb-2013-0017 Weber, H. (2002). Ärger. Psychologie einer alltäglichen Emotion [Anger. Psychology of an everyday emotion]. Weinheim, Germany: Juventa. Yoon, M., & Lai, M. H. C. (2018). Testing factorial invariance with unbalanced samples. Structural Equation Modeling, 25, 201– 213. https://doi.org/10.1080/10705511.2017.1387859 Yuan, K. H., & Bentler, P. M. (2000). Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. In M. E. Sobel & M. P. Becker (Eds.), Sociological methodology 2000 (pp. 165–200). Washington, DC: ASA. Zinbarg, R. E., Revelle, W., Yovel, I., & Li, W. (2005). Cronbach’s α, Revelle’s β, and McDonald’s ωH: Their relations with each other and two alternative conceptualizations of reliability. Psychometrika, 70, 1–11. History Received June 26, 2019 Revision received April 1, 2020 Accepted April 7, 2020 Published online July 1, 2020 ORCID Ana N. Tibubos https://orcid.org/0000-0003-4284-0019 Ana N. Tibubos Department of Psychosomatic Medicine and Psychotherapy University Medical Center Johannes Gutenberg University Mainz Untere Zahlbacher Str. 8 55131 Mainz Germany anananette.tibubos@unimedizin-mainz.de

European Journal of Health Psychology (2020), 27(2), 55–65

Understanding Medically Unexplained Symptoms and Bodily Distress Alexandra Martin / Omer Van den Bergh (Editors)

Medically Unexplained Symptoms and Bodily Distress (Series: Zeitschrift für Psychologie - Vol. 228/2) 2020, iv / 84 pp., large format US $49.00 / € 34.95 ISBN 978-0-88937-575-8 There is a high prevalence of adults and children presenting at health care services with bodily complaints that do not have an observable physiological cause, for example, pain, dizziness, fatigue, and bowel dysfunctions. A multitude of etiologic and pathogenic mechanisms and contributing factors play a role in medically unexplained symptoms (MUS) and bodily distress, representing a constant challenge to refine models of bodily distress to inform primary and secondary prevention, stepped care intervention, and specialized treatment.

This collection of contributions from around the world focuses on research that investigates four key areas: 1) detection and diagnosis, 2) etiologically relevant mechanisms, 3) clinical management and effective treatment, 4) implementation of treatment. Written from various research perspectives in psychology and medicine, this volume shows the necessity of interdisciplinary research to advance our understanding of MUS and bodily distress as a heterogeneous condition.

Contents and topics include • Medically Unexplained Symptoms and Bodily Distress: Four Challenges to Improve Understanding and Evidence-Based Care • Psychological Interventions for Health Anxiety and Somatic Symptoms • Functional Somatic Syndromes (FSS) in Children and Adolescents • Do Women With Severe Persistent Fatigue Present With Fatigue at the Primary Care Consultation?

www.hogrefe.com

• Somatic Symptom Perception and Interoception • C yberchondriasis: The Effect of Searching the Internet on Health Concerns • Stepped, Collaborative, Coordinated Care for Somatic Symptom and Related Disorders (Sofu-Net) • Efficacy of a Psychological Single-Session Intervention in Non-Cardiac • Health Anxiety: Conceptualization and Future Directions

Clear, up-to-date guidance for professionals working with obese children “This book should find a place on the bookshelf of all pediatric practitioners.” Marsha D. Marcus, PhD, Professor of Psychiatry and Psychology, University of Pittsburgh School of Medicine, PA

Denise E. Wilfley / John R. Best / Jodi Cahill Holland / Dorothy J. Van Buren†

Childhood Obesity (Series: Advances in Psychotherapy – Evidence-Based Practice - Volume 39) 2019, x + 80 pp. US $29.80 / € 24.95 ISBN 978-0-88937-406-5 Also available as eBook One in every six children, and more in some ethnic groups, are obese, which can lead to serious health problems in adulthood. Successful treatment of young patients is complex, requiring time-intensive, evidence-based care delivered by a multidisciplinary team. Help is at hand with this well written, compact book by leading experts, which gives health professionals – pediatricians, psychologists, other health workers – a clear overview of the current scientific knowledge on childhood obesity, from causality models and diagnosis to prevention and treatment.

www.hogrefe.com

In particular, the authors outline a family-based treatment method which is best supported by the evidence and meets the recommendations of the American Academy of Pediatrics and other organizations. The appendix provides the clinician with hands-on tools: a session plan, a pretreatment assessment form, self-monitoring forms, and a meal planning and physical activity worksheet. This book is essential reading for anyone who works with children and their families, equipping them to guide patients to appropriate and effective treatment.

Original Article

Affect Improvements and Measurement Concordance Between a Subjective and an Accelerometric Estimate of Physical Activity Björn Pannicke1, Julia Reichenberger1, Dana Schultchen2, Olga Pollatos2, and Jens Blechert1 1

Department of Psychology, Centre for Cognitive Neuroscience, Paris-Lodron-University of Salzburg, Austria

Department of Clinical & Health Psychology, Ulm University, Ulm, Germany

Abstract: Objectives: Physical activity (PA) positively influences several aspects of mental well-being including affect improvements. Yet, the fact that subjective and objective measures of PA often diverge challenges research on the relationship of PA and affect. Methods: Subjective (ecological momentary assessment, EMA) and objective (combined heart rate and accelerometric activity tracker) measures of PA alongside repeated ratings of positive and negative affects were obtained from 37 participants over 7 consecutive days. Results: Subjective and objective PA were significantly positively correlated. Affect improvements, that is, negative affect decrease as well as positive affect increase, were predicted by both subjective (EMA) and objective (activity tracker) data. Conclusions: Measurement concordance supports the validity of both assessment strategies. Affect improvements result from both subjective representations of one’s own activity as well as from physiological mechanisms of PA that one is not aware of, suggesting two independent routes to affect improvements. Keywords: physical activity, affect improvements, ecological momentary assessment, accelerometry, everyday life

Frequent physical activity (PA) is associated with broad physical health gains, for example, reduced risk for cardiovascular disease in primary prevention and slowed progression of coronary artery disease in secondary prevention (Warburton, Nicol, & Bredin, 2006). In addition to physiological health effects, PA also makes the onset of various mental disorders less likely and helps in reducing related psychological symptoms as a treatment component (Peluso & Andrade, 2005). Lifestyle modification treatments promoting PA show beneficial effects on depression and various stress related disorders (e.g., Cooney et al., 2013; Richardson et al., 2005; Rosenbaum, Tiedemann, Sherrington, Curtis, & Ward, 2014) as well as on non-clinical forms of depressed moods (Rebar et al., 2015). Importantly, besides planned and purposeful exercising, there are also various everyday activities of moderate intensity such as brisk walking or commuting by bicycle that can be beneficial for mental and physical health by contributing to the overall PA level (Audrey, Procter, & Cooper, 2014; Haskell et al., 2007; Martin, Kelly, Boyle, Corlett, & Reilly, European Journal of Health Psychology (2020), 27(2), 66–75 https://doi.org/10.1027/2512-8442/a000050

2016). The World Health Organization (2010) therefore recommends a minimum of 150 min of moderate PA or 75 min of vigorous PA per week, or respectively a combination of both, to achieve these benefits. Additionally, the benchmark of 10,000 steps per day was also suggested to ensure a certain degree of physical activity (Tudor-Locke & Bassett, 2004). Yet, research has repeatedly shown that large proportions of the population fail to reach these numbers due to multiple psychological and environmental barriers and constraints (e.g., Hallal, Andersen, Bull, Guthold, & Ekelund, 2012; Liu, Bennett, Harun, & Probst, 2008), although each episode of everyday PA might help reduce associated mental and physical health risks (Saris et al., 2003; World Health Organization, 2009; Zhai, Zhang, & Zhang, 2014). Thus, validly assessing and documenting everyday PA seems crucial as a starting point of effective prevention and intervention. Yet, methodological issues limit research on everyday PA: individuals are not always aware of their general fitness level (e.g., cardiorespiratory fitness level) as well as their Ó 2020 Hogrefe Publishing

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

daily extent of PA, especially regarding unplanned activities or shorter periods of PA. For instance, subjective moderate PA was found to be over-reported when compared to objective indicators in a college student sample (Downs, Van Hoomissen, Lafrenz, & Julka, 2014). Thus, the predictive validity of (mental) health effects from such subjective reports of PA might be limited due to the discordance between subjective and objective reports. It could therefore be problematic that research on health-related behaviors mostly uses self-reported PA (Sallis & Saelens, 2000). Such subjective reports are mainly operationalized through recall questionnaires or activity diaries, completed either on paper or electronically (Bert, Giacometti, Gualano, & Siliquini, 2013). However, there might be several limitations that bias such data such as effects of social desirability, memory limitations, or compliance with study instructions (Sallis & Saelens, 2000). Some of these disadvantages might be addressed through real-time assessment as afforded by smartphone apps (i.e., ecological momentary assessment, EMA): although subjective, they allow capturing activities right when they occur (Shiffman, Stone, & Hufford, 2008). Particularly memory bias might be reduced because of a relatively short time period between occurrence of behavior and assessment. While EMA reports are still subjective, they can be backed up with objective measures of PA (e.g., accelerometers, fitness trackers). These devices bypass any memory or reporting biases through continuously recording objective measures, for example, heart rate and accelerometry (Warren et al., 2010) to derive for instance metabolic equivalents of task (METs; Jetté, Sidney, & Blümchen, 1990). METs reflect the ratio of the energy expenditure rate relative to body weight for specific physical activities compared to resting (Jetté et al., 1990). In sum, subjective and objective markers of PA might have varying advantages and disadvantages and might differ from each other, especially regarding their reliability, but potentially also regarding their effects. Thus, the question arises as to whether data from simultaneous assessment of subjective (e.g., via EMA) and objective PA (e.g., via heart rate weighted accelerometry) agree and show measurement concordance. A systematic review comparing self-reported data of PA with objectively measured data found not only positive associations (resulting in a low-to-moderate mean correlation coefficient) but also zero and even negative associations (Prince et al., 2008). Thus, the present study will analyze the degree of withinperson measurement concordance between subjective, EMA-based estimates and objective, heart rate weighted accelerometric data of PA (Aim 1). Subjectively experienced and objectively measured PA might not only be discordant, but might also have different impacts on health outcomes. A particularly relevant health outcome concerns affect. According to psychophysiological Ó 2020 Hogrefe Publishing

theories, PA improves affect due to a reduction in the cortisol response from the hypothalamic-pituitary-adrenal axis (HPA axis) (Zschucke, Renneberg, Dimeo, Wüstenberg, & Ströhle, 2015) as well as to changes in specific brain regions and neurotransmitters (Bothe et al., 2013). Another mechanism for explaining PA mediated affect improvements posits that subjective awareness (a prerequisite for subjective reports; e.g., Ronda, Assema, & Brug, 2001) of one’s own successful engagement in PA might fuel positive experiences (e.g., of self-efficacy). To exemplify, research in youth has shown that only subjective (i.e., aware) reports in contrast to objective reports of PA were related to experienced self-efficacy (Kavanaugh, Moore, Hibbett, & Kaczynski, 2015). PA-related self-efficacy is in turn associated with affect (e.g., Bodin & Martinsen, 2004; Kwan & Bryan, 2010). Hence, the present study contrasted EMA-based and objective indices of PA in predicting subsequent affect changes (Aim 2). Particularly, we aimed to gain clarity on the nature of affect improvements: There is a plethora of evidence suggesting that PA increases subsequent positive affect. This effect has been found for subjectively reported PA (e.g., Kanning & Schlicht, 2010; Liao, Shonkoff, & Dunton, 2015; Schöndube, Kanning, & Fuchs, 2016) as well as objectively assessed PA (e.g., Bossmann, Kanning, Koudela, Hey, & Ebner-Priemer, 2013; Hogan, Mata, & Carstensen, 2013). However, the role of negative affect (reductions) is less clear (e.g., Liao et al., 2015; Puterman, Weiss, Beauchamp, Mogle, & Almeida, 2017), which is important as it might facilitate negative reinforcement.

The Present Study and Research Aims The present study used EMA measures to examine the two main study aims: smartphone supported subjective PA measurements were accompanied by objective PA measurements assessed by a heart rate/accelerometer monitoring device. Based on the literature, the main focus of the study was on moderate to vigorous activity (i.e., defined as subjectively being out of breath or sweating, respectively, objective METs > 3) to (a) ensure some degree of awareness and (b) include daily activities and not only planned exercise. Additionally, mental health effects are more frequently reported for these intensities (e.g., Asztalos, De Bourdeaudhuij, & Cardon, 2010). Therefore, we expected within-participant concordance between subjective and objective PA measures (Aim 1). Regarding affect improvements, we expected that both subjective PA (assessed via EMA) and objective PA (assessed via activity tracking measures) influence subsequent changes in positive and negative affects (Aim 2). Based on the literature reviewed above, we hypothesized that higher subjective and objective PA would result in higher positive and lower negative European Journal of Health Psychology (2020), 27(2), 66–75

affects. However, due to a lack of previous research, we had no specific hypothesis as to whether subjective or objective PA would be more important for affect improvements. In addition, to acknowledge individual differences potentially influencing the associations described above we controlled for body-mass index (BMI) as well as gender based on differences found in previous literature (Slootmaker, Schuit, Chinapaw, Seidell, & Mechelen, 2009; Thome & Espelage, 2004). We decided to investigate the described research aims in a sample of university students, given that PA is related to health variables and mood improvements in (university) student populations (Annesi, Porter, Hill, & Goldfine, 2017; Joseph, Royse, Benitez, & Pekmezi, 2014; Wunsch, Kasten, & Fuchs, 2017). Moreover, students are prone to regularly experience stress as well as negative affect and are even increasingly at risk for mental health problems (Storrie, Ahern, & Tuckett, 2010).

Method Participants Participants were recruited by means of a study announcement via e-mail, flyers, and by word of mouth. Initially, data of 51 participants were collected. Since not all participants were able to wear the activity tracking device (e.g., because of a sticking plaster allergy; constraints with regard to the number of devices), objective PA data could not be attained for all recruited participants (N = 12). Additionally, participants were excluded in case of substantial missing data in both EMA (< 50% of completed EMA questionnaires) and activity tracker assessment methods (N = 2). Thus, in total 37 participants (28 female) were included in the statistical analyses. Participants, mostly university students, had a mean age of 23.5 years (SD = 2.60 years, age range: 19– 28 years). All participants received written and oral information on the purpose of the study and signed an informed consent according to the relevant ethics committee that also granted the ethical approval for the present study.

Procedure Given the inclusion criteria, we only recruited participants ranging from 18 to 35 years, who owned a smartphone to complete EMA data and were willing and able to carry the Actihearts device (i.e., without sticking plaster allergy). Initially, the participants’ demographic variables and their body weight and height (for BMI) were measured in the laboratory. Afterward, participants were asked to answer the “Freiburger Fragebogen zur körperlichen Aktivität” (FFKA; Freiburg Questionnaire of Physical Activity; Frey & Berg, 2002; Frey, Berg, Grathwohl, & Keul, 1999). It European Journal of Health Psychology (2020), 27(2), 66–75

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

refers to physical activities at work and during leisure-time and allows classification of the recent level of physical activity for each participant in terms of body weight related energy metabolism per week. The FFKA shows acceptable psychometric properties (Frey et al., 1999). Then, participants received the monitoring device for objective PA with instructions on how to wear it (e.g., participants were told to wear the device all day and all night for 7 consecutive days unless they engaged in long-lasting water activities such as swimming) and how to install and use the smartphone app. The smartphone app enabled subjective data collection multiple times a day throughout the following 8 days. However, only the respective days of smartphone assessment with concurrent objective monitoring data were used for analyses. The first day functioned as a practice day for participants to become familiar with the surveys via app and the wearing of the device (data not used in the present study). The majority of participants wore the device for 8 days, two individuals only for 5 days, and four individuals for 7 days (practice day not taken into account). At the end of the study period, individuals answered final questionnaires which included compliance and reactivity items asking for altered behavior based on study participation. Reactivity concerning PA was assessed by asking “How much did the assessment itself influence your physical activity during the duration of the study?” (0 = not at all to 100 = very much). Reactivity concerning affects was assessed by asking “Did the assessment itself influence your mental well-being during the duration of the study?” (0–100, with 0 = negatively, 50 = not at all, 100 = positively). Moreover, compliance items (i.e., estimates of answered questionnaires as well as acceptability and feasibility of data entries) asked to what extent participants followed the study instructions. Note that other parts of the dataset have previously been analyzed and published in Schultchen et al. (2019) as well as Reichenberger et al. (2018).

Measures Smartphone App Participants received six app notifications (beeps) per day every 2.5 hr starting at 9:00 am and ending at 9:30 pm. With each of these beeps individuals were asked among other things (e.g., eating behavior, food craving, stress; not of relevance for the present study) to estimate their extent of moderate to vigorous PA since the last answered beep with the question “How many minutes have you been physically active since the last entry? Under physical activity we define activities that got you sweating or out of breath.” Answers were given on a continuous rating slider from 0 to 160 min. Participants were given 10 additional minutes for their subjective data entry because of a potential short delay in answering the questionnaire after the first Ó 2020 Hogrefe Publishing

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

initial beep. Moreover, each beep contained adjectives concerning current positive and negative affective states rated on a continuous rating slider from 0 (= not at all) to 100 (= very much). Some adjectives were chosen in the style of the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988), more specifically the German Version of the PANAS (Breyer & Bluemke, 2016). Additionally, we included items in the same style referring to emotions with a lower threshold (e.g., worried instead of afraid), items which might optimally cover a high and low arousal space (e.g., relaxed as a low-arousal positive emotion), or items that are relevant to the specific study content (e.g., bored). Positive affect adjectives used were “cheerful,” “enthusiastic,” “relaxed,” “calm,” “active,” and “awake,” whereas negative ones were “irritated,” “worried,” “troubled,” “bored,” “nervous/stressed,” and “dissatisfied with self.” The within-person reliability of affect items was calculated using the variance decomposition technique (Shrout & Lane, 2012) and was moderate for positive affect (Rc = .78) as well as for negative affect (Rc = .69) according to Shrout (1998). For safety reasons (e.g., while driving) and practicability, it was permitted to answer the app prompts with a maximum delay of 1 hr after the onset of each beep interval. Later entries were registered as missing. Objective PA Measures To measure objective parameters of PA, a combined heart rate and movement sensor (Actiheart; Cambridge Neurotechnology Ltd, 2010) was used. The first of two electrodes is located on the lower end of the sternum while the second one is attached laterally on the left side of the chest. The device determines heart rate (Brage, Brage, Franks, Ekelund, & Wareham, 2005) and activity impulses (using accelerometry) every 15 s and establishes METs classified in sedentary behavior (< 1.5 METs), light PA (1.5–3 METs), moderate PA (3–6 METs), and vigorous PA (> 6 METs). Based on these data and the Actiheart’s branched model of analysis, minutes of specific MET levels in the interval between two smartphone beeps were calculated for each participant. As described above, only intervals for which both subjective and objective data were available were included in the statistical analyses, that is, a maximum of 7 days due to the Actiheart’s battery life. In general, the Actiheart device is a technically reliable and valid instrument (Brage et al., 2005) for the measurement of heart rate and PA under natural conditions (Barreira, Kang, Caputo, Farley, & Renfrow, 2009). It has previously been used for PA research (Georgiou et al., 2015) and therefore served as the source of objective data in the present study.

Data Reduction and Statistical Analyses Due to the study’s focus on all, everyday (incidental) moderate to vigorous PA instead of exclusively intentional and Ó 2020 Hogrefe Publishing

planned exercising, analyses targeted minutes of moderate and vigorous PA (MVPA; > 3 METs) during a given 2.5-hour interval between smartphone beeps. Because of repeated measures for each participant, hierarchical multilevel modeling was used for analyses by means of the software HLM7 (Raudenbusch, Bryk, & Congdon, 2011). Repeated assessments of an individual on Level 1 were nested within participants on Level 2. Regarding the first research aim of subjective-objective measurement concordance, the subjective estimation of PA within the last 2.5 hr between t0 and t1 (reported at t1) was predicted on Level 1 by objectively measured data of PA, which were aggregated in the respective time interval. Secondary analyses were conducted including gender and BMI in prediction of the intercept on Level 2 to control for their influences. Moreover, similar to the measurement concordance analysis, an exploratory control analysis used sedentary behavior (< 1.5 METs) as the predictor of subjective PA. Continuous predictor variables were person-mean centered on Level 1 and grand-mean centered on Level 2, whereas the dichotomous variable, that is, gender (0 = female, 1 = male), was entered uncentered into the models. For the second research aim of affect improvements participants’ momentary mean positive (based on the 6 positive affect items above) and, respectively mean negative (based on the 6 negative affect items above) affective states were calculated for each beep at t1. Both mean affects were then predicted separately by subjective and objective measures of PA (person-mean centered) during the preceding 2.5 hr interval (between t0 and t1). To avoid obtaining general affect maintenance effects, the influence of the preceding affect (at t0) was simultaneously controlled for by using lagged affect variables (person-mean centered and autocorrelated). Additionally, exploratory control analyses used sedentary behavior as the predictor of changes in affect. The slopes and intercepts in all models were allowed to vary randomly.

Results Descriptives Table 1 reports participants’ descriptive statistics based on 1,155 records of both EMA- (subjective) and Actiheart(objective) based data. The overall compliance (percentage of completed EMA questionnaires) was 86%. Based on the FFKA questionnaire overall activity score, 27 participants were classified as having high physical activity levels, 7 medium PA levels, and 3 low PA levels during the weeks before the smartphone assessment. Participants reported that the influence of study participation on their PA was rather low, M = 10.8 (SD = 13.94, range = 0–46). Moreover, European Journal of Health Psychology (2020), 27(2), 66–75

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

Table 1. Descriptive statistics of the assessed variables Variable

Level 1 EMA: Subjective PA (0–160 min) Sedentary behavior (< 1.5 METs; min) Light PA (1.5–3 METs; min)

7.62

18.02

121.15

27.74

21.37

19.48

Moderate PA (3–6 METs; min)

6.76

11.84

Vigorous PA (> 6 METs; min)

0.63

3.93

Objective MVPA (> 3 METs; min)

7.39

13.21

Daily MVPA (> 3 METs; min)

45.05

42.29

Negative affect (0–100)

14.31

13.03

Positive affect (0–100)

41.06

17.81

22.28

2.72

PA (β10 = .205, SE = .045, p < .001) with an overall pseudo-R2 of .231. The longer the time sedentary, the lower participants’ subjective PA was. Additionally, to test whether participants improved in concordance between subjective PA and objective MVPA across study participation, the variable “day” (i.e., days in the study) was taken into account. No significant interaction was found meaning that consecutive days (i.e., the longer participants were part of the study) did not influence the association of subjective and objective PA (β30 = .88, SE = 2.04, p = .669). Controlling for gender and BMI did not influence the pattern or significance of all reported results.

Level 2 BMI (kg/m2)

Note. N = 37. Data based on 1,155 observations, except for “Daily MVPA” (222 observations) and BMI (one-time measurement). PA = Physical activity; MVPA = Moderate to vigorous physical activity; MET = Metabolic equivalent; BMI = body mass index. Range of daily MVPA: 0–299.00 min. All PA values (including sedentary behavior) refer to intervals of 2.5 hr, except the daily MVPA.

participants also reported that the study participation neither negatively nor positively changed their mental well-being, M = 49.2 (SD = 6.90, range = 38–73).

Affect Improvements In order to analyze potential associations between PA estimates and positive as well as negative affect within a day, the latter were predicted by subjective and objective PA separately. To account for autocorrelations or – stated differently – to focus on affect changes in a given interval, respectively preceding positive or negative affect was included as lagged control predictor. The models are expressed by the following equation (exemplified for subjective PA as predictor of positive affect): Level 1 (occasions)

Subjective-Objective Concordance A first set of analyses tested the relationship of objective moderate to vigorous PA (MVPA) and subjective PA. The models are expressed by the following equation: Level 1 (occasions)

Current positive affectij ¼ β0j þβ1j Subjective PAij þ β2j Preceding positive affectij þ r ij

Level 2 (participants)

Subjective PAtj ¼ β0j þ β1j Objective MVPAtj þ r tj

β0j ¼ γ00 þ u0j β1j ¼ γ10 þ u1j β2j ¼ γ20 þ u2j

Level 2 (participants)

ð2Þ

β0j ¼ γ00 þ u0j β1j ¼ γ10 þ u1j

ð1Þ

A significant positive association between subjective PA and objective data of MVPA was found (β10 = .497, SE = .097, p < .001) with an overall pseudo-R2 of .219. The higher the objective MVPA, the higher participants’ subjective PA was. Furthermore, there was no significant difference between the mean subjective PA and the mean objective MVPA (min/2.5 hr, objective: M = 7.45 min, SE = 4.14; subjective: M = 7.53, SE = 4.73) as revealed by a paired sample t-test, t(36) = .076, p = .936. Moreover, exploratory analyses were run with objective sedentary behavior as a form of (low) objective PA. Results revealed a significant negative association with subjective European Journal of Health Psychology (2020), 27(2), 66–75

Positive Affect A significant positive association between subjective PA estimates and positive affect was found (β10 = .193, SE = .041, p < .001) with an overall pseudo-R2 of .059. The higher participants estimated the amount of their own (preceding) PA, the higher their current positive affect was. A significant positive association was also found regarding objective MVPA and positive affect (β10 = .141, SE = .036, p < .001) with an overall pseudo-R2 of .027. When both assessment types were combined in one simultaneous model, subjective PA (β10 = .159, SE = .044, p < .001) as well as objective MVPA (β10 = .088, SE = .036, p < .001) remained significant with an overall pseudo-R2 of .070. In contrast, a significant negative association between Ó 2020 Hogrefe Publishing

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

objective sedentary behavior and positive affect was found (β10 = .077, SE = .021, p < .001) with an overall pseudo-R2 of .032. The longer the duration of sedentary behavior, the lower participants’ subsequent positive affect was. Again, controlling for gender and BMI did not influence the pattern or significance of all reported results. Negative Affect Participants showed lower negative affect subsequent to higher subjective PA estimates: analyses revealed a significant negative association between negative affect and subjective PA (β10 = .107, SE = .017, p < .001) with an overall pseudo-R2 of .039. Again, this result was also found for objective MVPA as the predictor of negative affect (β10 = .095, SE = .024, p < .001) with an overall pseudo-R2 of .020. When both assessment types were combined in one simultaneous model, subjective PA (β10 = .092, SE = .016, p < .001) as well as objective MVPA (β10 = .054, SE = .020, p < .001) remained significant with an overall pseudo-R2 of .043. On the other hand, higher negative affect was found after longer episodes of sedentary behavior, as objective sedentary behavior was significantly positively associated with subsequent negative affect (β10 = .051, SE = .011, p < .001) with an overall pseudo-R2 of .012. Again, controlling for gender and BMI did not influence the pattern or significance of all reported results.

Discussion In the context of assessing affect improvements as a result of PA, the present study used a naturalistic EMA design to examine two research aims. The first question examined the concordance of subjective EMA data and objective, heart rate weighted accelerometric measures of PA within a given individual across 7 days. The second question investigated both data sources of PA as predictors of affect improvements (increase in positive affect and decrease in negative affect examined separately).

Research Aim 1: Subjective-Objective (Measurement) Concordance Regarding the first research aim, we obtained measurement concordance as hypothesized: Results revealed a significant association between subjective and objective MVPA, as well as largely accurate MVPA estimations (nonsignificant t-test). This result contrasts with previous findings of overestimation of moderate (to vigorous) PA in a comparable student sample (Downs et al., 2014). In addition, the present results differ from results of over- or underestimation that used different methodological approaches and Ó 2020 Hogrefe Publishing

samples (e.g., in an adult sample: Dyrstad, Hansen, Holme, & Anderssen, 2014; Prince et al., 2008). However, our reported results correspond to findings of accurate (or even under-estimated) estimates by physically fit study participants in terms of overall concordance (Shook et al., 2016) as well as to time-specific within-subject concordance (Bexelius, Sandin, Trolle Lagerros, Litton, & Löf, 2011). One explanation of the accurate estimations in the present study might be that moderate to vigorous PA likely “breaks into awareness” by reaching an intensity that is subjectively noted and remembered, whereas “lighter” episodes of PA might go unnoticed. This interpretation is underpinned by a review observing that self-reports are more sensitive to high intensity objective PA compared to low-to-moderate ones (Prince et al., 2008). Thus, the observed measurement concordance could be the result of PA-awareness that enabled participants’ precise and systematic self-reports of PA episode durations (for a detailed view on PA-related awareness, see Ronda et al., 2001; van Sluijs, Griffin, & van Poppel, 2007). In the present study, neither BMI nor gender changed the pattern or significance of the results (in line with Dyrstad et al., 2014; Poole et al., 2011) suggesting a rather broad applicability within the present population.

Research Aim 2: Affect Improvements Regarding the second research aim, subjective estimates of PA emerged as significant predictors of improvements in both positive and negative affects (even in consideration of the controlled influence of the preceding affect and the respective objective PA). The same results were found regarding objective MVPA. Thus, both objective and subjective PA indices appear to be unique, non-redundant, and reliable predictors of both types of affect improvement (decrease in negative affect and increase in positive affect). Similarly, subjective and objective PA assessments have shown specific associations with health relevant outcomes such as different cardiovascular measures (Schmidt, Cleland, Thomson, Dwyer, & Venn, 2008). Complementing these findings, objective sedentary behavior was significant as a positive predictor of affect deterioration in the present study. Our reported effects for objective MVPA agree with previous findings regarding positive affect increases (Cox, Thomas, & Davis, 2001; Hogan et al., 2013) as well as negative affect decreases (Aggio et al., 2017). However, also the subjective perception of PA was associated with (positive and negative) affect improvements: the predictive potential of subjective PA in the present study corresponds to previous findings regarding positive affect increases (e.g., Kanning & Schlicht, 2010; Mata et al., 2012; Wichers et al., 2012). In contrast to other studies (Mata et al., 2012; Schwerdtfeger, Eberhardt, & European Journal of Health Psychology (2020), 27(2), 66–75

Chmitorz, 2008; Wichers et al., 2012), the present study was also able to find negative affect decreases after PA. The results suggest that introspection and self-monitoring/ self-awareness regarding PA translate into affect improvements independently of objective PA. At the same time, the results support the affect changing potential of actual (objective) PA, as the effects of PA on physical and mental health are undisputed. Together, these two associations suggest different mediating routes from PA to affect. Psychophysiological research on the one hand has shown that objective movement, for instance aerobic exercise, inhibits the feedback mechanism of the HPA axis and thereby reduces the cortisol response in stressful situations (i.e., stress buffering) while also increasing positive affect (Zschucke et al., 2015). Moreover, brain regions (e.g., ventral striatum) and neurotransmitters (e.g., dopamine), that are associated with reward systems, are influenced by PA (Bothe et al., 2013). Thus, psychophysiological or “bottom-up” mechanisms might likely be at play. In addition to physiological mechanisms, the subjective perception of one’s own PA seems to positively influence affect and thereby might contribute to the maintenance of health relevant behavior. As already described above, psychological aspects such as awareness and selfefficacy (or related experiences such as pride and goal attainment) might play important roles in this context (Bodin & Martinsen, 2004; Kavanaugh et al., 2015; McAuley et al., 2007; Pickett, Yardley, & Kendrick, 2012; van Sluijs et al., 2007). The fact that affect improvements can be attained through both positive affect increases and negative affect decreases suggests that (positive and negative) reinforcement should work for both negative and “neutral” baseline affects. Regarding both types of reinforcement, previous research has found that PA can lead to an increase in positive affect as well as to a decrease in negative affect in individuals with rather “neutral” baseline affects (i.e., healthy individuals; Annesi et al., 2017; Joseph et al., 2014; Wunsch et al., 2017) as well as individuals with rather negative baseline affects, that is, (subclinically) depressed individuals (Mata, Hogan, Joormann, Waugh, & Gotlib, 2013; Pickett et al., 2012).

Limitations, Future Directions, and Conclusions The following limitations must be acknowledged: Firstly, our sample consisted mostly of female university students. Therefore, results cannot be generalized, which appears plausible given that different results were found in different samples (e.g., Dyrstad et al., 2014). Hence, future studies may examine participants with broader sample

European Journal of Health Psychology (2020), 27(2), 66–75

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

characteristics, especially from vulnerable populations such as older adults (e.g., Shiroma et al., 2015) or individuals with chronic diseases (e.g., O’Neill et al., 2017; Thyregod & Bodtger, 2016). Additionally, the sample size of the present study was also relatively small and might be increased in future studies to improve statistical power on Level 2. In order to increase power on Level 1, future studies might also extend the assessment period (more days) and/or increase the frequency of assessment (more beeps per day). Secondly, our methodological approach was to assess subjective MVPA across the last 2.5 hr in order to balance study interest and participant burden. However, future research might profit from a more fine-grained assessment of PA (e.g., with regard to kind of activity, individual intensity according to fitness level, etc.) and variability in recall period (e.g., reporting PA directly when it occurs). This approach might also aid in mapping subjective PA more closely to the objective PA assessment. Moreover, future studies could additionally make use of objective physiological markers to examine PA and affect related biological mechanisms, for instance by means of an ambulatory sampling of cortisol. Thirdly, because of the monitoring device (and related battery life) the assessment period was shorter than the possible smartphone assessment period. Future research might utilize other kinds of PA tracking like an inbuilt accelerometer in modern smartphones, a wrist-worn fitness tracker or combinations of smartphone apps and fitness trackers to gather objective PA. To conclude, the present study shed light on the association between subjective EMA estimates and objective, heart rate weighted accelerometric PA measures. In this context, smartphone-based subjective assessment seems a promising and valid method (e.g., Knell et al., 2017) because of its high practicability and ubiquity. Importantly, a psychological advantage might be that the above-mentioned awareness of one’s individual extent of PA may be increased in the course of self-reporting (van Sluijs, van Poppel, Twisk, & Mechelen, 2006) and could thereby initiate a self-reflective process. Such competence appears to be of great importance for health-related awareness and increased engagement in PA (Lechner, Bolman, & van Dijke, 2006; Ronda et al., 2001; van Sluijs et al., 2007). Future studies could make use of individual (bio-)feedback regarding both subjective and objective assessment methods in order to examine an enhancement of participants’ PA-awareness as a potential form of health beneficial intervention. As the present study underlines the relevance of subjective perception and objective measures regarding health behavior, both sources of data are relevant parts of any comprehensive monitoring or intervention approach targeting PA and its affect improving benefits.

Ó 2020 Hogrefe Publishing

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

References Aggio, D., Wallace, K., Boreham, N., Shankar, A., Steptoe, A., & Hamer, M. (2017). Objectively measured daily physical activity and postural changes as related to positive and negative affect using ambulatory monitoring assessments. Psychosomatic Medicine, 79, 792–797. https://doi.org/10.1097/PSY.0000000000000485 Annesi, J. J., Porter, K. J., Hill, G. M., & Goldfine, B. D. (2017). Effects of instructional physical activity courses on overall physical activity and mood in university students. Research Quarterly for Exercise and Sport, 88, 358–364. https://doi.org/ 10.1080/02701367.2017.1336280 Asztalos, M., De Bourdeaudhuij, I., & Cardon, G. (2010). The relationship between physical activity and mental health varies across activity intensity levels and dimensions of mental health among women and men. Public Health Nutrition, 13, 1207– 1214. https://doi.org/10.1017/S1368980009992825 Audrey, S., Procter, S., & Cooper, A. R. (2014). The contribution of walking to work to adult physical activity levels: A cross sectional study. International Journal of Behavioral Nutrition and Physical Activity, 11, 37. https://doi.org/10.1186/14795868-11-37 Barreira, T., Kang, M., Caputo, J., Farley, R., & Renfrow, M. (2009). Validation of the Actiheart monitor for the measurement of physical activity. International Journal of Exercise Science, 2, 60–71. Bert, F., Giacometti, M., Gualano, M. R., & Siliquini, R. (2013). Smartphones and health promotion: A review of the evidence. Journal of Medical Systems, 38, 9995. https://doi.org/10.1007/ s10916-013-9995-7 Bexelius, C., Sandin, S., Trolle Lagerros, Y., Litton, J.-E., & Löf, M. (2011). Estimation of physical activity levels using cell phone questionnaires: A comparison with accelerometry for evaluation of between-subject and within-subject variations. Journal of Medical Internet Research, 13, e70. https://doi.org/10.2196/ jmir.1686 Bodin, T., & Martinsen, E. W. (2004). Mood and self-efficacy during acute exercise in clinical depression. A randomized, controlled study. Journal of Sport and Exercise Psychology, 26, 623–633. https://doi.org/10.1123/jsep.26.4.623 Bossmann, T., Kanning, M., Koudela, S., Hey, S., & Ebner-Priemer, U. (2013). The association between short periods of everyday life activities and affective states: A replication study using ambulatory assessment. Frontiers in Psychology, 4, 102. https://doi.org/10.3389/fpsyg.2013.00102 Bothe, N., Zschucke, E., Dimeo, F., Heinz, A., Wustenberg, T., & Strohle, A. (2013). Acute exercise influences reward processing in highly trained and untrained men. Medicine and Science in Sports and Exercise, 45, 583–591. https://doi.org/10.1249/ MSS.0b013e318275306f Brage, S., Brage, N., Franks, P. W., Ekelund, U., & Wareham, N. J. (2005). Reliability and validity of the combined heart rate and movement sensor Actiheart. European Journal of Clinical Nutrition, 59, 561–570. https://doi.org/10.1038/sj.ejcn.1602118 Breyer, B., & Bluemke, M. (2016). Deutsche Version der Positive and Negative Affect Schedule PANAS (GESIS Panel) [German version of the Positive and Negative Affect Schedule PANAS (GESIS Panel)]. Mannheim, Germany: GESIS – Leibniz-Institut für Sozialwissenschaften. Cambridge Neurotechnology Ltd. (2010). The Actiheart user manual (v. 4.0.35). Cambridgeshire, UK: Cambridge Neurotechnology. Cooney, G., Dwan, K., Greig, C., Lawlor, D., Rimer, J., Waugh, F., . . . Mead, G. (2013). Exercise for depression. The Cochrane Database of Systematic Reviews, 9, CD004366. https://doi. org/10.1002/14651858.CD004366.pub6

Ó 2020 Hogrefe Publishing

Cox, R. H., Thomas, T. R., & Davis, J. E. (2001). Postive and negative affect associated with an acute bout of aerobic exercise. Journal of Exercise Physiology Online, 4, 13–20. Downs, A., Van Hoomissen, J., Lafrenz, A., & Julka, D. L. (2014). Accelerometer-measured versus self-reported physical activity in college students: Implications for research and practice. Journal of American College Health, 62, 204–212. https://doi. org/10.1080/07448481.2013.877018 Dyrstad, S., Hansen, B., Holme, I., & Anderssen, S. (2014). Comparison of self-reported versus accelerometer-measured physical activity. Medicine and Science in Sports and Exercise, 46, 99–106. https://doi.org/10.1249/MSS.0b013e3182a0595f Frey, I., & Berg, A. (2002). Physical activity counseling: Assessment of physical activity by questionnaire. European Journal of Sport Science, 2, 1–6. https://doi.org/10.1080/17461390200072406 Frey, I., Berg, A., Grathwohl, D., & Keul, J. (1999). Freiburger Fragebogen zur körperlichen Aktivität – Entwicklung, Prüfung und Anwendung [Freiburg Questionnaire of Physical Activity – development, evaluation and application]. Sozial- und Präventivmedizin, 44, 55–64. https://doi.org/10.1007/bf01667127 Georgiou, E., Matthias, E., Kobel, S., Kettner, S., Dreyhaupt, J., Steinacker, J., & Pollatos, O. (2015). Interaction of physical activity and interoception in children. Frontiers in Psychology, 6, 502. https://doi.org/10.3389/fpsyg.2015.00502 Hallal, P., Andersen, L., Bull, F., Guthold, R., & Ekelund, U. (2012). Global physical activity levels: Surveillance progress, pitfalls, and prospects. The Lancet, 380, 247–257. https://doi.org/ 10.1016/S0140-6736(12)60646-1 Haskell, W. L., Lee, I.-M., Pate, R. R., Powell, K. E., Blair, S. N., Franklin, B. A., . . . Bauman, A. (2007). Physical activity and public health: Updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Medicine and Science in Sports and Exercise, 39, 1423–1434. https://doi.org/10.1249/mss.0b013e3180616b27 Hogan, C., Mata, J., & Carstensen, L. (2013). Exercise holds immediate benefits for affect and cognition in younger and older adults. Psychology and Aging, 28, 587–594. https://doi. org/10.1037/a0032634 Jetté, M., Sidney, K., & Blümchen, G. (1990). Metabolic Equivalents (METS) in exercise testing, exercise prescription, and evaluation of functional capacity. Clinical Cardiology, 13, 555– 565. https://doi.org/10.1002/clc.4960130809 Joseph, R. P., Royse, K. E., Benitez, T. J., & Pekmezi, D. W. (2014). Physical activity and quality of life among university students: Exploring self-efficacy, self-esteem, and affect as potential mediators. Quality of Life Research, 23, 659–667. https://doi. org/10.1007/s11136-013-0492-8 Kanning, M., & Schlicht, W. (2010). Be active and become happy: An ecological momentary assessment of physical activity and mood. Journal of Sport and Exercise Psychology, 32, 253–261. https://doi.org/10.1123/jsep.32.2.253 Kavanaugh, K., Moore, J. B., Hibbett, L. J., & Kaczynski, A. T. (2015). Correlates of subjectively and objectively measured physical activity in young adolescents. Journal of Sport and Health Science, 4, 222–227. https://doi.org/10.1016/j.jshs.2014. 03.015 Knell, G., Gabriel, K., Businelle, M., Shuval, K., Wetter, D., & Kendzor, D. (2017). Ecological momentary assessment of physical activity: Validation study. Journal of Medical Internet Research, 19, e253. https://doi.org/10.2196/jmir.7602 Kwan, B., & Bryan, A. (2010). Affective response to exercise as a component of exercise motivation: Attitudes, norms, selfefficacy, and temporal stability of intentions. Psychology of Sport and Exercise, 11, 71–79. https://doi.org/10.1016/j. psychsport.2009.05.010

European Journal of Health Psychology (2020), 27(2), 66–75

Lechner, L., Bolman, C., & van Dijke, M. (2006). Factors related to misperception of physical activity in the Netherlands and implications for health promotion programmes. Health Promotion International, 21, 104–112. https://doi.org/10.1093/ heapro/dal011 Liao, Y., Shonkoff, E., & Dunton, G. (2015). The acute relationships between affect, physical feeling states, and physical activity in daily life: A review of current evidence. Frontiers in Psychology, 6, 1975. https://doi.org/10.3389/fpsyg.2015.01975 Liu, J., Bennett, K., Harun, N., & Probst, J. (2008). Urban-rural differences in overweight status and physical inactivity among US children aged 10–17 years. The Journal of Rural Health, 24, 407–415. https://doi.org/10.1111/j.1748-0361.2008.00188.x Martin, A., Kelly, P., Boyle, J., Corlett, F., & Reilly, J. J. (2016). Contribution of walking to school to individual and population moderate-vigorous intensity physical activity: Systematic review and meta-analysis. Pediatric Exercise Science, 28, 353–363. https://doi.org/10.1123/pes.2015-0207 Mata, J., Hogan, C. L., Joormann, J., Waugh, C. E., & Gotlib, I. H. (2013). Acute exercise attenuates negative affect following repeated sad mood inductions in persons who have recovered from depression. Journal of Abnormal Psychology, 122, 45–50. Mata, J., Thompson, R. J., Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Gotlib, I. H. (2012). Walk on the bright side: Physical activity and affect in major depressive disorder. Journal of Abnormal Psychology, 121, 297–308. https://doi.org/10.1037/ a0023533 McAuley, E., Morris, K., Motl, R., Hu, L., Konopack, J., & Elavsky, S. (2007). Long-term follow-up of physical activity behavior in older adults. Health Psychology, 26, 375–380. https://doi.org/ 10.1037/0278-6133.26.3.375 O’Neill, B., McDonough, S., Wilson, J., Bradbury, I., Hayes, K., Kirk, A., . . . Tully, M. (2017). Comparing accelerometer, pedometer and a questionnaire for measuring physical activity in bronchiectasis: A validity and feasibility study. Respiratory Research, 18, 16. https://doi.org/10.1186/s12931-016-0497-2 Peluso, M., & Andrade, L. H. (2005). Physical activity and mental health: The association between exercise and mood. Clinics, 60, 61–70. https://doi.org/10.1590/S1807-59322005000100012 Pickett, K., Yardley, L., & Kendrick, T. (2012). Physical activity and depression: A multiple mediation analysis. Mental Health and Physical Activity, 5, 125–134. https://doi.org/10.1016/j.mhpa. 2012.10.001 Poole, L., Steptoe, A., Wawrzyniak, A. J., Bostock, S., Mitchell, E. S., & Hamer, M. (2011). Associations of objectively measured physical activity with daily mood ratings and psychophysiological stress responses in women. Psychophysiology, 48, 1165– 1172. https://doi.org/10.1111/j.1469-8986.2011.01184.x Prince, S., Adamo, K., Hamel, M., Hardt, J., Gorber, S., & Tremblay, M. (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review. The International Journal of Behavioral Nutrition and Physical Activity, 5, 56. https://doi.org/10.1186/1479-5868-5-56 Puterman, E., Weiss, J., Beauchamp, M., Mogle, J., & Almeida, D. (2017). Physical activity and negative affective reactivity in daily life. Health Psychology, 36, 1186–1194. https://doi.org/ 10.1037/hea0000532 Raudenbusch, S. W., Bryk, A. S., & Congdon, R. (2011). HLM7 for Windows [Computer software]. Skokie, IL: Scientific Software International. Rebar, A., Stanton, R., Geard, D., Short, C., Duncan, M., & Vandelanotte, C. (2015). A meta-meta-analysis of the effect of physical activity on depression and anxiety in non-clinical adult populations. Health Psychology Review, 9, 366–378. https://doi.org/10.1080/17437199.2015.1022901

European Journal of Health Psychology (2020), 27(2), 66–75

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

Reichenberger, J., Richard, A., Smyth, J. M., Fischer, D., Pollatos, O., & Blechert, J. (2018). It’s craving time: Time of day effects on momentary hunger and food craving in daily life. Nutrition, 55– 56, 15–20. https://doi.org/10.1016/j.nut.2018.03.048 Richardson, C., Faulkner, G., McDevitt, J., Skrinar, G., Hutchinson, D., & Piette, J. (2005). Integrating physical activity into mental health services for persons with serious mental illness. Psychiatric Services, 56, 324–331. https://doi.org/10.1176/appi.ps.56.3.324 Ronda, G., Assema, P., & Brug, J. (2001). Stages of change, psychological factors and awareness of physical activity levels in the Netherlands. Health Promotion International, 16, 305– 314. https://doi.org/10.1093/heapro/16.4.305 Rosenbaum, S., Tiedemann, A., Sherrington, C., Curtis, J., & Ward, P. (2014). Physical activity interventions for people with mental illness: A systematic review and meta-analysis. Journal of Clinical Psychiatry, 75, 964–974. https://doi.org/10.4088/JCP.13r08765 Sallis, J., & Saelens, B. (2000). Assessment of physical activity by self-report: Status, limitations, and future directions. Research Quarterly for Exercise and Sport, 71, 1–14. https://doi.org/ 10.1080/02701367.2000.11082780 Saris, W., Blair, S. N., van Baak, M., Eaton, S., Davies, P., Pietro, L., . . . Wyatt, H. (2003). How much physical activity is enough to prevent unhealthy weight gain? Outcome of the IASO 1st Stock Conference and consensus statement. Obesity Reviews, 4, 101– 114. https://doi.org/10.1046/j.1467-789X.2003.00101.x Schmidt, M., Cleland, V., Thomson, R., Dwyer, T., & Venn, A. (2008). A comparison of subjective and objective measures of physical activity and fitness in identifying associations with cardiometabolic risk factors. Annals of Epidemiology, 18, 378– 386. https://doi.org/10.1016/j.annepidem.2008.01.005 Schöndube, A., Kanning, M., & Fuchs, R. (2016). The bidirectional effect between momentary affective states and exercise duration on a day level. Frontiers in Psychology, 7, 1414. https://doi. org/10.3389/fpsyg.2016.01414 Schultchen, D., Reichenberger, J., Mittl, T., Weh, T. R. M., Smyth, J. M., Blechert, J., & Pollatos, O. (2019). Bidirectional relationship of stress and affect with physical activity and healthy eating. British Journal of Health Psychology, 24, 315–333. https://doi.org/10.1111/bjhp.12355 Schwerdtfeger, A., Eberhardt, R., & Chmitorz, A. (2008). Gibt es einen Zusammenhang zwischen Bewegungsaktivität und psychischem Befinden im Alltag? [Is there an association between physical activity and mental well-being in everyday life?] Zeitschrift für Gesundheitspsychologie, 16, 2–11. https://doi. org/10.1026/0943-8149.16.1.2 Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological momentary assessment. Annual Review of Clinical Psychology, 4, 1–32. Shiroma, E. J., Cook, N. R., Manson, J. E., Buring, J. E., Rimm, E. B., & Lee, I.-M. (2015). Comparison of self-reported and accelerometer-assessed physical activity in older women. PLoS One, 10, e0145950. https://doi.org/10.1371/journal.pone.0145950 Shook, R., Gribben, N., Hand, G., Paluch, A., Welk, G., Jakicic, J., . . . Blair, S. (2016). Subjective estimation of physical activity using the IPAQ varies by fitness level. Journal of Physical Activity and Health, 13, 79–86. https://doi.org/10.1123/jpah.2014-0543 Shrout, P. (1998). Measurement reliability and agreement in psychiatry. Statistical Methods in Medical Research, 7, 301– 317. https://doi.org/10.1191/096228098672090967 Shrout, P., & Lane, S. P. (2012). Psychometrics. In M. R. Mehl & T. S. Conner (Eds.), Handbook of research methods for studying daily life (pp. 302–320). New York, NY: The Guilford Press. Slootmaker, S., Schuit, A., Chinapaw, M., Seidell, J., & Mechelen, W. (2009). Disagreement in physical activity assessed by accelerometer and self-report in subgroups of age, gender, education and weight status. The International Journal of

Ó 2020 Hogrefe Publishing

B. Pannicke et al., Subjective / Objective Physical Activity and Affect

Behavioral Nutrition and Physical Activity, 6, 17. https://doi.org/ 10.1186/1479-5868-6-17 Storrie, K., Ahern, K., & Tuckett, A. (2010). A systematic review: Students with mental health problems – A growing problem. International Journal of Nursing Practice, 16, 1–6. https://doi. org/10.1111/j.1440-172X.2009.01813.x Thome, J., & Espelage, D. L. (2004). Relations among exercise, coping, disordered eating, and psychological health among college students. Eating Behaviors, 5, 337–351. https://doi.org/ 10.1016/j.eatbeh.2004.04.002 Thyregod, M., & Bodtger, U. (2016). Coherence between selfreported and objectively measured physical activity in patients with chronic obstructive lung disease: A systematic review. International Journal of Chronic Obstructive Pulmonary Disease, 11, 2931–2938. https://doi.org/10.2147/COPD.S116422 Tudor-Locke, C., & Bassett, D. (2004). How many steps/day are enough? Sports Medicine, 34, 1–8. https://doi.org/10.2165/ 00007256-200434010-00001 van Sluijs, E., Griffin, S., & van Poppel, M. (2007). A cross-sectional study of awareness of physical activity: Associations with personal, behavioral and psychosocial factors. The International Journal of Behavioral Nutrition and Physical Activity, 4, 53. https://doi.org/10.1186/1479-5868-4-53 van Sluijs, E., van Poppel, M., Twisk, J., & Mechelen, W. (2006). Physical activity measurements affected participants’ behavior in a randomized controlled trial. Journal of Clinical Epidemiology, 59, 404–411. https://doi.org/10.1016/j.jclinepi.2005.08.016 Warburton, D. E., Nicol, C. W., & Bredin, S. S. (2006). Health benefits of physical activity: The evidence. Canadian Medical Association Journal, 174, 801–809. Warren, J., Ekelund, U., Besson, H., Mezzani, A., Geladas, N., & Vanhees, L. (2010). Assessment of physical activity – A review of methodologies with reference to epidemiological research: A report of the exercise physiology section of the European Association of Cardiovascular Prevention and Rehabilitation. European Journal of Cardiovascular Prevention and Rehabilitation, 17, 127–139. https://doi.org/10.1097/HJR.0b013e32832ed875 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1070. Wichers, M., Peeters, F., Rutten, B., Jacobs, N., Derom, C., Thiery, E., . . . van Os, J. (2012). A time-lagged momentary assessment study on daily life physical activity and affect. Health Psychology, 31, 135–144. https://doi.org/10.1037/a0025688 World Health Organization. (2009). Global health risks: Mortality and burden of disease attributable to selected major risks. Geneva, Switzerland: World Health Organization.

Ó 2020 Hogrefe Publishing

World Health Organization. (2010). Global recommendations on physical activity for health. Geneva, Switzerland: World Health Organization. Wunsch, K., Kasten, N., & Fuchs, R. (2017). The effect of physical activity on sleep quality, well-being, and affect in academic stress periods. Nature and Science of Sleep, 9, 117–126. https://doi.org/10.2147/NSS.S132078 Zhai, L., Zhang, Y., & Zhang, D. (2014). Sedentary behaviour and the risk of depression: A meta-analysis. British Journal of Sports Medicine, 49(11), 705–709. https://doi.org/10.1136/ bjsports-2014-093613 Zschucke, E., Renneberg, B., Dimeo, F., Wüstenberg, T., & Ströhle, A. (2015). The stress-buffering effect of acute exercise: Evidence for HPA axis negative feedback. Psychoneuroendocrinology, 51, 414–425. https://doi.org/10.1016/j.psyneuen. 2014.10.019 History Received August 29, 2019 Revision received April 8, 2020 Accepted April 17, 2020 Published online July 1, 2020 Conflict of Interest The authors declare that there is no conflict of interest. Publication Ethics All participants received written and oral information on the purpose of the study and signed an informed consent according to the relevant ethics committee that also granted the ethical approval for the present study. Funding This work was supported by the Peer-Mentoring Team-Program (Line A) of the German Psychological Society (DGPs, section Health Psychology). In addition, this work was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (ERC-StG-2014 639445 NewEat) and the Austrian Science Fund (FWF): [I 02130B27]. Björn Pannicke Department of Psychology Paris-Lodron-University of Salzburg Hellbrunner Straße 34 5020 Salzburg Austria bjoern.pannicke@sbg.ac.at

European Journal of Health Psychology (2020), 27(2), 66–75

News and Announcements Meeting Calendar August 6–9, 2020 APA 2020 – Virtual Meeting Washington, DC, USA Contact: American Psychological Association, E-mail convention@apa. org, Web https://convention.apa.org/ August 24–25, 2020 8th World Congress on Nursing and Healthcare London, UK Contact: Scientific Future Group, E-mail secretary@congressnursing. info, Web http://congressnursing. info/2020/index.php July 6–9, 2021 17th European Congress of Psychology (ECP2021) Ljubljana, Slovenia Contact: Slovenian Psychological Association, Web https://www.sloveniaconvention.com/european-congresspsychology-2021-held-ljubljana/, and European Federation of Psychologists’ Associations, E-mail headoffice@efpa. eu

July 18–23, 2021 32th International Congress of Psychology (ICP) Prague, Czech Republic Contact: Congress Secretariat, Computer System Group a.s., 5. Kvetna 65, 140 21 Prague, Czech Republic, E-mail secretariat@icp2020.com, Web http:// www.icp2020.com/ August 12–15, 2021 APA 2021 San Diego, CA, USA Contact: American Psychological Association, E-mail convention@apa.org, Web https://convention.apa.org/

November 3–6, 2021 Work, Stress, and Health Conference Miami, FL, USA Contact: American Psychological Association, Washington, DC, https:// www.apa.org/wsh/ June 29–July 2, 2022 10th European Conference on Positive Psychology Reykjavik, Iceland Contact: SENA, Reykjavik, Iceland, E-mail gudrun@sena.is, Web https:// ecpp2020.com/ecpp/

August 22–28, 2021 EHPS Conference 2021 Bratislava, Slovakia Contact: European Health Psychology Society, E-mail info@easyconferences. eu, Web https://ehps.net/conferences/

European Journal of Health Psychology (2020), 27(2), 76 https://doi.org/10.1027/2512-8442/a000051

Ó 2020 Hogrefe Publishing

Instructions to Authors 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 practiceoriented articles, current methodological developments, and comprehensive critical reviews of the scientiﬁc literature. European Journal of Health Psychology publishes the following types of article: Original Articles, Review Articles, Conceptual Articles, Short Reports, Commentaries, and Book Reviews.

Manuscript submission: All manuscripts should in the first instance be submitted electronically at http://www.editorialmanager. com/zgp. Detailed instructions to authors are provided at http:// www.hogrefe.com/periodicals/zgp/advice-for-authors/ Copyright Agreement: By submitting an article, the author confirms and guarantees on behalf of themselves and any coauthors that the manuscript has not been submitted or published elsewhere, and that they holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches, and tables, and that the article and its contents do not infringe in any way on the rights of third parties. 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, or other form, in the journal to which it has been submitted and in other independent publications, with no limitations on the number of copies or on the form or the extent of distribution. These rights are transferred for the duration of copyright as defined by international law. Furthermore, the author

Ó 2020 Hogrefe Publishing

transfers to the publisher the following exclusive rights to the article and its contents: 1. The rights to produce advance copies, reprints, or offprints of the article, in full or in part, to undertake or allow translations into other languages, to distribute other forms or modified versions of the article, and to produce and distribute summaries or abstracts. 2. The rights to microfilm and microfiche editions or similar, to the use of the article and its contents in videotext, teletext, and similar systems, to recordings or reproduction using other media, digital or analog, including electronic, magnetic, and optical media, and in multimedia form, as well as for public broadcasting in radio, television, or other forms of broadcast. 3. The rights to store the article and its content in machine readable or electronic form on all media (such as computer disks, compact disks, magnetic tape), to store the article and its contents in online databases belonging to the publisher or third parties for viewing or downloading by third parties, and to present or reproduce the article or its contents on visual display screens, monitors, and similar devices, either directly or via data transmission. 4. The rights to reproduce and distribute the article and its contents by all other means, including photomechanical and similar processes (such as photocopying or facsimile), and as part of so-called document delivery services. 5. The right to transfer any or all rights mentioned in this agreement, as well as rights retained by the relevant copyright clearing centers, including royalty rights to third parties. Online Rights for Journal Articles: Guidelines on authors’ rights to archive electronic versions of their manuscripts online are given in the ’’Guidelines on sharing and use of articles in Hogrefe journals’’ on the journal’s web page at http://www.hogrefe.com/j/zgp. June 2020

European Journal of Health Psychology (2020), 27(2)

Quick and comprehensive information on psychotropic drugs for adults Also e ilabl ava e onlin

“As a psychiatrist who routinely prescribes medications, I find this handbook invaluable. It is the most comprehensive book on this topic that I have come across.” Corey Goldstein, MD, Rush University Medical Center

Ric M. Procyshyn / Kalyna Z. Bezchlibnyk-Butler / J. Joel Jeffries (Editors)

Clinical Handbook of Psychotropic Drugs 23rd ed. 2019, iv + 454 pp. + 63 pp. of printable PDF patient information sheets, spiral bound, large format US $99.80 / € 79.95 ISBN 978-0-88937-561-1 The Clinical Handbook of Psychotropic Drugs has become a standard reference and working tool for psychiatrists, psychologists, physicians, pharmacists, nurses, and other mental health professionals. This book is a must for everyone who needs an up-todate, easy-to-use, comprehensive summary of all the most relevant information about psychotropic drugs. • Independent, unbiased, up-to-date • Packed with unique, easy-to-read comparison charts and tables (dosages, side effects, pharmacokinetics, interactions…) for a quick overview of treatment options

www.hogrefe.com

• Succinct, bulleted information on all classes of medication: on- and off-label indications, side effects, interactions, pharmacodynamics, precautions in the young, the elderly, and pregnancy, nursing implications, and much more – all you need to know for each class of drug • Potential interactions and side effects summarized in comparison charts • With instantly recognizable icons and in full color throughout, allowing you to find at a glance all the information you seek • Clearly written patient information sheets available for download as printable PDF files

Behavioral sciences for the next generation of health care providers “This latest edition continues the editors’ outstanding success in bridging the exciting worlds of the physiological and behavioral-social contributions to quality health care.” Patrick DeLeon, PhD, MPH, JD, former President of the American Psychological Association

Olle Jane Z. Sahler / John E. Carr / Julia B. Frank / João V. Nunes (Editors)

The Behavioral Sciences and Health Care 4th edition 2018, xiv + 578 pp. US $69.00 / € 54.95 ISBN 978-0-88937-486-7 Also available as eBook

www.hogrefe.com

A clearly structured, compact text on the behavioral sciences for medical and other health professions, including practical features such as chapter review questions and an annotated practice exam.

related disorders, and theories of social relations and interprofessional collaboration as well as a new appendix on psychological testing as well as integrating the new classification of the DSM-5.

This edition of this easy-to-use text presents succinct information about the wide variety of social and psychological sciences that interact with the biological sciences in contributing to health and illness. Based around but expanding on the Integrated Sciences Model, and focusing on the Institute of Medicine’s key themes for medical training, this latest edition includes new chapters on pain, obsessive compulsive and

Each chapter begins with guidance questions and ends with current recommended readings and review questions. A complete 335 question-and-answer multiple choicetype review section not only allows readers to assess how well they have learned the material, but also highlights important points and adds additional specific information to supplement the text.

First papers now published! Be part of this exciting journey and submit your paper now!

https://www.editorialmanager.com/ptad

OA New al Journ

“PTAD is a unique academic journal, filling the burgeoning gap in the field of test adaptations and development – I’m very excited to announce the release of our first papers and look forward to even more submissions!” Matthias Ziegler, Editor-in-chief

Psychological Test Adaptation and Development (PTAD) is the first open access, peer-reviewed journal publishing papers on adaptations of tests to specific (e.g., cultural) needs, test translations, or the development of existing measures. For more information, visit https://eu.hogrefe.com/products/journals/ptad Why you should publish in PTAD: • A unique outlet for research papers portraying adaptations (e.g., translations) and developments (e.g., state to trait) of individual tests – the backbone of assessment • Official open-access journal of the European Association of Psychological Assessment (EAPA) • With an expert editor-in-chief, supported by a stellar cast of internationally renowned associate editors • Fully embraces open science – including registered reports

Benefits for authors: • Clear guidance on the structure of papers • Fast peer-review • With the optional registered report format you can get expert advice from seasoned reviewers to help improve your research • Open access publication, with a choice of Creative Commons licenses • Widest possible dissemination of your paper – and thus of qualified information about your test and your research • Generous APC waiver program

Introductory Editorials: Matthias Ziegler, Psychological Test Adaptation and Development – A New Beginning https://doi.org/10.1027/2698-1866/a000001

Original Article: Ziwen Teuber, Qichen Wang, Yanjie Su, Arnold Lohaus, & Fridtjof W. Nussbeck, Human Resources in Chinese Youngsters – A Chinese Adaptation of the QARCA https://doi.org/10.1027/2698-1866/a000003

Matthias Ziegler, Psychological Test Adaptation and Development – How Papers Are Structured and Why https://doi.org/10.1027/2698-1866/a000002