Zsp 2017 48issue 1 by Hogrefe

Volume 48 / Number 1 / 2017

Social Psychology

Editor-in-Chief Kai Epstude Associate Editors Julia Becker Adam Fetterman Malte Friese Ilka Gleibs Michael Häfner Hans J. IJzerman Ulrich Kühnen Toon Kuppens Ruth Mayo Christian Unkelbach Michaela Wänke

How to meet people’s increasing need to develop and manage their own lives and careers “A must read for all career development professionals and students alike. The Handbook of the Life Design is a new and essential resource for those working to improve career services in line with today’s challenges and conditions.” Sara Santilli, PhD, writing in Career Convergence, February 2015

Laura Nota / Jérôme Rossier (Editors)

Handbook of Life Design

From Practice to Theory and from Theory to Practice 2015, vi + 298 pp., hardcover US $54.00 / € 38.95 ISBN 978-0-88937-447-8 Also available as eBook Our lives and careers are becoming ever more unpredictable. The “lifedesign paradigm” described in detail in this ground-breaking handbook helps counselors and others meet people’s increasing need to develop and manage their own lives and careers. Life-design interventions, suited to a wide variety of cultural settings, help

www.hogrefe.com

individuals become actors in their own lives and careers by activating, stimulating, and developing their personal resources. This handbook first addresses life-design theory, then shows how to apply life designing to different age groups and with more at-risk people, and looks at how to train life-design counselors.

Social Psychology

Volume 48, No. 1, 2017

Editor-in-Chief

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

Editorial Ofﬁce

Wim Meerholz, University of Groningen, Department of Psychology, Grote Kruisstraat 2/1, 9712 TS Groningen,The Netherlands, Tel. +31 50 363-6393, Fax + 31 50 363-4581, E-mail SocialPsych.EditorialOfﬁce@gmail.com Hans IJzerman, Vrije Universiteit Amsterdam, Julia Becker, Universita¨t Osnabru¨ck, Germany The Netherlands Adam Fetterman, University of Texas, USA Ulrich Ku¨hnen, Jacobs University, Germany Malte Friese, Universita¨t des Saarlandes, Toon Kuppens, University of Groningen, Germany The Netherlands Ilka Gleibs, London School of Economics, Ruth Mayo, Hebrew University of Jerusalem, Israel United Kingdom Christian Unkelbach, University of Cologne, Germany Michael Ha¨fner, Universita¨t der Ku¨nste Berlin, Michaela Wa¨nke, Universita¨t Mannheim, Germany Germany

Associate Editors

Consulting Editors

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

Publisher

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

Production

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

Subscriptions

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

Advertising/Inserts

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

ISSN

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

Ó 2017 Hogrefe Publishing. This journal as well as the individual contributions and illustrations contained within it are protected under international copyright law. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microﬁlming, recording or otherwise, without prior written permission from the publisher. All rights, including translation rights, reserved.

Publication

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

Subscription Prices

Calendar year subscriptions only. Rates for 2017: Institutions 374.00/US $478.00; Individuals 159.00/US $223.00 (all plus US $24.00/ 18.00 shipping & handling; Germany: 5.00).

Payment

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

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

Electronic Full Text

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

Abstracting Services

Abstracted/indexed in Current Contents/Social and Behavioral Sciences (CC/S&BS), Social Sciences Citation Index (SSCI), PsycINFO, PASCAL, PSYNDEX, ERIH, Scopus, and EM Care. Impact Factor (2015): 1.979

Social Psychology (2017), 48(1)

Ó 2017 Hogrefe Publishing

Contents Editorial

Towards a Replicable and Relevant Social Psychology Kai Epstude

Original Articles

The Malleability of Stereotype Effects on Spontaneous Trait Inferences: The Moderating Role of Perceiversâ&#x20AC;&#x2122; Power Meifang Wang and Feng Yang

Relational Utility Affects Self-Punishment in Direct and Indirect Reciprocity Situations Ruida Zhu, Tao Jin, Xueyi Shen, Shen Zhang, Xiaoqin Mai, and Chao Liu

Stressing the Advantages of Female Leadership Can Place Women at a Disadvantage Joris Lammers and Anne Gast

Stepping in the Shoes of Leaders of Populist Right-Wing Parties: Promoting Anti-Immigrant Views in Times of Economic Prosperity Jolanda Jetten, Rachel Ryan, and Frank Mols

Glucose Increases Risky Behavior and Attitudes in People Low in Self-Control: A Pilot Study Michaela Pfundmair, Eva Lermer, and Dieter Frey

Projection of Visceral Needs: Satisfaction of the Need and Similarity of the Target Person as Moderators Janet N. Ahn, Gabriele Oettingen, and Peter M. Gollwitzer

Research Reports

Replication

Ă&#x201C; 2017 Hogrefe Publishing

Social Psychology (2017), 48(1)

Editorial Towards a Replicable and Relevant Social Psychology Kai Epstude Department of Psychology, University of Groningen, The Netherlands

In the last few months, I had many awkward conversations. When people heard that I will take on the role of Editor-inChief the responses varied from “Why do you do this to yourself?” to “Oh. The replication journal. You’re one of THEM now?” Here are my standard answers: I like to read and discuss research. And, while Social Psychology became known to many non-German colleagues through the Special Issue on replications (Nosek & Lakens, 2014), it publishes a mix of original research and replications. Those anecdotes also illustrate lingering issues in the field. There are still many uncertainties on how to deal with various factors that make social psychological research challenging. On the one hand, there are concerns about the replicability of findings in social psychology. Those attract considerable attention in traditional and social media. On the other hand, researchers in many countries are faced with a changing academic environment that asks for research that has actual societal impact. This factor is rarely debated outside of the academic environment despite its direct influence on the topics investigated. An unfortunate fact is that those two main issues ask for research strategies that are not always easy to combine. The efforts in replicating existing findings have led to a general consensus that sample sizes need to increase, and that the documentation of the research process and the findings need to be improved. Many colleagues have outlined the path forward in excellent contributions (e.g., Fiedler, 2017). However, inherent to the field of social psychology is the problem that we examine a person’s behavior in a certain situation or social context (Lewin, 1946). Though this idea is covered in every introductory psychology textbook, it has often been neglected, leading to sweeping generalizations when interpreting original findings as well as some replication attempts. In line with this classic idea, replication attempts focusing on studies with less (social) context seem to replicate better (Van Bavel, MendeSiedlecki, Brady, & Reinero, 2016). Despite the fact that the latter argument leads to a separate debate (e.g., Inbar, 2016), it is apparent that the nature of concepts studied Ó 2017 Hogrefe Publishing

in our field oftentimes (e.g., cultural stereotypes) makes a direct replication without any adjustments difficult. It is therefore tempting to reduce a research question to its most fundamental core and examine it in highly controlled conditions. However, a highly replicable finding is not necessarily an impactful finding (see also Maner, 2016). In order to be impactful, it also needs to be tested in terms of its influence on behavior outside of the lab. In various countries, it becomes more and more important (if not mandatory) that knowledge produced in a project has demonstrable implication for the “real world.” In other words, findings are expected to have an impact. This pressure is coming from funding organizations, as well as from universities aiming to establish a meaningful profile that attracts students. For example, in the Netherlands the former Marxist term “valorization” received a new meaning by becoming a criterion for the evaluation of grant proposals. Applicants have to demonstrate the relevance of their findings for society and the economy. Moreover, a stronger collaboration between the corporate sector and universities is emphasized in order to increase the societal impact of research (NWO, 2017). In Germany, the introduction of differentiated master programs, oftentimes with an applied focus, has raised concerns about the marginalization of more fundamental subdisciplines of psychology (Bermeitinger et al., 2016). Both types of pressures do point to the necessity for individual researchers to study phenomena in more applied settings, both to secure funding as well as to have a professional future in academia. It is of course not the case that replicability is incompatible with impact. In line with Maner (2016), I would argue that they complement each other. Fundamental research is needed to establish an effect, while applied research helps to test the implications and the impact of the effect. Combining both goals oftentimes leads to research that might just not be applied or fundamental enough to be suitable for journals that are only interested in one or the other. Social Psychology is open for such an approach. Especially in a time when many societal questions also require social Social Psychology (2017), 48(1), 1–2 DOI: 10.1027/1864-9335/a000303

psychologists to provide answers, it would be detrimental for the future of the field to be completely absorbed by a debate that only focuses on replicability while at the same time losing sight of the (societal) impact of our research. I’m completely aware of the fact that such a statement can be met with relentless cynicism. However, I also think that the standing of the field can be improved if we show that our findings do matter outside of the lab. The editorial team of Social Psychology will continue to strive for a replicable (Unkelbach, 2016) as well as impactful social psychology.

Adjustments in the Editorial Policy I would also like to slightly specify some aspects of our editorial policy. For several years now, Social Psychology publishes replication reports. Starting with a whole replication issue, this format has gained a lot of attention. By now, many psychological journals offer the opportunity to publish replication reports. Despite the broadened interest in replications, the standards for what is a “good” replication and which original studies should be replicated only develop slowly. This leads to some uncertainties both on the side of researchers as well on the side of editors. We aim to publish replication reports. However, Social Psychology reserves the right to decide whether a replication is suitable for the journal. This means that like regular submissions replication attempts are evaluated in terms of whether they are within the scope of the journal, and whether they are of interest to the readers. Social Psychology strongly encourages open data and open materials to become the default for publications in the journal. In case of preregistered studies, the respective paper will receive an additional note at the end of the paper stating the details of the preregistration. In light of the growing importance of collaborative research projects, future submissions should also include a note stating the contribution of each author.

Comings and Goings I would like to express my gratitude to the previous Editor-inchief, Christian Unkelbach, and the former Editorial Assistant Juliane Burghardt. Both did a great job in the time of transition and were responsible for much of the success of the journal in recent years. I’m even more grateful, that Christian has decided to remain with the journal until this spring in the role of an Associate Editor. I would also like to thank Julia Becker, Malte Friese, and Markus Kemmelmeier for their work. They served as Associate Editors for many years and thereby helped to shape the journal’s profile.

Social Psychology (2017), 48(1), 1–2

Editorial

I would like to welcome Adam Fetterman (University of Texas at El Paso, USA), Ilka Gleibs (London School of Economics and Political Science, London, UK), and Toon Kuppens (University of Groningen, The Netherlands) as new Associate Editors. All of them started in 2016. Later this year, Anna Baumert (Max Planck Institute for Collective Goods, Bonn, Germany) and Kim Peters (University of Queensland, Brisbane, Australia) will join the editorial team as Associate Editors. Wim Meerholz (University of Groningen, The Netherlands) is the new Editorial Assistant. Together with the already experienced Associate Editors Michael Häfner, Hans IJzerman, Ulrich Kühnen, Ruth Mayo, and Michaela Wänke the new editorial team is complete. I hope that we can continue the success of Social Psychology as a journal and help to strengthen social psychology as a field.

References Bermeitinger, C., Kaup, B., Kiesel, A., Koch, I., Kunde, W., Müsseler, J., . . . Ulrich, R. (2016). Positionspapier zur Lage der Allgemeinen Psychologie [On the state of experimental psychology]. Psychologische Rundschau, 67, 175–179. doi: 10.1026/0033-3042/a000317 Fiedler, K. (2017). What constitutes strong psychological science? The (neglected) role of diagnosticity and a priori theorizing. Perspectives on Psychological Science, 12, 46–61. doi: 10.1177/ 1745691616654458 Inbar, Y. (2016). Association between contextual dependence and replicability in psychology may be spurious. Proceedings of the National Academy of Sciences, 113, E4933–E4934. doi: 10.1073/pnas.1608676113 Lewin, K. (1946). Behavior and development as a function of the total situation. In L. Carmichael (Ed.), Manual of child psychology (pp. 791–844). Hoboken, NJ: Wiley. Maner, J. K. (2016). Into the wild: Field research can increase both replicability and real-world impact. Journal of Experimental Social Psychology, 66, 100–106. doi: 10.1016/j.jesp.2015.09.018 Nosek, B. A., & Lakens, D. (2014). Registered reports. Social Psychology, 45, 137–141. doi: 10.1027/1864-9335/a000192 NWO. (2017, January 27). Reactie NWO op Kamerbrief “Wetenschap met impact” [Response of the NWO to the parliamentary letter on science with an impact]. Unkelbach, C. (2016). Increasing replicability. Social Psychology, 47, 1–3. doi: 10.1027/1864-9335/a000270 Van Bavel, J. J., Mende-Siedlecki, P., Brady, W. J., & Reinero, D. A. (2016). Contextual sensitivity in scientific reproducibility. Proceedings of the National Academy of Sciences, 113, 6454–6459. doi: 10.1073/pnas.1521897113

Kai Epstude Department of Psychology University of Groningen Grote Kruisstraat 2/1 9712 TS Groningen The Netherlands k.epstude@rug.nl

Ó 2017 Hogrefe Publishing

Original Article

The Malleability of Stereotype Effects on Spontaneous Trait Inferences The Moderating Role of Perceivers’ Power Meifang Wang and Feng Yang Department of Psychology, Shandong Normal University, Jinan, PR China

Abstract: Past research has demonstrated that perceivers are more likely to draw spontaneous trait inferences (STIs) from stereotypeconsistent behaviors than from stereotype-inconsistent behaviors. Four studies were conducted to examine the moderating role of power in stereotype effects on STIs. Priming power using the scrambled sentence task, Study 1 found that high-power participants drew STIs from elderly stereotype-consistent but not from elderly stereotype-inconsistent sentences, while low-power participants did not draw STIs from elderly stereotype-consistent or stereotype-inconsistent behaviors. Study 2 replicated the findings of Study 1 by exploring the moderating role of power in gender stereotype effects on STIs. Measuring participants’ dispositional power via the Personal Sense of Power, Study 3 found that dispositional power also moderated the effects of gender stereotype on STIs. Study 4 found that compared with the baseline condition (nopower manipulation), the low-power condition inhibited STIs from stereotype-consistent behaviors, but the high-power condition did not facilitate STI formation from stereotype-consistent behaviors. The current study is the first to show that power influences the reliance on stereotypes when spontaneously inferring traits from behaviors. Keywords: power, stereotypes, spontaneous trait inferences, impression formation

In social life, people often spontaneously form an impression of others without having a particular goal or even a general intention to do so (Todorov & Uleman, 2002). For example, when people read the sentence “the secretary solved the mystery halfway through the book,” they spontaneously infer the trait “clever” (Winter & Uleman, 1984). Winter and Uleman (1984) dubbed such unintentional trait inferences as spontaneous trait inferences (STIs). According to Uleman, Newman, and Moskowitz’s (1996) definition, STIs refer to trait inferences that are drawn from behaviors in the absence of specific intentions to infer traits or to form an impression of a target person. Past research has found that an actor’s salient features (e.g., gender, age, and race) can automatically activate corresponding stereotypes and influence the formation of subsequent STIs (Wigboldus, Dijksterhuis, & van Knippenberg, 2003; Wigboldus, Sherman, Franzese, & van Knippenberg, 2004; Yan, Wang, & Zhang, 2012). When people read the sentence “the skinhead [girl] hits the saleswoman,” they are more likely to infer that the skinhead is aggressive than that the girl is aggressive (Wigboldus et al., 2003). However, the effects of stereotypes on STIs may vary according to perceivers’ power. Ample research has demonstrated that compared with low-power individuals, high-power individuals are Ó 2017 Hogrefe Publishing

more likely to form an impression of a target person by relying on stereotype-consistent information (e.g., Goodwin, Gubin, Fiske, & Yzerbyt, 2000; Guinote & Phillips, 2010; Weick & Guinote, 2008). The current study is the first to extend previous work on stereotypes and STIs by exploring the potential moderating role of perceivers’ power in the effects of stereotypes on STIs.

Stereotypes and STIs Stereotypes, defined as a set of beliefs about the characteristics, attributes, and behaviors of members of certain groups (Hilton & Von Hippel, 1996), have a profound effect on individuals’ cognitive processes. A large body of research has documented that stereotypes can be automatically activated when social category cues (e.g., age, race, and gender) are presented, even when these cues are subliminally presented (for a review, see Blair, 2002). Once activated, stereotypes can guide the subsequent person perception tasks (e.g., Dijksterhuis & van Knippenberg, 1996; Perdue & Gurtman, 1990; Rudman, & Borgida, 1995). For instance, Dijksterhuis and van Knippenberg (1996) primed participants’ stereotypes about soccer hooligans by asking them to imagine a typical soccer hooligan for 5 min and to record that person’s behaviors, Social Psychology (2017), 48(1), 3–18 DOI: 10.1027/1864-9335/a000288

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

lifestyle, and physical attributes. The researchers then administered a lexical decision task and found that participants responded faster to stereotype-consistent traits (e.g., aggressive, violent, and prejudiced) and slower to stereotype-inconsistent traits (e.g., intelligent, friendly, and understanding), suggesting that stereotypes may foster the formation of stereotype-confirming impressions and undermine the formation of stereotype-disconfirming impressions. Wigboldus et al. (2003) were the first researchers to examine the effects of stereotypes on STIs. In their study, participants were presented with three types of traitimplying behavior sentences (stereotype-consistent, stereotype-inconsistent, and stereotype-neutral). Prior to the presentation of each sentence, stereotypes were activated by a category label. For example, the sentence “X hits the saleswoman without any apparent reason” was preceded by the label “skinhead,” “girl,” or “human.” The researchers found that participants were more likely to draw STIs from stereotype-consistent behaviors than from stereotype-inconsistent behaviors. Recently, Yan et al. (2012) primed stereotypes with photos rather than a semantic stimulus and examined the effects of gender stereotypes on STIs. They also found that perceivers more readily drew STIs from stereotype-consistent behaviors than from stereotype-inconsistent behaviors (Yan et al., 2012). According to the accessibility theory of knowledge activation (Dijksterhuis & van Knippenberg, 1996; Förster & Liberman, 2007; Higgins, 2012), once activated, stereotypes can enhance access to stereotype-consistent traits and reduce access to stereotype-inconsistent traits. The accessibility of traits may influence the trait inferences made from behaviors (Wigboldus et al., 2003). As a result, perceivers are more likely to make STIs from stereotypeconsistent behaviors than from stereotype-inconsistent behaviors.

Power and Stereotypes In social psychology, power is often defined as the capacity to alter others’ states by providing or withholding resources and administering punishments (Emerson, 1962; Keltner, Gruenfeld, & Anderson, 2003). Power is a fundamental part of our social life and has profound effects on individuals’ psychological and behavioral processes (Anderson, John, & Keltner, 2012). Previous theoretical and empirical work has suggested that power promotes individuals’ reliance on stereotypes when forming social perceptions (Fiske, 1993; Fiske & Dépret, 1996; Goodwin et al., 2000; Guinote & Phillips, 2010; Keltner et al., 2003). Generally speaking, high-power individuals live in environments with abundant resources and perceive fewer constraints from others, whereas low-power individuals have Social Psychology (2017), 48(1), 3–18

less access to resources and are more subject to social threats and punishments from others (Domhoff, 1998; French & Raven, 1959; Weber, 1947). In terms of the approach/inhibition theory of power, high-power individuals construe social events in a rapid and effortless manner because such individuals are less motivated to care about the consequences of their judgments for others, whereas low-power individuals interpret social behaviors in a deliberate and effortful manner because they need to take into account multiple sources of information to increase the accuracy of their social judgments (Keltner et al., 2003). Although stereotypes can simplify information processing by serving as cognitive shortcuts and allowing individuals to handle issues more effortlessly and efficiently, they may lead to less accurate perceptions of others (Bargh, 1999; Fiske, 1998). Therefore, according to the approach/inhibition theory of power, high-power individuals’ social cognition should rely more on stereotypes than low-power individuals’ social cognition. The situated focus theory of power (Guinote, 2007a, 2010; Weick & Guinote, 2008), another prevailing theory about power and cognition, proposes that power promotes stereotyping when stereotypes are accessible in a given situation. Powerful individuals perceive fewer constraints from external environments and can attain their desired outcomes more easily. Thus, they can allocate their cognitive resources more selectively, leading to more cognitive flexibility. By contrast, low-power individuals perceive more constraints from external environments and cannot easily attain their desired outcomes. Therefore, they need to pay attention to multiple sources of information to increase their prediction of and control over outcomes, resulting in less cognitive flexibility. Due to powerful individuals’ cognitive flexibility, their cognitive processes are more likely to be guided by accessible constructs in a given situation than are those of low-power individuals (Guinote, 2007b; Weick & Guinote, 2008). Therefore, when stereotype-relevant cues are presented in a situation, high-power individuals should be more susceptible to stereotypes than low-power individuals. Ample empirical research has demonstrated that powerful individuals are more likely than powerless individuals to stereotype others when forming an impression of them (Goodwin et al., 2000; Guinote & Phillips, 2010; Guinote, Willis, & Martellotta, 2010; Weick & Guinote, 2008). For example, Fiske and her colleagues (Goodwin et al., 2000) found that powerful individuals attended more to stereotype-consistent information than to stereotypeinconsistent information when they were asked to evaluate targets, whereas powerless individuals paid equal attention to stereotype-consistent and stereotype-inconsistent information. Guinote and Phillips’ (2010) research examining the hotel industry provides more convincing evidence. Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

In their study, 39 managers were selected as powerful people and 39 subordinates were selected as powerless people. The participants were presented with stereotypeconsistent and stereotype-inconsistent information about a target, and their task was to judge the target’s suitability for a job. The results showed that managers tended to consider targets who performed stereotype-consistent roles more suitable for the job than targets who performed stereotype-inconsistent roles, but subordinates’ judgments did not vary according to stereotype consistency. Recent research by Guinote et al. (2010) suggests that power may also increase individuals’ stereotyping of others even when they do not intend to form an impression. In this study, participants viewed photos of White/Black people and were then asked to categorize adjectives according to their valence (positive/negative). The results showed that high-power participants, but not low-power participants, responded faster to stereotype-consistent than to stereotype-inconsistent words after viewing photos of Black/White people, indicating that power promotes the expression of implicit racial stereotypes.

Power, Stereotypes, and STIs According to the theory of Person Situation interaction (Kunda & Spencer, 2003; Shoda & Mischel, 1993), a person’s responses to the external social environment are commonly determined by his/her features and contextual cues. Supporting the proposition of the theory, past research in the field of STIs has suggested that the effects of stereotypes on STIs are flexible and can be moderated by factors relevant to perceivers (Wang, Xia, & Yang, 2015; Wigboldus et al., 2004). For instance, Wigboldus et al. (2004) found that participants with high cognitive load were more likely to draw STIs from stereotype-consistent than stereotype-inconsistent behaviors, whereas participants with low cognitive load did not exhibit variation in STIs as a function of stereotypes. More recently, Wang et al. (2015) found that perceivers’ mood also moderated the effects of stereotypes on STIs. Individuals with positive mood were more likely to form STIs from stereotypeconsistent behaviors than from stereotype-inconsistent behaviors, whereas those with negative mood did not draw STIs from stereotype-consistent or stereotype-inconsistent behaviors. In the present research, we reasoned that perceivers’ power would moderate the effects of stereotypes on STIs. Based on prior literature (e.g., Fiske, 1993; Goodwin et al., 2000; Keltner et al., 2003), we predicted that high-power individuals would be more likely to draw STIs from stereotype-consistent behaviors than from stereotypeinconsistent behaviors, whereas low-power individuals

Ó 2017 Hogrefe Publishing

would be impervious to the influences of stereotypes. Four studies were conducted to test our hypothesis. In Studies 1 and 2, we manipulated power through conceptual priming and explored the moderating role of power in elderly/gender stereotype effects on STIs. In Study 3, we aimed to replicate the findings obtained in the previous two studies by measuring participants’ dispositional power. In Study 4, we added a baseline condition (no-power manipulation) to further clarify the direction of the effects of power on the relationship between stereotypes and STIs.

Study 1 The primary objective of Study 1 was to explore whether perceivers’ power moderates the effects of elderly stereotypes on STIs. We employed the probe recognition paradigm, which was developed by McKoon and Ratcliff (1986) and has been shown to be well suited for detecting STIs (Ham & Vonk, 2003; Uleman, Hon, Roman, & Moskowitz, 1996; Wigboldus et al., 2003; Yan et al., 2012). In this paradigm (McKoon & Ratcliff, 1986), participants are shown trait-implying and control sentences on a computer screen. Each sentence is immediately followed by a probe word. The participants’ task is to quickly and accurately indicate whether the probe word was included in the preceding sentence. In the experimental trials, probe words are traits that are implied by traitimplying sentences (e.g., “the secretary solved the mystery halfway through the book,” followed by “clever”). If participants spontaneously infer the trait during their encoding of the trait-implying sentence, it should be relatively difficult for them to indicate that the probe was not in the preceding sentence. Thus, participants display longer reaction times (RTs) (and, occasionally, more errors) when responding to trait-implying sentences than to control sentences (Yan et al., 2012).

Method Participants and Design A total of 81 undergraduate students (24 males and 57 females, M = 19.30, SD = 0.87) who attended a university in China volunteered to participate in the current study. The experimental design was a 2 (Power: high vs. low) 3 (Sentence type: stereotype-consistent vs. stereotype-inconsistent vs. control) mixed design, with power as a between-subjects factor and sentence type as a within-subjects factor.

Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Materials

Procedure

Experimental Trials In Study 1, the experimental trials included elderly stereotype-consistent/-inconsistent sentences and corresponding probe traits. The experimental trials proceeded as follows. Six elderly stereotype-consistent traits (grandfatherly, calm, prestigious, forgetful, weak, garrulous) and six elderly stereotype-inconsistent traits (strong, flexible, lively, adventurous, unorthodoxy, hasty) were selected based on the results of a pretest. Correspondingly, six elderly stereotype-consistent and six elderly stereotypeinconsistent sentences were created, and each sentence implied the corresponding trait. For example, the elderly stereotype-consistent sentence “the elder repeated the same thing all afternoon” implies the trait “garrulous,” and the elderly stereotype-inconsistent sentence “the elder carried a gas cylinder to the 5th floor” implies the trait “strong.” In a pilot study, we asked 33 participants to indicate the extent to which each of the 12 behaviors described elderly people on a 5-point scale (1 = not at all, 5 = extremely). The results indicated that elderly stereotype-consistent sentences were rated as more descriptive of elderly people than elderly stereotypeinconsistent sentences (Mcon = 4.23, SD = 0.39; Mincon = 2.12, SD = 0.38), F(1, 32) = 651.84, p < .001, η2p = .953. The probes following the elderly stereotype-consistent/ -inconsistent sentences were traits implied by the corresponding sentences.

Upon their arrival, participants were seated at individual computer stations and were instructed that they would engage in two (ostensibly) unrelated tasks: a scrambled sentence task and a memory task. The scrambled sentence task consisted of 16 items, and each item included five words listed in random order. Participants were asked to use those five words to construct a grammatically correct sentence. For the high-power prime, of the 16 items, 8 were fillers and 8 contained a word related to the possession of power (control, dominate, command, govern, influenced, authority, captain, executive). For the low-power prime, 8 of the 16 items were fillers, and the remaining 8 items contained a word related to a lack of power (submit, yield, comply, obey, compromise, subordinate, dustman, dependent). After completing the scrambled sentence task, participants performed the so-called memory test. During the memory task, participants viewed a series of sentences presented on the computer screen. Participants were informed that a word would appear on the screen after each sentence and that they simply needed to indicate as quickly and accurately as possible whether the probe word was present in the preceding sentence. Participants were instructed to indicate a “yes” answer by pressing the key marked “Yes” and to indicate a “no” answer by pressing the key marked “No.” Before the experiment started, participants completed four practice trials to gain familiarity with the procedure. As illustrated in Figure 1, the experiment began with a stimulus sentence presented on the screen for 1,500 ms (ms). Then, a blank screen was presented for 1,000 ms, followed by a probe word presented in the center of the screen. After each response, a blank screen appeared for 800 ms before the next trial started. The order of the presentation of the 60 trials was randomized. A computer recorded participants’ responses and RTs to the nearest millisecond. Following the memory test, participants were checked for suspicion, debriefed, and thanked.

Control Trials Twelve control sentences corresponding to the 12 traitimplying sentences were also developed. The control sentences contained most of the words in the trait-implying sentences but did not imply the target traits. The probes following control sentences were traits implied by the corresponding trait-implying sentences (e.g., the control sentence “the elder addressed an important thing in the afternoon” was followed by the trait “forgetful”). Filler Trials If we presented only experimental and control trials, the correct answer for all trials would be “No.” To balance the “Yes” and “No” responses, 36 filler trials were added. The filler sentences were similar in length to experimental sentences, but they did not imply specific traits (e.g., the filler sentence “the elder encountered an old friend while walking in the street” was followed by the probe “friend”). Among the filler trials, 30 should elicit “Yes” responses and 6 should elicit “No” responses. To ensure that participants paid attention to the entire sentence rather than only adjectives included in the sentence, adjectives, verbs, and nouns were used as probes in the filler trials. Social Psychology (2017), 48(1), 3–18

Results and Discussion Reaction Times Only experimental and control trials were included in the data analysis. The principal dependent variable was the participants’ RTs, which refers to the amount of time participants required to indicate that a probe word either was or was not presented in the preceding sentence (Wigboldus et al., 2003; Yan et al., 2012). Four participants (1 male, 3 females) with high error rates (greater than 50%) and who may have failed to follow our instructions were Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Figure 1. The schema of the probe recognition task (so-called memory test) procedure in Study 1. The behavioral sentences included 12 trait-implying sentences (6 elderly stereotype-consistent and 6 elderly stereotype-consistent sentences), 12 control sentences, and 36 filler sentences. The computer presented these sentences in random order.

dropped from the data analysis; their inclusion did not change the pattern of the data or any critical inferential test. The analyses of the full data set can be found in the Electronic Supplementary Materials. Following previous research (Yan et al., 2012), we excluded incorrect “yes” responses during the experimental trials and RTs that were regarded as outliers (greater than 2.5 SDs above or below the mean) from the data analyses. As a result, 3.13% of the total data were disregarded. Participants’ RTs were submitted to a 2 (Power: high vs. low) 3 (Sentence type: consistent vs. inconsistent vs. control) mixed-design analysis of variance (ANOVA) with sentence type as a within-subjects variable. Table 1 shows mean RTs in milliseconds and error rates as a function of sentence type and power priming. The ANOVA revealed a significant main effect of sentence type, F(2, 150) = 7.05, p = .002, η2p = .086, indicating that participants’ STI formation was influenced by the activated elderly stereotypes. Specific comparisons found that participants required more time to correctly reject probes following elderly stereotype-consistent sentences than those following control sentences (M = 790 and 750, respectively), F(1, 76) = 11.94, p = .001, η2p = .136. However, RTs did not differ between elderly stereotype-inconsistent and control sentences (M = 755 and 750, respectively), F(1, 76) = 0.19, p = .667, η2p = .002. More importantly, as we predicted, this main effect was qualified by a significant interaction effect between power and sentence type, F(2, 150) = 3.20, p = .048, η2p = .041. Simple main effects analyses showed that participants’ RTs varied according to sentence type in the high-power priming condition, F(2, 78) = 9.05, p < .001, η2p = .188, whereas no significant differences were found between different sentence types in the low-power priming condition, F(2, 72) = 0.44, p = .646, η2p = .012.1 Specific comparisons revealed that high-power

Table 1. Mean reaction times in milliseconds and error rates to trait probes as a function of sentence type and power priming in Study 1 Sentence type Power priming

Stereotypeconsistent

Stereotypeinconsistent

Control

Reaction times High power

822 (149)

766 (127)

757 (118)

Low power

755 (154)

743 (126)

744 (143)

High power

.02 (0.05)

.01 (0.04)

.01 (0.03)

Low power

.02 (0.06)

.02 (0.07)

.01 (0.02)

Error rates

Note. Standard deviations are given in parentheses.

participants had longer RTs to probes following elderly stereotype-consistent sentences than those following control sentences, F(1, 39) = 15.39, p < .001, η2p = .283, whereas no significant differences were found between elderly stereotype-inconsistent sentences and control sentences, F(1, 39) = 0.46, p = .504, η2p = .012. This result suggests that high-power participants made STIs from elderly-consistent sentences but not from stereotypeinconsistent sentences.

Error Rates The error rates were analyzed in a 2 (Power) 3 (Sentence type) mixed-design ANOVA. The main effects of power and sentence type were not significant, F(1, 75) = 0.39, p = .534, η2p = .005, and F(2, 150) = 1.80, p = .169, η2p = .023, respectively, and the interaction effect between power and sentence type was not significant, F(2, 150) = 0.38, p = .688, η2p = .005. These results are not surprising because the probe recognition task was quite easy for

In post hoc tests, we first examined whether participants made STIs from stereotype-consistent/ stereotype-inconsistent sentences by comparing the RTs for those sentences (stereotype-consistent and stereotype-inconsistent sentences) with the RTs for control sentences. We further compared the magnitude of the STIs between stereotype-consistent and stereotype-inconsistent sentences only when participants made STIs from both types of behavioral sentences (showing longer RTs than control behavioral sentences). Similarly, the magnitude of the STIs between high-power and low-power participants was compared only when both groups of participants made STIs from stereotype-consistent/ stereotype-inconsistent sentences.

Ó 2017 Hogrefe Publishing

Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

undergraduates (Wigboldus et al., 2003) and the majority of participants in both conditions performed well on the task. In this study, we explored the effects of elderly stereotypes on STIs and the moderating role of power. We found that participants tended to draw STIs from stereotype-consistent sentences rather than from stereotype-inconsistent sentences. More importantly, consistent with our prediction, the effects of stereotypes on STIs were moderated by perceivers’ power. Specifically, high-power participants drew STIs from stereotype-consistent sentences rather than from stereotype-inconsistent sentences, whereas low-power participants seemingly did not draw STIs from stereotype-consistent or stereotypeinconsistent behaviors. Study 1 provided initial evidence for our hypothesis that the effects of stereotypes on STIs are moderated by perceivers’ power.

Study 2 Similar to age, gender is a basic social category (Weick & Guinote, 2008). Previous research has indicated that when gender-relevant cues are presented, individuals’ gender stereotypes are automatically activated (e.g., Most, Sorber, & Cunningham, 2007; Yan et al., 2012). In Study 2, we explored the effects of gender stereotypes on STIs and the moderating role of power. In Study 1, elderly stereotypes were primed by the semantic stimulus “the elder.” By contrast, in Study 2, we primed participants’ gender stereotype using male/female photos.

(Yan et al., 2012). All of the photos were 2.5 3.6 cm in size. All faces in the photos had neutral expressions and clothing (Yan et al., 2012). Experimental Sentences Based on previous research (Yan et al., 2012), we selected four typical masculine traits (adventurous, aggressive, fearless, rude) and four typical feminine traits (tender, dexterous, shy, emotional). Correspondingly, we developed 8 trait-implying sentences (4 male stereotype-consistent sentences and 4 female stereotype-consistent sentences), each describing one specific behavior of one actor. The actor in each sentence was simply denoted by the letter “X” (e.g., the sentence “X hits the saleswoman without any apparent reason” implies the trait “aggressive”). In a pilot study, 30 participants were asked to indicate the extent to which each of 12 behaviors represented typical male and female behaviors on a 5-point scale (1 = not at all, 5 = extremely). The results showed that 4 male stereotype-consistent behaviors were rated as more typical for males (M = 4.44, SD = 0.36) than females (M = 2.12, SD = 0.68), t(29) = 16.10, p < .001, and that 4 female stereotypeconsistent behaviors were rated as more typical for females (M = 4.64, SD = 0.33) than for males (M = 2.32, SD = 0.70), t(29) = 15.24, p < .001. Control Sentences Corresponding to eight trait-implying sentences, eight control sentences were also developed. These sentences contained the key words included in the trait-implying sentences but did not imply the target traits (e.g., the control sentence that corresponded to the previous traitimplying example sentence was “X took the change from the salesman after the settlement of accounts”).

Method Participants and Design Seventy-two undergraduates from a public university volunteered to participate in this study (10 men, 62 women). The average age was 19.28 years (SD = 0.76). The experimental design was a 2 (Power: high vs. low) 3 (Sentence type: stereotype consistent vs. stereotype inconsistent vs. control) mixed design, with power as a between-subjects factor and sentence type as a withinsubjects factor.

Materials Photos Sixty-eight head-and-shoulder photos of Chinese undergraduates (36 males and 36 females) selected from previous research were used to prime gender stereotypes Social Psychology (2017), 48(1), 3–18

Filler Sentences As in Study 1, in this study, 16 filler sentences corresponding to the 16 experimental sentences were created to balance participants’ “Yes” and “No” responses. These filler sentences were similar in length to the experimental sentences, but they did not imply specific traits. Photo-Behavior Sentence Pairs Each of the 16 key sentences (8 experimental sentences and 8 control sentences) was randomly assigned to a male photo or a female photo. Thus, a total of 32 experimental photo-behavior pairs were created: 8 stereotype-consistent pairs (pairing male photos with masculine behaviors and female photos with feminine behaviors), 8 stereotypeinconsistent pairs (pairing male photos with feminine behaviors and female photos with masculine behaviors), and 16 control pairs (pairing male and female photos with Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

control sentences).2 Likewise, each of the 16 filler sentences was paired with one male face and one female face to create 32 filler pairs.

Table 2. Mean reaction times in milliseconds and error rates to trait probes as a function of sentence type and power priming in Study 2 Sentence type Power priming

Stereotypeconsistent

Procedure The procedure was similar to that of Study 1. Participants were asked to complete a scrambled sentence task and a memory task. The scrambled sentence was identical to that used in Study 1. The so-called memory task was also similar to that employed in Study 1, with one exception: Before each sentence was shown, a male/female photo was presented for 33 ms. Similar to the procedures of Study 1, participants completed four practice trials to gain familiarity with the procedure before the formal experiment started. Responses and RTs were recorded to the nearest millisecond by a computer. After completing the memory test, participants were checked for suspicion, debriefed, and thanked.

Results and Discussion Reaction Times Three participants (1 male, 2 females) with high error rates (greater than 50%) and who may have failed to follow our instructions were excluded from all analyses. The inclusion of these participants did not change the pattern of the data or any critical inferential test. The analyses of the full data set can be found in the Electronic Supplementary Materials. Incorrect “yes” responses during the experimental trials were excluded from the RT analysis (2.08% of the total data). To reduce extreme variance, we regarded the RTs that were greater than 2.5 standard deviations above or below the mean of a particular item as outliers and excluded them from the RT analysis (3.07% of the total data). Participants’ RTs were analyzed using a 2 (Power: high vs. low) 3 (Sentence type: consistent vs. inconsistent vs. control) mixed-model ANOVA with sentence type as a within-subjects variable. Table 2 shows mean RTs in milliseconds and error rates as a function of sentence type and power priming. The ANOVA revealed a significant main effect of sentence type, F(2, 134) = 11.11, p < .001, η2p = .142. Participants were more likely to draw STIs from 2

Stereotypeinconsistent

Control

Reaction times High power

918 (188)

841 (183)

805 (133)

Low power

809 (172)

804 (166)

775 (113)

High power

.04 (0.07)

.02 (0.06)

.02 (0.03)

Low power

.02 (0.04)

Error rates

Note. Standard deviations are given in parentheses.

gender stereotype-consistent behaviors than from gender stereotype-inconsistent behaviors (M = 866 and 823, respectively, F(1, 68) = 5.09, p = .027). However, this main effect was qualified by a significant interaction between power and sentence type, F(2, 134) = 3.98, p = .027, η2p = .056. Simple main effects analyses showed that for the high-power primed participants, RTs varied as a function of the gender stereotype consistency of photo-behavior pairs, F(2, 70) = 12.04, p < .001, η2p = .256, whereas for the low-power primed participants, RTs did not vary across types of photo-behavior pairs, F(2, 64) = 1.64, p = .202, η2p = .049. Specific comparisons revealed that high-power participants required more time to correctly react to probes following gender stereotype-consistent sentences than those following control sentences, F(1, 35) = 28.35, p < .001, η2p = .448. However, the differences between the gender stereotype-inconsistent and control sentences were only marginally significant, F(1, 35) = 3.40, p = .074, η2p = .089. Specifically, high-power participants were inclined to draw STIs from gender stereotype-consistent sentences but not from gender stereotype-inconsistent sentences, whereas lower-power participants did not draw STIs from gender stereotype-consistent or stereotypeinconsistent behaviors. In other words, the effects of gender stereotypes on STIs were moderated by participants’ power.

Error Rates The error rates were also analyzed in a 2 (Power) 3 (Sentence type) mixed-design ANOVA. The results showed that the effects of power and sentence type were not significant, F(1, 67) = 1.60, p = .210, η2p = .023, and F(2, 134) = 1.26, p = .286, η2p = .018, respectively, and that

In contrast to Study 1, in which elderly stereotypes were activated via semantic words (the elder) embedded in each sentence, Study 2 used male/female photos to activate gender stereotypes and paired each sentence with a male or a female photo. To run the same number of experimental trials run in Study 1 (six stereotype-consistent trials, six stereotype-inconsistent trials, and 12 control trials), we would need to select three male traits and three female traits. However, we experienced difficulty in balancing the valences of the selected traits. Thus, we selected four male traits (two positive traits and two negative traits) and four female traits (two positive traits and two negative traits), resulting in 32 experimental trials (eight stereotype-consistent trials, eight stereotype-inconsistent trials, and 16 control trials).

Ó 2017 Hogrefe Publishing

Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

the interaction effect between power and sentence type was not significant, F(2, 134) = 2.07, p = .130, η2p = .030. Consistent with the findings of Study 1, the results of Study 2 demonstrated that the effects of gender stereotypes on STIs were moderated by perceivers’ power. That is, high-power individuals, but not low-power individuals, tended to draw STIs from stereotype-consistent behaviors but not from stereotype-inconsistent behaviors. Thus, the results of Study 1 and Study 2 consistently confirmed our prediction that the effects of stereotypes on STIs are moderated by perceivers’ power. In both studies, we manipulated participants’ power via conceptual priming. To further assess whether our findings can be generalized, we measured participants’ dispositional power and examined the moderating role of such power in Study 3.

translated from English to Chinese by a psychology Ph.D. student who was fluent in English. This version was then translated from Chinese to English by a different psychology Ph.D. student. The new English version was then compared with the original version; any discrepancies arising from the back-translation process were then adjusted. The generalized version of the Personal Sense of Power Scale asks participants to report their generalized beliefs about the power that they hold in their relationships with others. Participants rated their agreement with each of eight items, including “in my relationships with others, I think I have a great deal of power,” on a 7-point scale (1 = strongly disagree, 7 = strongly agree). As in previous research (Anderson & Galinsky, 2006), in the present study, the scale showed high internal consistency (α = .848, M = 4.95, SD = 0.73).

Study 3 In Study 3, we measured participants’ sense of power with the Personal Sense of Power Scale (Anderson, John, et al., 2012), which has been widely used in previous research (e.g., Anderson & Galinsky, 2006; Anderson, Kraus, Galinsky, & Keltner, 2012; Joshi & Fast, 2013).

Method Participants A total of 118 undergraduate students (41 males and 76 females, one participant’s sex information was lost, Mage = 19.52, SD = 0.71) who attended a university in China volunteered to participate in the current study.

Procedure and Materials Upon their arrival, participants were informed that they would complete a memory test and a questionnaire about interpersonal relationships. The so-called memory test was the same as that employed in Study 2. Participants needed to indicate as quickly and accurately as possible whether the probe was presented in the preceding sentence. All 64 trials were presented on the computer in random order. Following the memory test, participants completed the Personal Sense of Power Scale. Finally, participants were checked for suspicion, debriefed, and thanked. The Personal Sense of Power Scale We measured participants’ sense of power with the generalized version of the Personal Sense of Power Scale (Anderson, John, et al., 2012). Following the backtranslation procedure (Brislin, 1980), the scale was first Social Psychology (2017), 48(1), 3–18

Results and Discussion Preliminary Analyses Four participants (2 males and 2 females) who may have failed to follow our instructions were excluded from the analysis because of their excessive error rates (greater than 50%); the inclusion of these individuals did not change the pattern of the data or any critical inferential test. The data of the remaining 114 participants were further analyzed. Participants’ sense of power scores was calculated by averaging across the items (items 2, 4, 6, and 7 were reverse scored), with higher scores indicating a higher sense of power. Their scores were sorted in descending order. To test whether the effects of stereotypes on STIs differed between powerful and powerless individuals, we selected individuals with higher scores (at the top end of the distribution) as powerful participants and individuals with lower scores (at the bottom end of the distribution) as powerless participants. A power analysis (80% power, a large effect size, two-tailed, at the 0.05 significance level; Cohen, 1988) indicated that a sample size of at least 40 participants was sufficient for the present data analysis. Hence, the top 20% of participants (n = 23, 9 males and 14 females) were selected as high-power participants and the bottom 20% of participants (n = 23, 6 males and 17 females) were selected as low-power participants (similar to the approach used to select participants in Zou, Hudson, & Rapee, 2007).

Reaction Times As in Studies 1 and 2, in Study 3, incorrect “yes” responses (4.10%) and outliers (greater than 2.5 standard deviations Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

150

Table 3. Mean reaction times in milliseconds and error rates to trait probes as a function of sentence type and sense of power in Study 3

Stereotype-consistent behaviors

Sentence type Stereotypeconsistent

Stereotypeinconsistent

Stereotype-inconsistent behaviors

100

Control

Reaction times High power

799 (167)

742 (155)

737 (118)

Low power

802 (151)

817 (185)

780 (150)

High power

.05 (0.16)

.04 (0.13)

.04 (0.16)

Low power

.06 (0.11)

.06 (0.10)

.02 (0.04)

Error rates

The magnitude of STIs

Power sense

Note. Standard deviations are given in parentheses. -50

above or below the mean for a particular item, 2.08%) were excluded from the RT analysis. The RTs were analyzed in a 2 (Power: high vs. low) 3 (Sentence type: consistent vs. inconsistent vs. control) mixed-model ANOVA with repeated measures on the sentence type. Table 3 shows mean RTs in milliseconds and error rates as a function of sentence type and sense of power. The ANOVA revealed a main effect of sentence type, F(2, 88) = 4.30, p = .017, η2p = .089, which was qualified by a significant interaction between power and sentence type, F(2, 88) = 3.17, p = .047, η2p = .067. Simple effects analyses of the interaction showed that the RTs to the probes varied according to sentence type for high-power participants, F(2, 44) = 4.91, p = .012, η2p = .182, while the RTs did not significantly differ between different types of sentences for low-power participants, F(2, 44) = 2.08, p = .138, η2p = .086. As we predicted, specific comparisons revealed that high-power participants spent more time correctly reacting to the probes following the gender-stereotype-consistent sentences than those following the control sentences, F(1, 22) = 7.27, p = .013, η2p = .248, whereas no significant differences were found between gender-inconsistent sentences and control sentences, F(1, 22) = 0.08, p = .784, η2p = .004.

Error Rates We submitted participants’ error rates to a 2 (Power) 3 (Sentence type) mixed-model ANOVA. The results showed that the main effects of power and sentence type were not significant, F(1, 44) = 0.01, p = .921, η2p < .001, and F(2, 88) = 1.05, p = .353, η2p = .023, respectively, and that the interaction effect between power and sentence type was not significant, F(2, 88) = 0.26, p = .770, η2p = .006.

Supplemental Analyses We also computed two regression equations (including all 118 participants) to provide additional evidence for the moderating role of power in stereotype effects on STIs. Ó 2017 Hogrefe Publishing

Sense of power

Figure 2. The magnitude of STIs as a function of sense of power and sentence type (Study 3). The magnitude of STIs was calculated by subtracting the RTs for control behaviors from the RTs for stereotypeconsistent/-inconsistent behaviors.

In these regressions, the independent variable was participants’ sense of power and the dependent variable was “the magnitude of STIs” for stereotype-consistent behaviors (or for stereotype-inconsistent behaviors). The dependent variable was created by subtracting the RTs for control behaviors from the RTs for stereotype-consistent behaviors (or from stereotype-inconsistent behaviors; for a similar approach, see Zhang & Wang, 2013; also see Newman, 1991). The above ANOVA for RTs showed that high-power participants, but not low-power participants, made STIs from stereotype-consistent behaviors, whereas neither high-power nor low-power participants made STIs from stereotype-inconsistent behaviors. Based on this finding, we predicted that power was positively correlated with the magnitude of STIs for stereotype-consistent behaviors but that no significant correlation existed between power and the magnitude of STIs for stereotype-inconsistent behaviors. Consistent with our prediction, as shown in Figure 2, regression analyses showed that power positively predicted the magnitude of STIs for stereotype-consistent behaviors (β = 0.19, t = 2.08, p = .040) but did not predict the magnitude of STIs for stereotype-inconsistent behaviors (β = 0.08, t = 0.86, p = .394). Measuring participants’ dispositional power with the Personal Sense of Power Scale, Study 3 replicated the findings of Study 2 and found that participants’ power moderated the effects of gender stereotypes on STI formation. Specifically, high-power individuals drew STIs from stereotype-consistent behaviors but not from stereotypeinconsistent behaviors, while low-power individuals did not draw STIs from stereotype-consistent or stereotypeinconsistent behaviors. Thus, the three studies provide Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

convergent evidence that perceivers’ power moderates the effects of stereotypes on STIs.

occurrence of STIs. Finally, participants were checked for suspicion, debriefed, and thanked.

Study 4

Results and Discussion

In the previous three studies, we consistently found that stereotypes affected STI formation only in the high-power condition and not in the low-power condition, supporting the notion that possessing power promotes a reliance on stereotypes when perceiving others (Fiske, 1993; Keltner et al., 2003). However, due to the lack of a baseline condition (no-power manipulation), we cannot determine whether high levels of power promoted stereotype effects on STI formation, low levels of power inhibited stereotype effects on STI formation, or both. To address this issue, we added a baseline condition to further clarify the direction of the effects of power on the relationship between stereotypes and STI in Study 4. We predicted that compared with the baseline condition, the high-power condition would exacerbate the effects of stereotypes on STIs and the low-power condition would reduce the effects of stereotypes on STIs. In concrete terms, compared with the baseline, the high-power condition would promote STIs from stereotype-consistent behaviors and the low-power condition would inhibit STIs from stereotype-consistent behaviors. As in the previous three studies, in this study, we predicted that participants in the three conditions would not make STIs from stereotype-inconsistent behaviors.

Reaction Times

Participants and Design Ninety-three undergraduate students volunteered to participate in this study. The experiment employed a 3 (Power: high vs. low vs. no power) 3 (Sentence type: stereotype-consistent vs. stereotype-inconsistent vs. control) factorial design with repeated measures on the last factor. Participants were randomly assigned to the power conditions.

Materials and Procedure As in Study 2, in this study, participants were told that they would engage in two different tasks: a scrambled sentence task and a so-called memory task. The scrambled sentence task comprised the experimental manipulation of power. Participants in the high-power and low-power conditions completed the items utilized in Study 2. Participants in the baseline condition completed 16 items that contained only power-irrelevant words. Immediately after performing the scramble sentence task, participants completed the so-called memory task, which was used to detect the Social Psychology (2017), 48(1), 3–18

Four participants (1 male, 3 females) who may have failed to follow our instructions were excluded from all analyses because of their excessive error rates (greater than 50%). The inclusions of these participants did not change the pattern of the data or any critical inferential test. The analyses of the full data set are presented in the Electronic Supplementary Materials. Incorrect “yes” responses and extreme values (greater than 2.5 standard deviations above or below the mean) were also excluded from the RT analysis. In total, 5.16% of the RTs were dropped from data analysis. The RTs were submitted to a 3 (Power: high vs. low vs. no power) 3 (Sentence type: consistent vs. inconsistent vs. control) ANOVA with repeated measures on the last factor. Table 4 shows mean RTs in milliseconds and error rates as a function of sentence type and power priming. As predicted, the AVONA yielded a significant Power Sentence type interaction, F(4, 172) = 2.85, p = .026, η2p = .062. The main effects of power and sentence type were not significant, F(2, 86) = 0.30, p = .742, η2p = .007, and F(2, 172) = 2.23, p = .110, η2p = .025, respectively. Simple main effects analyses of the interaction between power and sentence type revealed a significant effect of sentence type for the high-power primed participants, F(2, 56) = 4.65, p = .013, η2p = .243, and a marginally significant effect of sentence type for the no-power primed participants, F(2, 60) = 2.51, p = .090, η2p = .077. However, for the low-power primed participants, RTs did not vary as a function of sentences type, F(2, 56) = 1.19, p = .311, η2p = .041. Specific comparisons found that participants in both the high-power priming and the baseline conditions took more time to correctly reject probes following stereotype-consistent sentences than those following control sentences, F(1, 28) = 10.60, p = .003, η2p = .275, and F(1, 30) = 6.23, p = .018, η2p = .172, respectively. Nevertheless, no significant differences were found between stereotype-inconsistent sentences and control sentences in the high-power priming or the baseline conditions, F(1, 28) = 0.69, p = .414, η2p = .024, and F(1, 30) = 2.81, p = .104, η2p = .086, respectively. That is to say, participants in the high-power priming and baseline conditions made STIs from stereotype-consistent behaviors but not from stereotype-inconsistent behaviors, whereas participants in the low-power priming condition did not make STIs from either stereotype-consistent or Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Table 4. Mean reaction times in milliseconds and error rates to trait probes as a function of sentence type and power priming in Study 4 Sentence type Power priming

Stereotypeconsistent

Stereotypeinconsistent

Control

Reaction times High power

869 (188)

828 (202)

810 (136)

Low power

787 (169)

815 (178)

817 (177)

Baseline

826 (169)

817 (210)

780 (156)

Error rates High power

.06 (0.16)

.08 (0.19)

.02 (0.04)

Low power

.02 (0.04)

.01 (0.05)

.002 (0.01)

Baseline

.02 (0.04)

.04 (0.06)

.01 (0.01)

stereotype-inconsistent behaviors, F(1, 88) = 2.02, p = .158, η2p = .022. No other effects were found. In Study 4, in addition to applying the high-power and low-power priming conditions, we added the baseline condition to explore the possible facilitation or inhibition effects of power on STIs from stereotype-consistent behaviors. As we predicted, the low-power prime inhibited STIs from stereotype-consistent behaviors. However, inconsistent with our prediction, the high-power prime did not promote STIs from stereotype-consistent behaviors. Potential explanations for this unexpected finding are presented in the General Discussion.

Note. Standard deviations are given in parentheses.

General Discussion stereotype-inconsistent behaviors. In short, STIs from stereotype-consistent behaviors emerged only among the high-power and no-power primed participants and the low-power primed participants. In this sense, low-power priming actually inhibited STIs from stereotype-consistent behaviors. Given that participants in both the high-power priming and baseline conditions made STIs from gender stereotype-consistent behaviors, we further examined whether the high-power prime increased the magnitude of STIs compared with the baseline condition. As done in Study 3, a “STI strength score” was calculated for each participant by subtracting the RTs for control sentences from the RTs for stereotype-consistent sentences. An independent t test showed no significant differences in the magnitude of STIs between the high-power priming and the baseline conditions (M = 59 and 45, respectively), t(58) = 0.54, p = .590, d = 0.14. That is to say, the high-power prime did not facilitate STIs compared with the baseline condition.

Error Rates A 3 (Power) 3 (Sentence type) ANOVA was performed on the mean error rates, with repeated measures on the last factor. This data analysis revealed only a significant main effect of sentence type, F(2, 172) = 6.97, p = .001, η2p = .075. Post hoc tests showed that both the error rates for stereotype-consistent behaviors (M = 0.03, SD = 0.10) and stereotype-inconsistent behaviors (M = 0.04, SD = 0.02) were significantly higher than the error rates for control sentences (M = 0.01, SD = 0.02), F(1, 88) = 6.91, p = .01, η2p = .073; and F(1, 88) = 8.94, p = .004, η2p = .092, respectively. There were no significant differences in error rates between stereotype-consistent and

Ó 2017 Hogrefe Publishing

The current research examined the role of power in stereotype effects on STIs. Across four studies, we consistently found that power moderated the effects of stereotypes on STIs. High-power individuals tended to draw STIs from stereotype-consistent behaviors rather than from stereotype-inconsistent behaviors, whereas low-power individuals did not draw STIs from stereotype-consistent or stereotype-inconsistent behaviors. As mentioned, Guinote and colleagues (2010) suggested that power promotes individuals’ stereotyping of others even when they do not intend to form an impression. However, the present research extended beyond previous research by first demonstrating that power promotes a reliance on stereotypes in the formation of spontaneous trait impressions, which are predominant in the formation of spontaneous impressions (Schneid, Carlston, & Skowronski, 2015).

Stereotype Effects on STIs Past research has found that perceivers are more likely to draw STIs from stereotype-consistent behaviors than from stereotype-inconsistent behaviors (Wigboldus et al., 2003, 2004; Yan et al., 2012). Wigboldus and colleagues (2003) proposed a potential mechanism underlying the effects of stereotypes on STIs. According to this proposition, in the present research, stereotypes would be automatically activated when stereotype-relevant cues were presented. The activated stereotypes, in turn, would enhance the accessibility of stereotype-consistent traits and reduce the accessibility of stereotype-inconsistent traits (Dijksterhuis & van Knippenberg, 1996). As a result, participants would more easily draw accessible stereotype-consistent traits from stereotype-consistent behaviors than less accessible stereotype-inconsistent traits from stereotype-inconsistent

Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

behaviors. Correspondingly, participants were more likely to draw STIs from stereotype-consistent behaviors than from stereotype-inconsistent behaviors.

Power Moderates Stereotype Effects on STIs Extending previous work on stereotypes and STIs, the present research further specified the conditions under which STIs are susceptible to the influences of stereotypes. Across the four studies, we found that effects of stereotypes on STIs among powerful participants but not among powerless participants, demonstrating the moderating role of perceivers’ power in stereotype effects on STIs. The present research did not directly explore the mechanism underlying the moderating effects of power. However, based on prior literature (Rim, Uleman, & Trope, 2009; Smith & Trope, 2006), we speculated that the construal level may account for the moderating effects of power in stereotype effects on STIs. The theory holds that powerful individuals construe others on a relatively high level due to their distal psychological distance to others, while powerless individuals construe others on a relatively low level due to their proximal psychological distance to others (Magee & Smith, 2013; Smith & Trope, 2006). Prior research has indicated that individuals with a highconstrual level rely more on categorical information when forming impressions, such as traits and stereotypes, than those with a low-construal level (McCrea, Wieber, & Myers, 2012; Rim et al., 2009). For instance, priming construal level by asking participants to consider increasingly highorder superordinate goals or increasingly lower-level behaviors that could fulfill a particular goal, McCrea and colleagues (2012) found that participants were more likely to make stereotypic judgments about a person when they were placed in a high-construal mindset than in a low-construal mindset. With regard to the present research, we speculated that high levels of power may elicit a highconstrual level and that low levels of power may elicit a low-construal level. As a result, stereotypes were more likely to be activated under the high-power condition than under the low-power condition even though the same stereotypic cues were presented under both conditions. Accordingly, the effects of stereotypes on STIs were observed among powerful participants but not among powerless participants. That is, under the high-power condition, participants tended to make STIs from stereotype-consistent behaviors but not from stereotypeinconsistent behaviors; under the low-power condition, participants did not make STIs from stereotype-consistent or stereotype-inconsistent behaviors. According to the situated focus theory of power (Guinote, 2010, 2013), the moderating effects of power Social Psychology (2017), 48(1), 3–18

on the relationship between stereotypes and STIs may be flexible and malleable. The theory holds that power promotes a reliance on stereotypes only when stereotypes are accessible in a given situation. When stereotypes are less accessible, powerful people rely less on stereotypes than powerless people. In a recent study conducted by Weick and Guinote (2008), compared with powerless participants, powerful participants exhibited more stereotypic perceptions of men and women when stereotypes were accessible (in conditions in which participants could easily retrieve stereotypic information). However, when stereotypes were less accessible (in conditions in which participants experienced difficulty in retrieving stereotypic information), powerful participants displayed less stereotypic perceptions of the gender groups compared with powerless participants. In our research, we presented stereotypic cues in each behavioral sentence or before each behavioral sentence was presented; thus, stereotypic cues were highly accessible. According to the situated focus theory of power (Guinote, 2010, 2013), if we reduce the accessibility of stereotypic cues, high-power conditions should reduce individuals’ reliance on stereotypes in STIs compared with low-power conditions. For example, considering that the presentation of counterstereotypic information before stereotypic cues can reduce the temporary accessibility of stereotypes (Dasgupta & Asgari, 2004; Huntsinger, Sinclair, Dunn, & Clore, 2010), high-power conditions may not promote stereotype effects on STIs compared with low-power conditions when counter-stereotypic information is presented in advance. This speculation should be tested in future research on STIs.

The Direction of the Moderating Effects of Power In Study 4, we added a baseline condition (no-power manipulation) to clarify the direction of the moderating effect of power in the relationship between stereotypes and STIs. As expected, we found that participants in the baseline condition made STIs from stereotype-consistent behaviors, while participants in the low-power priming condition did not make such inferences, indicating that low-power priming actually inhibited STIs from stereotype-consistent behaviors. Although we found that participants in both the high-power priming and baseline conditions made STIs from stereotype-consistent behaviors, the magnitude of STIs did not differ between the groups. This result indicates that high-power priming did not promote STIs from stereotype-consistent behaviors compared with the baseline condition and is inconsistent with our expectations. In the low-power priming condition, participants did not draw STIs from stereotype-consistent behaviors. This result Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

may be partially explained by the social distance theory of power (Magee & Smith, 2013; Smith & Trope, 2006). As mentioned above, low-power individuals tend to construct others’ behaviors in a concrete, item-level manner due to their psychologically proximal distance to others (Smith & Trope, 2006). The low level construal of behaviors, in turn, inhibits the occurrence of STIs (Rim, Trope, Liberman, & Shapira, 2013; Rim et al., 2009). Rim and colleagues’ (2009) recent study that primed participants’ high/ low-construal level through a word generation task (asking participants to generate a series of category words or example words) found that high-construal level priming inhibited STI formation compared with low-construal level priming. Thus, in our research, low-power priming may elicit low level construal for behaviors, thereby inhibiting STIs. In addition, the unique culture of power in China may exacerbate the inhibition effects of low power on STI formation. For over 2,000 years, Chinese societies have followed Confucianism and, thus, have emphasized hierarchy and submission to authority (Farh & Cheng, 2000; Cheng, Chou, Wu, Huang, & Farh, 2004; Farh, Hackett, & Liang, 2007). This focus differs from most Western countries’ emphasis on equity (Triandis & Gelfand, 1998). As a result, compared with low-power individuals in Western societies, low-power individuals in Chinese societies may be more likely to suffer from constraints and threats from the external environment, especially from high-power individuals (e.g., Ma & Han, 2009). Accordingly, compared with powerless Westerners, powerless individuals in Chinese societies may be more likely to interpret others’ behaviors in a concrete and detailed manner to avoid possible threats and punishments. Thus, the formation of STIs among these individuals in Chinese societies may be inhibited to a greater extent than that among individuals in Western societies, resulting in powerless participants’ lack of STIs. We speculated that we did not find the expected facilitation effects under the high-power priming condition due to ceiling effects. In concrete terms, previous research has indicated that a sense of control is a basic need for human beings and that most people have control over the surrounding context in most situations (Whitson & Galinsky, 2008). Therefore, in the baseline condition, participants may actually have a sense of control even though power was not manipulated. As a consequence, gender stereotypes may be sufficiently activated when gender-relevant cues are presented in the baseline condition, which may lead to less room to enhance the accessibility of stereotype-consistent traits in the highpower priming condition compared with the baseline condition. Accordingly, the finding that stereotype-consistent behaviors did not facilitate STIs to a greater extent in the high-power condition than in the baseline condition is not Ó 2017 Hogrefe Publishing

surprising. To some extent, this finding provides a potential explanation for why most previous experiments that have not manipulated power have found STI effects and stereotype effects – most people do not feel powerless most of the time; rather, they have a sense of control. One finding warrants further consideration. Specifically, participants seemingly did not draw STIs from stereotypeinconsistent behaviors across the four studies. Prior research has indicated that, once activated, stereotypes not only enhance the accessibility of stereotype-consistent traits but also reduce the accessibility of stereotypeinconsistent traits (Dijksterhuis & van Knippenberg, 1996). In the latter case, inferring a trait that is inconsistent with an activated stereotype may be difficult because the trait that must be inferred is temporarily inaccessible (Wigboldus et al., 2003). Correspondingly, drawing STIs from stereotype-inconsistent behaviors should be difficult. Of note, in previous research (e.g., Ramos, Garcia-Marques, Hamilton, Ferreira, & Acker, 2012; Wigboldus et al., 2003), participants formed STIs from stereotype-inconsistent behaviors even though stereotypes reduced the strength of STIs from such behaviors. Given these results, the lack of STI formation from stereotype-inconsistent behaviors is somewhat surprising. However, to our knowledge, most research concerning STIs has been conducted in individualistic cultures, whereas the present research was conducted in collectivistic cultures. Prior literature has suggested that unlike individuals in individualistic cultures, individuals in collectivistic cultures place greater emphasis on situational constraints than on internal dispositions when interpreting social behaviors (Duff & Newman, 1997; Na & Kitayama, 2011). As a consequence, the tendency to draw STIs from trait-implying behaviors may be relatively weak for individuals in collectivistic cultures compared with individuals in individualistic cultures. In this sense, our finding that participants did not make STIs from stereotype-inconsistent behaviors may be culture specific.

Limitations and Future Directions There are several limitations to our research. First, STIs are generally assumed to involve two processes: (a) spontaneously inferring the corresponding trait from behaviors (trait activation) and (b) ascribing the trait to the actor (trait bonding). Theoretically, the probe recognition paradigm, which was used to detect STIs in the present research, may capture only the first stage of STI formation (trait activation). Thus, future research should utilize multiple research paradigms to capture the cumulative effects of trait activation and trait bonding, such as the false recognition paradigm and the relearning paradigm (Carlston & Skowronski, 1994; Todorov & Uleman, 2002). Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Second, all participants involved in the research are undergraduates, who have little experience with power (Guinote & Phillips, 2010), potentially limiting the generalizability of our findings. In future research, we can manipulate power by assigning undergraduates to high-/ low-power roles in the laboratory (e.g., Rucker, Hu, & Galinsky, 2014) or by selecting power holders and subordinates in real contexts of power (e.g., Prooijen, Coffeng, & Vermeer, 2014). Despite these limitations, this study is the first to explore how power influences stereotype effects on STIs. We found that power moderated the effects of stereotypes on STIs. Stereotype effects on STIs were found among high-power individuals but not among low-power individuals. We connected research on power to research on STIs and provided new evidence for the malleability of the effects of stereotypes on STIs.

Acknowledgments This research was supported by a grant from the National Natural Science Foundation of China (31271104). Electronic Supplementary Material The electronic supplementary material is available with the online version of the article at http://dx.doi.org/10.1027/ 1864-9335/a000288 ESM 1. Tables and Text (PDF). Experimental, control sentences, and the corresponding probes in Studies 1–4. Further explanations and analyses.

References Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risktaking. European Journal of Social Psychology, 36, 511–536. doi: 10.1002/ejsp.324 Anderson, C., John, O. P., & Keltner, D. (2012). The personal sense of power. Journal of Personality, 80, 313–344. doi: 10.1111/ j.1467-6494.2011.00734.x Anderson, C., Kraus, M. W., Galinsky, A. D., & Keltner, D. (2012). The local-ladder effect social status and subjective wellbeing. Psychological Science, 23, 764–771. doi: 10.1177/ 0956797611434537 Bargh, J. A. (1999). The cognitive monster: The case against the controllability of automatic stereotype effects. In S. Chaiken & Y. Trope (Eds.), Dual-process theories in social psychology (pp. 361–382). New York, NY: Guilford Press. Blair, I. V. (2002). The malleability of automatic stereotypes and prejudice. Personality and Social Psychology Review, 6, 242–261. doi: 10.1207/S15327957PSPR0603_8 Brislin, R. W. (1980). Translation and content analysis of oral and written material. In H. C. Triandis & J. W. Berry (Eds.), Handbook of cross-cultural psychology (Vol. 2, pp. 389–444). Boston, MD: Allyn & Bacon. Carlston, D. E., & Skowronski, J. J. (1994). Saving in the relearning of trait information as evidence for spontaneous inference

Social Psychology (2017), 48(1), 3–18

generation. Journal of Personality and Social Psychology, 66, 840–856. doi: 10.1037/0022-3514.66.5.840 Cheng, B. S., Chou, L. F., Wu, T. Y., Huang, M. P., & Farh, J. L. (2004). Paternalistic leadership and subordinate responses: Establishing a leadership model in Chinese organizations. Asian Journal of Social Psychology, 7, 89–117. doi: 10.1111/ j.1467-839X.2004.00137.x Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum. Dasgupta, N., & Asgari, S. (2004). Seeing is believing: Exposure to counterstereotypic women leaders and its effect on the malleability of automatic gender stereotyping. Journal of Experimental Social Psychology, 40, 642–658. doi: 10.1016/ j.jesp.2004.02.003 Dijksterhuis, A., & van Knippenberg, A. D. (1996). The knife that cuts both ways: Facilitated and inhibited access to traits as a result of stereotype activation. Journal of Experimental Social Psychology, 32, 271–288. doi: 10.1006/jesp.1996.0013 Domhoff, G. W. (1998). Who rules America. Mountain View, CA: Mayfield Publishing. Duff, K. J., & Newman, L. S. (1997). Individual differences in the spontaneous construal of behavior: Idiocentrism and the automatization of the trait inference process. Social Cognition, 15, 217–241. doi: 10.1521/soco.1997.15.3.217 Emerson, R. M. (1962). Power dependence relations. American Sociological Review, 27, 31–41. doi: 10.2307/2089716 Farh, J. L., & Cheng, B. S. (2000). A cultural analysis of paternalistic leadership in Chinese organizations. In J. T. Li, A. S. Tsui, & E. Weldon (Eds.), Management and organizations in the Chinese context (pp. 357–411). London, UK: MacMillan. Farh, J. L., Hackett, R. D., & Liang, J. (2007). Individual-level cultural values as moderators of perceived organizational support–employee outcome relationships in China: Comparing the effects of power distance and traditionality. Academy of Management Journal, 50, 715–729. doi: 10.5465/AMJ.2007. 25530866 Fiske, S. T. (1993). Controlling other people. The American Psychologist, 48, 621–628. doi: 10.1037/0003-066X.48.6.621 Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (pp. 357–411). New York, NY: McGraw-Hill. Fiske, S. T., & Dépret, E. (1996). Control interdependence and power: Understanding social cognition in its social context. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (Vol. 7, pp. 31–61). New York, NY: Wiley. Förster, J., & Liberman, N. (2007). Knowledge activation. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (2nd ed., pp. 201–231). New York, NY: Guilford. French, J., & Raven, B. (1959). The bases of social power. In D. Cartwright (Ed.), Studies in social power (pp. 150–167). Ann Arbor, MI: Institute for Social Research. Goodwin, S. A., Gubin, A., Fiske, S. T., & Yzerbyt, V. (2000). Power can bias impression formation: Stereotyping subordinates by default and by design. Group Processes and Intergroup relations, 3, 227–256. doi: 10.1177/1368430200003003001 Guinote, A. (2007a). Behavior variability and the situated focus theory of power. In W. Stroebe & M. Hewstone (Eds.), European review of social psychology (pp. 256–295). New York, NY: Wiley. Guinote, A. (2007b). Power and the suppression of unwanted thoughts: Does control over others decrease control over the self? Journal of Experimental Social Psychology, 43, 433–440. doi: 10.1016/j.jesp.2006.03.003 Guinote, A. (2010). The situated focus theory of power. In A. Guinote & T. K. Vescio (Eds.), The social psychology of power (pp. 141–173). New York, NY: Guilford Press.

Ó 2017 Hogrefe Publishing

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Guinote, A. (2013). Social power and cognition. In D. E. Carlston (Ed.), Oxford handbook of social cognition. New York, NY: Oxford University Press. Guinote, A., & Phillips, A. (2010). Power can increase stereotyping: Evidence from managers and subordinates in the hotel industry. Social Psychology, 41, 3–9. doi: 10.1027/1864-9335/ a000002 Guinote, A., Willis, G. B., & Martellotta, C. (2010). Social power increases implicit prejudice. Journal of Experimental Social Psychology, 46, 299–307. doi: 10.1016/j.jesp.2009.11.012 Ham, J., & Vonk, R. (2003). Smart and easy: Co-occurring activation of spontaneous trait inferences and spontaneous situational inferences. Journal of Experimental Social Psychology, 39, 434–447. doi: 10.1016/S0022-1031(03)00033-7 Higgins, E. T. (2012). Accessibility Theory. In P. A. M. Van Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 11–31). Thousand Oaks, CA: SAGE. Hilton, J. L., & Von Hippel, W. (1996). Stereotypes. Annual Review of Psychology, 47, 237–271. doi: 10.1146/annurev.psych. 47.1.237 Huntsinger, J. R., Sinclair, S., Dunn, E., & Clore, G. L. (2010). Affective regulation of stereotype activation: It’s the (accessible) thought that counts. Personality and Social Psychology Bulletin, 36, 564–577. doi: 10.1177/0146167210363404 Joshi, P. D., & Fast, N. J. (2013). Power and reduced temporal discounting. Psychological Science, 24, 432–438. doi: 10.1177/ 0956797612457950 Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach, and inhibition. Psychological Review, 110, 265–284. doi: 10.1037/0033-295X.110.2.265 Kunda, Z., & Spencer, S. J. (2003). When do stereotypes come to mind and when do they color judgment? A goal-based theoretical framework for stereotype activation and application. Psychological Bulletin, 129, 522–544. doi: 10.1037/00332909.129.4.522 Ma, Y., & Han, S. (2009). Self-face advantage is modulated by social threat-boss effect on self-face recognition. Journal of Experimental Social Psychology, 45, 1048–1051. doi: 10.1016/ j.jesp.2009.05.008 McKoon, G., & Ratcliff, R. (1986). Inferences about predictable events. Journal of Personality and Social Psychology: Learning, Memory and Cognition, 15, 326–338. doi: 10.1037/02787393.12.1.82 Magee, J. C., & Smith, P. K. (2013). The social distance theory of power. Personality and Social Psychology Review, 17, 158–186. doi: 10.1177/1088868312472732 McCrea, S. M., Wieber, F., & Myers, A. L. (2012). Construal level mind-sets moderate self- and social stereotyping. Journal of Personality & Social Psychology, 102, 51–68. doi: 10.1037/ a0026108 Most, S. B., Sorber, A. V., & Cunningham, J. G. (2007). Auditory Stroop reveals implicit gender associations in adults and children. Journal of Experimental Social Psychology, 43, 287–294. doi: 10.1016/j.jesp.2006.02.002 Na, J., & Kitayama, S. (2011). Spontaneous trait inference is culture-specific: Behavioral and neural evidence. Psychological Science, 22, 1025–1032. doi: 10.1177/0956797611414727 Newman, L. S. (1991). Why are traits inferred spontaneously? A developmental approach. Social Cognition, 9, 221–253. doi: 10.1521/soco.1991.9.3.221 Perdue, C. W., & Gurtman, M. B. (1990). Evidence for the automaticity of ageism. Journal of Experimental Social Psychology, 26, 199–216. doi: 10.1016/0022-1031(90)90035-K Prooijen, J. W. V., Coffeng, J., & Vermeer, M. (2014). Power and retributive justice: How trait information influences the fairness

Ó 2017 Hogrefe Publishing

of punishment among power holders. Journal of Experimental Social Psychology, 50, 190–201. doi: 10.1016/j.jesp.2013. 10.004 Ramos, T., Garcia-Marques, L., Hamilton, D. L., Ferreira, M., & Acker, K. V. (2012). What I infer depends on who you are: The influence of stereotypes on trait and situational spontaneous inferences. Journal of Experimental Social Psychology, 48, 1247–1256. doi: 10.1016/j.jesp.2012.05.009 Rim, S., Trope, Y., Liberman, N., & Shapira, O. (2013). The highs and lows of mental representation: A construal level perspective on the structure of knowledge. In D. E. Carlston (Ed.), Oxford handbook of social cognition. New York, NY: Oxford University Press. Rim, S., Uleman, J. S., & Trope, Y. (2009). Spontaneous trait inference and construal level theory: Psychological distance increases nonconscious trait thinking. Journal of Experimental Social Psychology, 45, 1088–1097. doi: 10.1016/j.jesp.2009. 06.015 Rucker, D. D., Hu, M., & Galinsky, A. D. (2014). The experience versus the expectations of power: A recipe for altering the effects of power on behavior. Journal of Consumer Research, 41, 381–396. doi: 10.1086/676598 Rudman, L. A., & Borgida, E. (1995). The afterglow of construct accessibility: The behavioral consequences of priming men to view women as sexual objects. Journal of Experimental Social Psychology, 31, 493–517. doi: 10.1006/jesp.1995.1022 Schneid, E. D., Carlston, D. E., & Skowronski, J. J. (2015). Spontaneous evaluative inferences and their relationship to spontaneous trait inferences. Journal of Personality and Social Psychology, 108, 681–696. doi: 10.1037/a0039118 Shoda, Y., & Mischel, W. (1993). Cognitive social approach to dispositional inferences: What if the perceiver is a cognitive social theorist? Personality and Social Psychology Bulletin, 19, 574–585. doi: 10.1177/0146167293195009 Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90, 578–596. doi: 10.1037/0022-3514.90.4.578 Todorov, A., & Uleman, J. S. (2002). Spontaneous trait inferences are bound to actors’ faces: Evidence from a false recognition paradigm. Journal of Personality and Social Psychology, 83, 1051–1065. doi: 10.1037/0022-3514.83.5.1051 Triandis, H. C., & Gelfand, M. J. (1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality and Social Psychology, 74, 118–128. doi: 10.1037/0022-3514.74.1.118 Uleman, J. S., Hon, A., Roman, R., & Moskowitz, G. B. (1996). On-line evidence for spontaneous trait inferences at encoding. Personality and Social Psychology Bulletin, 22, 377–394. doi: 10.1177/0146167296224005 Uleman, J. S., Newman, L. S., & Moskowitz, G. B. (1996). People as flexible interpreters: Evidence and issues from spontaneous trait inference. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 28, pp. 211–279). San Diego, CA: Academic Press. Wang, M., Xia, J., & Yang, F. (2015). Flexibility of spontaneous trait inferences: The interactive effects of mood and gender stereotypes. Social Cognition, 33, 345–358. doi: 10.1521/soco. 2015.33.4.1 Weber, M. (1947). The theory of social and economic organization (A. M. Henderson & T. Parsons, Trans.). New York: Oxford University Press. Weick, M., & Guinote, A. (2008). When subjective experiences matter: Power increases reliance on the ease of retrieval. Journal of Personality and Social Psychology, 94, 956–970. doi: 10.1037/0022-3514.94.6.956

Social Psychology (2017), 48(1), 3–18

M. Wang & F. Yang, Power, Stereotypes, and Spontaneous Trait Inferences

Whitson, J. A., & Galinsky, A. D. (2008). Lacking control increases illusory pattern perception. Science, 322(5898), 115–117. doi: 10.1126/science.1159845 Wigboldus, D. H. J., Dijksterhuis, A., & van Knippenberg, A. (2003). When stereotypes get in the way: Stereotypes obstruct stereotype-inconsistent trait inferences. Journal of Personality and Social Psychology, 84, 470–484. doi: 10.1037/0022-3514. 84.3.470 Wigboldus, D. H., Sherman, J. W., Franzese, H. L., & van Knippenberg, A. V. (2004). Capacity and comprehension: Spontaneous stereotyping under cognitive load. Social Cognition, 22, 292–309. doi: 10.1521/soco.22.3.292.35967 Winter, L., & Uleman, J. S. (1984). When are social judgments made? Evidence for the spontaneousness of trait inferences. Journal of Personality and Social Psychology, 47, 237–252. doi: 10.1037/0022-3514.47.2.237 Yan, X., Wang, M., & Zhang, Q. (2012). Effects of gender stereotypes on spontaneous trait inferences and the moderating role of gender schematicity: Evidence from Chinese undergraduates. Social Cognition, 30, 220–231. doi: 10.1521/soco. 2012.30.2.220

Social Psychology (2017), 48(1), 3–18

Zhang, Q., & Wang, M. (2013). The development of spontaneous trait inferences: Evidence from Chinese children. Psychological Reports, 112, 887–899. doi: 10.2466/21.07.PR0.112.3.887-899 Zou, J. B., Hudson, J. L., & Rapee, R. M. (2007). The effect of attentional focus on social anxiety. Behaviour Research & Therapy, 45, 2326–2333. doi: 10.1016/j.brat.2007.03.014 Received September 5, 2015 Revision received August 2, 2016 Accepted August 4, 2016 Published online March 15, 2017

Meifang Wang Department of Psychology Shandong Normal University No. 88 East Wenhua Road Jinan, 250014 PR China meifangw@hotmail.com or wangmf@sdnu.edu.cn

Ó 2017 Hogrefe Publishing

Original Article

Relational Utility Affects Self-Punishment in Direct and Indirect Reciprocity Situations Ruida Zhu,1,2 Tao Jin,1,2 Xueyi Shen,1,2 Shen Zhang,1,2 Xiaoqin Mai,3 and Chao Liu1,2 1

State Key Laboratory of Cognitive Neuroscience and Learning & IDG/McGovern Institute for Brain Research,

Beijing Normal University, PR China 2

Center for Collaboration and Innovation in Brain and Learning Sciences, Beijing Normal University, PR China

Department of Psychology, Renmin University of China, Beijing, PR China

Abstract: Previous studies of self-punishment focused on negative emotions and information transmission between wrongdoers and victims. We propose that self-punishment can be moderated by relational utility and can work not only in direct but also indirect reciprocity. In Studies 1 and 2, participants were more inclined to punish themselves when the victim could benefit the participants in future interactions than when the victim could not. In Study 3, participants were more inclined to punish themselves when the bystander could potentially offer lots of benefits to them in the future compared to when the bystander could only offer few or no benefits. These findings support our hypothesis, suggesting that wrongdoers strategically use self-punishment to pursue profits through repairing damaged relationships which are really conducive to achieve their personal goals. It helps us to understand self-punishment better in real life. Keywords: self-punishment, relational utility, direct reciprocity, indirect reciprocity

In human society, wrongdoers sometimes exert economic loss or physical damage to themselves after violating social norms (Bastian, Jetten, & Fasoli, 2011; Inbar, Pizarro, Gilovich, & Ariely, 2013; Nelissen, 2011; Nelissen & Zeelenberg, 2009; Tanaka, Yagi, Komiya, Mifune, & Ohtsubo, 2015; Watanabe & Ohtsubo, 2012). Such a phenomenon, referred to as self-punishment, attracts researchers’ interest as it seems to diminish wrongdoers’ own benefit and does not benefit anyone. Researchers have proposed two possible explanations. First, self-punishment may be driven by negative emotions, especially guilt, due to harming others (Nelissen, 2011; Nelissen & Zeelenberg, 2009; Tanaka et al., 2015; Watanabe & Ohtsubo, 2012). Studies have found that the guiltier the participants felt, the more severe punishment they inflicted on themselves; in turn, the more severe punishment the participants gave themselves, the more their guilt was relieved (Bastian et al., 2011; Inbar et al., 2013). Second, self-punishment may be a method for wrongdoers to express their remorse to their victims. If the option of direct compensation is unavailable, wrongdoers can choose to engage in self-punishment to express their remorse after committing a transgression, which is conductive to relationship maintenance (Nelissen & Zeelenberg, 2009). Nelissen (2011) further found that Ó 2017 Hogrefe Publishing

wrongdoers are more willing to engage in self-punishment in the presence of a victim than in the presence of a general audience. It implies self-punishment is used as a signal of remorse for victims specifically. Existing studies suggest that self-punishment is driven by the guilt of violating moral standards and is used to express remorse (Bastian et al., 2011; Nelissen, 2011; Nelissen & Zeelenberg, 2009; Watanabe & Ohtsubo, 2012), which implies that self-punishment is mainly motivated by moral and other-focused considerations (e.g., victims are harmed). However, self-punishment may also be motivated by some self-interested and self-focused considerations (e.g., the potential benefits that self-punishment may bring to wrongdoers). It is generally assumed that moral behaviors benefit people in the long run by maintaining reciprocal rewarding relationships (Baumard, André, & Sperber, 2013; Haidt, 2003; Trivers, 1971). However, this hypothesis is correct only if the long-term benefits of the moral behavior surpass its costs. Thus, it is necessary for people to adjust their responses according to the potential benefits. Supporting this hypothesis, recently Nelissen (2014) found that moral emotional guilt, which is closely related to moral behavior, is moderated by relational utility, the utility of others for the achievement of one’s personal aims through social Social Psychology (2017), 48(1), 19–27 DOI: 10.1027/1864-9335/a000291

interaction. This finding was replicated in Ohtsubo and Yagi (2015)’s research, which showed that the wrongdoers’ feelings of guilt following interpersonal transgressions increased with an increase of the victims’ relational utility. Because guilt serves as an emotional mechanism for protecting interpersonal relationships (Haidt, 2003), one may infer that with regard to long-term benefits, wrongdoers are more likely to repair a damaged relationship that could bring them more benefits. A wrongdoer’s guilt per se could not repair a relationship, while guilt-related moral behaviors such as self-punishment that signal remorse, could (Nelissen, 2011; Ohtsubo & Watanabe, 2009). Therefore, it is probable that relational utility affects guilt-related behavior such as self-punishment as well. So our first aim in this study is to test whether the effect of relational utility on the feelings of guilt could extend to self-punishment. Most previous research on relational utility focused on the interaction between wrongdoers and victims, and thus only studied the effect of direct reciprocity (Nelissen, 2014; Ohtsubo & Yagi, 2015). However, according to the indirect reciprocity theory, wrongdoers tend to display their benign intentions through specific behaviors to maintain a positive reputation in public, as bystanders also judge wrongdoers’ behavior, and many of them are willing to bear some costs to punish wrongdoers (Nowak & Sigmund, 1998; Fehr & Fischbacher, 2004). So wrongdoers may also take the relational utility of bystanders into consideration when making a decision about self-punishment in front of them. Our second aim in this study thus is to test whether relational utility influences wrongdoers’ self-punishment in both direct and indirect reciprocity contexts.

Overview of the Current Research In the present research, relational utility refers to the potential monetary benefits the victim or bystander could offer to the wrongdoer in the future. Self-punishment is defined as participants’ behavior of abandoning their own monetary benefits. In Studies 1 and 2, we tested whether relational utility affects self-punishment in direct reciprocity by manipulating participants’ future chance of reciprocating with the victim. The future chance of reciprocity increases the utility of a relationship as people could benefit from future cooperation (Trivers, 1971). In Study 3, we tested whether relational utility affects self-punishment in indirect reciprocity. Besides the future opportunity for reciprocity, we also manipulated the amount of benefits the bystander was able to offer. The relational utility increases with increases in the amount of benefits that the bystander could provide (Baumard et al., 2013). We predict that relational utility promotes self-punishment in both direct and indirect reciprocity situations. All studies were Social Psychology (2017), 48(1), 19–27

R. Zhu et al., Relational Utility Affects Self-Punishment

approved by the Institutional Review Board of Beijing Normal University.

Study 1 Methods Participants and Design Forty-three undergraduate students (23 females, Mage = 22.1 years, SDage = 2.1 years) participated in the study for payment. The study had a one-factor (Future opportunity for reciprocity: Future vs. No future), between-subject design. Procedure The participants came in groups and performed tasks on the computer alone in separate rooms. The participants in each group did not know each other before the experiment. Personality Measurements The Interpersonal Reactivity Index (IRI) scale (Davis, 1980), which measures participants’ dispositional capacity for empathy, was completed by the participants. Tangney and Dearing (2003) found that people’s dispositional capacity for empathy is positively correlated to their feelings of guilt. As guilt is closely related to self-punishment, it is necessary to measure people’s capacity for empathy. Damage to Relationship The participants who were assigned role A played three rounds of a time-estimation game, ostensibly with another player (B) (adapted from Nelissen & Zeelenberg, 2009, Study 2). In each round, the participants and B played 10 trials of the game independently to earn points for themselves or for each other. They were told that the more points the participants owned (the points that the participants earned for themselves and the points that the other player earned for them), the more monetary rewards they would receive. In each trial, a red lamp appeared for 2,000 ms, turned green, and remained lit until a corresponding key was pressed. The participants were asked to press the key when they estimated that the green lamp had been lit for 3,000 ms. Any estimation between 2,700 ms and 3,300 ms was considered correct. A correct estimation earned 10 points. Before the formal experiment, a practice round was provided to help participants familiarize themselves with the game. During the formal experiment, at the beginning of the first round, the participants were informed that they would earn points for themselves. Regardless of their real performance, the predetermined overall feedback – that both the participants and B earned 80 points for themselves – was shown after the 10 trails of the first round were finished. Ó 2017 Hogrefe Publishing

R. Zhu et al., Relational Utility Affects Self-Punishment

The feedback served as a reference for the participants’ performance in the second round. At the beginning of the second round, the participants were informed that they would earn points for each other. The predetermined feedback showed that B earned the participant 80 points, while the participant only earned B 30 points. According to the performance in the first round, the feedback from this round implied that the participants had damaged B’s benefits, and according to the results, it seemed that the participants cared more about their own benefits than B’s benefits. In such a situation, the participants’ cooperative relationship with B could be damaged. Future Opportunity for Reciprocity At the beginning of the third round, the participants were informed that they would earn points for each other (the Future condition) or for themselves (the No future condition) and that this was the final round. Self-Punishment After learning of the recipient of the points in the third round, the participants were informed that all players could decide whether to deduct their own points, which ranged from 0 to all the points they had at that time. They were told that those deducted points just disappeared and would NOT be given to B, but a message about how many points they deducted would be sent to B (e.g., “A deducted himself or herself 5 points”). If they deducted zero points, no message would be sent. After participants deducted their points, they would find that B did not deduct any points. Afterwards, the third round was played without feedback. At the end, the participants were told that the feedback would be shown after they finished the other measurements. Emotion Measurements When the time-estimation task of the third round was finished, the participants were asked to rate (1 = very slightly or not at all, 5 = extremely) how guilty and how distressed and upset (two guilt-like emotions) they felt when they saw the feedback of the second round. Debriefing The participants were probed using a funnelled procedure that tested the participants’ comprehension of the instructions, general suspicions about the authenticity of the feedback, and interaction.

Results Three participants were excluded due to not understanding the experimental instructions or having suspicions about the authenticity of the feedback, leaving 40 (20 in the Ó 2017 Hogrefe Publishing

Future condition and 20 in the No future condition) in the subsequent analyses. Guilt To test whether Nelissen’s (2014) finding that relational utility affects guilt could be replicated, an independentsamples t-test on guilt ratings was run. The feeling of guilt was not significantly different between Future (M = 3.60, SD = 1.31) and No future (M = 3.55, SD = 1.50) conditions, t(38) = .11, p = .911, Cohen’s d = .04. Self-Punishment To test whether future reciprocity affects self-punishment, a one-way (Future opportunity for reciprocity) ANOVA on the number of self-deducted points was run. Consistent with our hypothesis that relational utility promotes selfpunishment, the number of self-deducted points in the Future condition (M = 42.70, SD = 27.97) was significantly higher than that in the No future condition (M = 20.75, SD = 22.14), F(1, 38) = 7.57, p = .009, η2 = .070. As studies have found that empathy is positively correlated to guilty feelings (Tangney & Dearing, 2003) and that negative feelings are positively correlated to selfpunishment (Inbar et al., 2013), we ran an analysis of covariance (ANCOVA) to test whether relational utility could affect self-punishment independent of these emotion-related factors. After controlling for the ratings of guilt, distress, upset, and empathy, the ANCOVA revealed that the difference in self-deducted points between the Future and No future conditions was still significant: F(1, 34) = 6.33, p = .017, η2 = .065. There were no significant effects from these covariates on self-deducted points (all Fs < 1.28, ns). To test whether wrongdoers punish themselves when no relational utility is involved, a one-sample t-test was run on the number of self-deducted points in the No future condition. The number of self-deducted points in the No future condition was significantly higher than 0, t(19) = 4.19, p < .001.

Discussion Confirming our prediction, the participants were significantly more inclined to engage in self-punishment in the Future than No future condition, even when the effects of negative emotions and empathy were controlled. Because the future chance for reciprocity increases the utility of a relationship, these results suggest that relational utility promotes self-punishment. Additionally, in the No future condition, the number of self-deducted points was significantly higher than 0, which indicated that the participants engaged in self-punishment when the victim could not offer any benefits in the future. The results are consistent with Social Psychology (2017), 48(1), 19–27

previous research that found that self-punishment is (partly) driven by guilt due to transgression (Bastian et al., 2011; Inbar et al., 2013). In Study 1, some methodological limitations may be present. First, emotional feelings were measured after the participants made their self-punishment decisions, which meant the participants’ feelings might have been regulated by the self-punishment. That may be why we did not replicate Nelissen’s (2014) finding that rational utility affects feelings of guilt. Second, we did not have control conditions in which the reciprocal relationship was not damaged. Third, we did not directly ask the participants to judge their reciprocal relationship with their partner, so our hypothesis that the participants could realize the damage to their reciprocal relationship was not supported directly. We would attempt to resolve these problems in Study 2.

Study 2 Methods Participants and Design A total of 138 undergraduate students (106 females, Mage = 21.6, SDage = 2.3) participated in the study for payment.1 The study had a 2 (Relationship status: Damaged vs. Non-damaged) 2 (Future opportunity for reciprocity: Future vs. No future) between-subject design. Procedure The basic rules and procedures of Study 2 were exactly the same as those of Study 1 except for the following elements: (1) The personality measure was finished at least two days before the experiment. (2) A new factor (Relationship status: Damaged vs. Non-damaged) was added. In the two new control conditions (Non-damaged-Future and Non-damaged-No future conditions), the participants and B earned themselves 80 points in the first round and earned each other 80 points in the second round. In such cases, the reciprocal relationships were not damaged. The other two conditions (Damaged-Future and Damaged-No future conditions) were the same as the two conditions (Future and No future conditions) in Study 1. (3) The participants’ emotions were measured after the participants knew for whom they would earn points in the third round and before their self-punishment decision. Here, we also asked

R. Zhu et al., Relational Utility Affects Self-Punishment

the participants to rate their reciprocal relationship with B at the moment (1 = very negative, 7 = very positive). (4) When the game was finished, self-punishers in the DamagedFuture and Non-damaged-Future conditions answered two questions in order to examine whether they knew self-punishment might benefit them: (a) After B knew how many points you deducted, how many points did you expect to receive from B in the third round? and (b) Assuming that you did not deduct any points, how many points would you expect to receive from B in the third round? The participants who did not deduct any points from themselves did not need to answer these questions.

Results Twelve participants were excluded due to misunderstanding the experimental instructions or having suspicions regarding the authenticity of the feedback, leaving 126 (33 in the Damaged-Future condition, 32 in the Damaged-No future condition, 31 in the Non-damagedFuture condition, and 30 in the Non-damaged-No future condition) in the subsequent analyses. Reciprocal Relationship To check the manipulation of relationship status, a 2 (Relationship status) 2 (Future opportunity for reciprocity) analysis of variance (ANOVA) on reciprocal relationship ratings was conducted. The main effect of the relationship status was significant, F(1, 122) = 380.55, p < .001, η2 = .100 (Table 1), which meant participants could realize the damage to their reciprocal relationship. Neither the main effect of future opportunity for reciprocity, F(1, 122) = .74, p = .393, η2 < .001, nor the interaction effect was significant F(1, 122) = 1.19, p = .277, η2 < .001. Guilt To examine whether relational utility affects guilt, a 2 (Relationship status) 2 (Future opportunity for reciprocity) ANOVA on guilt ratings was run. It revealed that the main effect of the relationship status was significant, F(1, 122) = 285.96, p < .001, η2 = .198, which implied the participants who harmed the other’s economic benefits were more guilty than the participants who did not. The main effect of the future opportunity for reciprocity was not significant, F(1, 122) = 2.11, p = .15, η2 = .001. Importantly, the interaction effect was significant, F(1, 122) = 7.54, p = .007, η2 = .005. A simple effects

In the beginning, we recruited 88 and 120 participants in Studies 2 and 3, respectively. As the sample size was small, we decided to run an additional population of participants to ensure the reliability of our results. Using G*Power 3 software (Faul, Erdfelder, Lang, & Buchner, 2007), we determined that the minimum sample size was 29 participants per condition, which could provide adequate power (1 β > .80) and mediumsized effect (f = .25). To meet this standard, we ran additional 50 participants in Study 2 and additional 6 participants in Study 3. The additional samples did not change the statistical results of Studies 2 and 3, so we only report the results when the additional samples were included.

Social Psychology (2017), 48(1), 19–27

Ó 2017 Hogrefe Publishing

R. Zhu et al., Relational Utility Affects Self-Punishment

Table 1. Means (and standard deviations) of negative feelings, empathy, reciprocal relationship ratings, and self-deducted points in Study 2 Damaged

Non-damaged

Future

No future

Future

No future

Guilt

4.09 (1.04)

3.44 (1.24)

1.03 (0.18)

1.23 (0.57)

Distress

2.42 (0.90)

2.22 (1.26)

1.26 (0.58)

1.53 (0.78)

Upset

2.18 (1.18)

2.00 (1.11)

1.42 (0.85)

1.83 (1.05)

Empathy

3.43 (0.38)

3.33 (0.45)

3.56 (0.28)

3.49 (0.31)

Relationship

3.30 (1.40)

3.34 (1.00)

6.87 (0.34)

6.53 (0.78)

32.27 (16.82)

12.19 (19.30)

0.16 (0.90)

1.67 (5.92)

Deducted points

analysis showed that in the Damaged condition, participants felt marginally more guilt when the future opportunity for reciprocity was present compared to it was absent, F(1, 123) = 3.08, p = .082, η2 = .024, but in the No-damaged condition they did not feel more guilt when the future opportunity for reciprocity was present compared to when it was absent, F(1, 123) = .36, p = .547, η2 = .003. As the relational utility is manipulated by the future opportunity for reciprocity, these results suggest relational utility affects guilt when people do commit a transgression, which is consistent with Nelissen (2014)’s finding. Self-Punishment To examine whether relational utility affects selfpunishment, a 2 (Relationship status) 2 (Future opportunity for reciprocity) ANOVA on self-deducted points was run. There was a significant main effect of the future opportunity for reciprocity, F(1, 122) = 15.31, p < .001, η2 = .045, and a significant main effect of the relationship status, F(1, 122) = 80.58, p < .001, η2 = .234. Importantly, the interaction effect was also significant, F(1, 122) = 20.67, p < .001, η2 = .060 (Figure 1). A simple effects analysis showed that in the Damaged condition, participants deducted themselves more points when the future opportunity for reciprocity was present compared to when it was absent, F(1, 123) = 23.15, p < .001, η2 = .158, but in the No-damaged condition they did not deduct themselves more points when the future opportunity for reciprocity was present compared to when it was absent, F(1, 123) = .18, p = .671, η2 = .001. To test whether relational utility could affect selfpunishment independently, a follow-up ANCOVA, controlling for ratings of guilt, distress, upset, and empathy, was run. It revealed that the main effect of the future opportunity for reciprocity and the interaction effect were still significant, F(1, 118) = 13.03, p < .001, η2 = .034, F(1, 118) = 13.09, p < .001, η2 = .034. The main effect of the relationship status was marginally significant, F(1, 118) = 3.57, p = .061, η2 = .009. Guilt had a significant effect on the self-deducted points, F(1, 118) = 12.48, p = .001, η2 = .033. Not surprisingly, the guiltier the Ó 2017 Hogrefe Publishing

Figure 1. Mean self-deducted points (± SE) in different conditions in Study 2.

wrongdoers feel, the more severe are the punishments they inflicted on themselves (Inbar et al., 2013). There were no significant effects of the other covariates on the selfdeducted points (all Fs < 1.28, ns). These results suggest relational utility, independent of emotional factors, could affect self-punishment when people do commit a transgression. To test whether wrongdoers punish themselves when no relational utility is involved, a one-sample t-test was run on the number of self-deducted points in the Damaged-No future condition. The number of self-deducted points in the Damaged-No future condition was significantly higher than 0, t(31) = 3.57, p = .001. Expectation To test whether self-punishers know self-punishment might benefit them, a paired-samples t-test was run. In the Damaged-Future condition, the self-punishers (30 participants), who deducted themselves more than 0 point, expected that in the third round B would earn them more points after their self-punishment (M = 69.67, SD = 12.72) compared to if they had not punished themselves (M = 56.67, SD = 21.38), t(30) = 3.29, p = .003, Cohen’s d = .677. These data in the Non-damaged-Future condition were not analyzed, as there was only one self-punisher. Social Psychology (2017), 48(1), 19–27

Discussion Consistent with Study 1, the participants punished themselves more severely for their transgressions if the victim was of higher relational utility. It is noteworthy that when negative emotions were measured before selfpunishment, we replicated Nelissen’s (2014) finding that relational utility intensified feelings of guilt following a transgression. After controlling the effects of negative emotions and empathy, the main effect of future opportunity for reciprocity and the interaction effect on self-punishment were still significant, which implies that relationship utility could directly affect self-punishment independent of negative emotions. Confirming that the participants realized that their reciprocal relationships with B were damaged when they earned B only 30 points, participants in the Damaged conditions, compared to the participants in the Non-damaged conditions, reported that their reciprocal relationships were more negative. In addition, the self-punishers in the Damaged-Future condition expected that B would earn them more points after they punished themselves than if they had not punished themselves. It indicates that in their minds, self-punishers know self-punishment may benefit them. Together, Studies 1 and 2 suggest that relational utility can influence self-punishment independent of guilt in a direct reciprocity situation. Would this type of interestoriented self-punishment be found in an indirect reciprocity situation? We would explore this question in Study 3.

Study 3 Methods

R. Zhu et al., Relational Utility Affects Self-Punishment

for himself or herself. The feedback showed that B earned the participant 80 points, while the participant earned B only 30. The Amount of Potential Benefits and Future Opportunity for Reciprocity The feedback showed that in the first and second rounds, C earned himself or herself 70 and 90 points (the More condition) or 40 and 30 points (the Less condition). At the beginning of the third round, the participants were informed that they and C would earn points for each other while B would earn points for himself (the Future condition) or that all players would earn points for themselves (the No future condition). Self-Punishment The participants were informed that the players could decide whether to deduct points from themselves. The participants were also told that randomly decided by a computer program a message about how many points they deducted would be sent to C (bystander) only, but NOT to B (victim). Personality, Emotion Measurements, and Debriefing The measurements and debriefing procedure were identical to those in Study 1.

Results Ten participants were excluded from the following analysis due to either misunderstanding the experimental instructions or having suspicions about the authenticity of the interaction, leaving 116 participants (29 participants in each condition) in the subsequent analyses.

Participants and Design A total of 126 undergraduate students (90 females, Mage = 22.0, SDage = 2.2) participated in the study for payment. The study had a 2 (Amount of potential benefits: More vs. Less) 2 (Future opportunity for reciprocity: Future vs. No future) between-subject design.

Guilt To test whether relational utility affects guilt in an indirect reciprocity situation, a 2 (Future opportunity for reciprocity) 2 (Amount of potential benefits) ANOVA on guilt ratings was conducted. There was no significant main effect or interaction effect (all Fs < 1, ns) (Table 2).

Procedure The basic rules and procedures of the game in Study 3 were similar to those of Study 1. The difference was that the participant played the game ostensibly with two other players (B as the victim and C as the bystander).

Self-Punishment To examine whether relational utility affects selfpunishment in an indirect reciprocity situation, a 2 (Future opportunity for reciprocity) 2 (Amount of potential benefits) ANOVA on self-deducted points was run. There was a significant main effect of the future opportunity for reciprocity, F(1, 112) = 7.40, p = .008, η2 = .037, and a marginally significant main effect of the amount of potential benefits, F(1, 112) = 3.75, p = .055, η2 = .019) (Figure 2). The interaction effect was not significant, F(1, 112) = 1.41, p = .237, η2 = .007.

Damage to Relationship In the first round, all players earned points for themselves. The feedback showed that the participant and B earned 80 points for themselves. In the second round, the participant and B earned points for each other while C earned points Social Psychology (2017), 48(1), 19–27

Ó 2017 Hogrefe Publishing

R. Zhu et al., Relational Utility Affects Self-Punishment

Table 2. Means (and Standard Deviations) of negative feelings, empathy, reciprocal relationship ratings, and self-deducted points in Study 3 More

Less

Future

No future

Future

No future

Guilt

4.03 (0.94)

3.86 (1.33)

3.83 (1.44)

3.69 (1.20)

Distress

2.55 (1.24)

2.00 (1.04)

2.21 (1.26)

2.48 (1.43)

Upset

2.83 (1.44)

2.17 (1.36)

2.55 (1.15)

2.45 (1.43)

3.37 (0.41)

3.44 (0.45)

3.55 (0.44)

3.36 (0.47)

20.52 (18.39)

9.52 (11.80)

11.72 (15.60)

7.41 (14.05)

Empathy Deducted points

Discussion

Figure 2. Mean self-deducted points (± SE) in different conditions in Study 3.

To test whether relational utility could affect selfpunishment independently, a follow-up ANCOVA, controlling for ratings of guilt, distress, upset, and empathy, was conducted. It revealed that the main effect of the future opportunity for reciprocity was significant, F(1, 108) = 7.18, p = .009, η2 = .033. The main effect of the amount of potential benefits was marginally significant F(1, 108) = 3.04, p = .084, η2 = .014. The interaction effect was not significant, F(1, 108) = .924, p = .339, η2 = .004. Guilt had a significant effect on the self-deducted points, F(1, 108) = 6.40, p = .013, η2 = .030. There were no significant effects of the other covariates on the selfdeducted points (all Fs < 2.41, ns). As the bystander’s relational utility is manipulated by the future opportunity for reciprocity and the amount of potential benefits, these results suggest relational utility, independent of emotional factors, could affect self-punishment in an indirect reciprocity situation. To test whether wrongdoers punish themselves when no relational utility is involved in an indirect reciprocity situation, one-sample t-tests were run on the number of selfdeducted points in the More-No future and the Less-No future conditions. The number of deducted points in the More-No future and the Less-No future conditions was significantly higher than 0, t(28) = 4.34, p < .001, t(28) = 2.84, p = .008. Ó 2017 Hogrefe Publishing

As the participants’ self-punishment was shown only to the bystander, the significant effects of relational utility on selfpunishment provide evidence that the participants took the relational utility of the bystander into consideration when making a decision about self-punishment. The results support our prediction that relational utility also exacerbates self-punishment in indirect reciprocal relationships. In addition, in the No future conditions, the participants’ self-deducted points were significantly more than 0. This indicates that wrongdoers punish themselves even when the bystander could not provide any benefits for them, which confirms previous research findings that selfpunishment is partly driven by guilt due to transgression (Inbar et al., 2013). In Study 3, feelings of guilt were not affected by the relational utility of the bystander. This is reasonable because the source of guilt is the harm inflicted on the victim. Guilty feelings are closely related to the victim rather than the bystander. Some researchers have proposed that shame instead of guilt may be affected by the relational utility of the bystander in an indirect reciprocity context (Nelissen, 2014). This needs to be confirmed in future studies.

General Discussion The present research investigated the influence of relational utility on self-punishment. In Studies 1 and 2, the participants were more inclined to punish themselves when the victim could benefit them in the future compared to when the victim could not. In Study 3, the participants were more likely to engage in self-punishment when the bystander could potentially offer many benefits to them in the future compared to when the bystander could offer only a few or no benefits. These effects were still robust after controlling negative emotions and empathy. These results demonstrate that relational utility can directly influence self-punishment in both direct and indirect reciprocity situations. Thus we extend the effect of relational utility on feelings of guilt (Nelissen, 2014) to guilt-related behavior (self-punishment), which further Social Psychology (2017), 48(1), 19–27

supports the notion that moral behaviors bring long-term benefits to people through the maintenance of reciprocal rewarding relationships (Baumard et al., 2013). Our results confirm the close relationship between selfpunishment and guilt. After being induced to feel guilty, the participants punished themselves even when no benefits were involved. These results are consistent with previous findings that self-punishment is partly driven by guilt due to transgression (Bastian et al., 2011; Inbar et al., 2013). It indicates that the effects of emotions and cost-benefit analyses may be intermingled in a selfpunishment decision. Our findings that both the victim and the bystander are considered as potential receivers of self-punishment by participants appear to be inconsistent with the results of Nelissen’s study (2011), in which participants seemed to use self-punishment as a signal of remorse for victims specifically. This difference is likely to be caused by different experimental manipulations. In our Study 3, the participants were clearly aware that their transgressions were known by the bystander. However, in the study by Nelissen (2011), the participants in the general audience condition were not sure whether the bystander knew about their transgression and thus did not inflict extra punishment on themselves. There are some limitations of the present studies. First, there is not a good cover story for the possibility of deducting points from the participants’ own private pool. Nevertheless, our findings that the number of self-deducted points was affected by guilt and relational utility implied that the participants did understand how to make use of this operation. Secondly, the present studies focus only on financial self-punishment. As Tanaka et al. (2015) suggested, in addition to inflicting a financial loss on themselves, the wrongdoers could also punish themselves by suffering physical pain. Although the cost-benefit perspective implies that physical self-punishment would be affected by benefits as well, one may argue differently according to the taboo trade-off hypothesis, which suggests that body and money are incommensurable, and the sacred value of the body would be desecrated by being weighed against money (Fiske & Tetlock, 1997). Even if some wrongdoers insist on making a profit through physical self-punishment, the strong negative emotion triggered by the taboo trade-off itself could cause uncertainty in wrongdoers’ behavior (Fiske & Tetlock, 1997). Therefore, it is difficult to predict whether or how relational utility influences physical self-punishment. It is an interesting topic for future studies. In summary, self-punishment is not as empathic and moral as it appears. The present studies demonstrate that relational utility affects self-punishment. In both direct

Social Psychology (2017), 48(1), 19–27

R. Zhu et al., Relational Utility Affects Self-Punishment

and indirect reciprocity contexts, people strategically use self-punishment to pursue profits through the repair of damaged relationships. These results help us understand self-punishment in real life. Acknowledgments This work was supported by the National Basic Research Program of China (2011CB711000, 2013CB837300), the National Natural Science Foundation of China (NSFC) (31170971, 61210010, 31400888), the Major Project of National Social Science Foundation (12&ZD228), the Beijing Municipal Science & Technology Commission (Z151100003915122), and the Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China (15XNLQ05).

References Bastian, B., Jetten, J., & Fasoli, F. (2011). Cleansing the soul by hurting the flesh: The guilt-reducing effect of pain. Psychological Science, 22, 334–335. doi: 10.1177/ 0956797610397058 Baumard, N., André, J.-B., & Sperber, D. (2013). A mutualistic approach to morality: The evolution of fairness by partner choice. Behavioral and Brain Sciences, 36, 59–78. doi: 10.1017/ S0140525X11002202 Davis, M. H. (1980). A multidimensional approach to individual differences in empathy. JSAS Catalog of Selected Documents in Psychology, 10, 85. Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. doi: 10.3758/BF03193146 Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social norms. Evolution and Human Behavior, 25, 63–87. doi: 10.1016/S1090-5138(04)00005-4 Fiske, A. P., & Tetlock, P. E. (1997). Taboo trade-offs: Reactions to transactions that transgress the spheres of justice. Political Psychology, 18, 255–297. doi: 10.1111/0162-895X.00058 Haidt, J. (2003). The moral emotions. In R. J. Davidson, K. R. Scherer, & H. Hill Goldsmith (Eds), Handbook of affective sciences (pp. 852–870). Oxford, UK: Oxford University Press. Inbar, Y., Pizarro, D. A., Gilovich, T., & Ariely, D. (2013). Moral masochism: On the connection between guilt and self-punishment. Emotion, 13, 14–18. doi: 10.1037/a0029749 Nelissen, R. M. A. (2011). Guilt-induced self-punishment as a sign of remorse. Social Psychological and Personality Science, 3, 139–144. Nelissen, R. M. A. (2014). Relational utility as a moderator of guilt in social interactions. Journal of Personality and Social Psychology, 106, 257–271. doi: 10.1037/a0034711 Nelissen, R. M. A., & Zeelenberg, M. (2009). When guilt evokes self-punishment: Evidence for the existence of a Dobby Effect. Emotion, 9, 118–122. doi: 10.1037/a0014540 Nowak, M. A., & Sigmund, K. (1998). The dynamics of indirect reciprocity. Journal of Theoretical Biology, 194, 561–574. doi: 10.1006/jtbi.1998.0775

Ó 2017 Hogrefe Publishing

R. Zhu et al., Relational Utility Affects Self-Punishment

Ohtsubo, Y., & Watanabe, E. (2009). Do sincere apologies need to be costly? Test of a costly signaling model of apology. Evolution and Human Behavior, 30, 114–123. doi: 10.1016/ j.evolhumbehav.2008.09.004 Ohtsubo, Y., & Yagi, A. (2015). Relationship value promotes costly apology-making: Testing the valuable relationships hypothesis from the perpetrator’s perspective. Evolution and Human Behavior, 36, 232–239. doi: 10.1016/j.evolhumbehav. 2014.11.008 Tanaka, H., Yagi, A., Komiya, A., Mifune, N., & Ohtsubo, Y. (2015). Shame-prone people are more likely to punish themselves: A test of the reputation-maintenance explanation for selfpunishment. Evolutionary Behavioral Sciences, 9, 1–7. doi: 10.1037/ebs0000016 Tangney, J. P., & Dearing, R. L. (2003). Shame and guilt (pp. 78–89). New York, NY: Guilford Press. Trivers, R. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35–57. doi: 10.1086/406755 Watanabe, E., & Ohtsubo, Y. (2012). Costly apology and selfpunishment after an unintentional transgression. Journal of Evolutionary Psychology, 10, 87–105. doi: 10.1556/JEP.10. 2012.3.1

Ó 2017 Hogrefe Publishing

Received March 31, 2016 Revision received September 6, 2016 Accepted September 7, 2016 Published online March 15, 2017 Chao Liu State Key Laboratory of Cognitive Neuroscience and Learning Beijing Normal University No. 19 Xinjiekouwai Street Beijing, 100875 PR China liuchao@bnu.edu.cn Xiaoqin Mai Department of Psychology Renmin University of China No. 59 Zhongguancun Street Beijing, 100872 PR China chinamaixq@ruc.edu.cn

Social Psychology (2017), 48(1), 19–27

Original Article

Stressing the Advantages of Female Leadership Can Place Women at a Disadvantage Joris Lammers and Anne Gast Social Cognition Center Cologne, University of Cologne, Germany

Abstract: Women are still underrepresented in management and men hold the majority of higher positions. Nonetheless, one often-heard claim in popular media is that female people-centered leadership skills (empathy, communication, etc.) are a better match for the business world – especially in the future. Furthermore, a related idea is that women may use this advantage to take over men’s dominant position in leadership. Four studies show that such claims paradoxically maintain gender inequality, by undermining support for affirmative action to reduce female underrepresentation in leadership. Where earlier research shows that positive stereotypes can hurt women by suggesting that they are unqualified for leadership, the current findings show that even positive stereotypes that claim that women are particularly well qualified for leadership can hurt women in their chances for gaining leadership positions. Although it is good to highlight the advantages of female leadership, exaggerated and sensationalist claims contribute to a perpetuation of gender inequality. Keywords: gender, sexism, stereotypes, biases, leadership

Introduction A common idea that recurs in contemporary popular media is that women differ from men in their leadership style and that these female leadership skills offer women important advantages – especially in postindustrial society. Specifically, women are thought to display more people skills – such as empathy, talent for communication, and increased cooperativeness – in their leadership, and these are thought to be increasingly important for effective leadership in the 21st century. Therefore, the argument goes, where in the old days men enjoyed an advantage with their aggressive, dominant, and directive leadership style, women will have an advantage over men in postindustrial society. The Economist, for example, recently devoted an article to these ideas and called men “the new weaker sex” (2015). A related idea, that often goes hand in hand with the first idea, but is theoretically distinct, is that due to these advantages, men will soon lose their advantaged position and women will take over the majority of leadership positions in business, industry, and society at large. After all, if women make better leaders than men, then ceteris paribus they should be more likely to perform well, receive promotions, etc., and therefore slowly but surely take over the majority position. Both Bloomberg Business (2012) and Business Week (2003) made such claims and anticipated a Social Psychology (2017), 48(1), 28–39 DOI: 10.1027/1864-9335/a000292

reversal of the gender gap. As another example, when The Telegraph (2015) recently discussed Hillary Clinton’s nomination as presidential candidate for the Democrats, it could not resist the temptation to headline that women “finally take over the world.” Furthermore, a series of popular scientific bestsellers retells this same story of how the era of men is about to end and how women are about to take over (e.g., Hymowitz, 2012; Mundy, 2012; Rosin, 2012). Although combining common gender stereotypes with strong claims that society is about to radically change makes for entertaining reading, the empirical support behind these two ideas is quite weak. For sure, there is certainly support for the first idea – that women on average exercise a more productive leadership style than men – but the size of that gender effect is small and similarities between the genders far outweigh their differences (Eagly, 2007; Eagly & Carli, 2003; Hyde, 2005; Vecchio, 2002). For example, although women are more likely than men to exercise transformational leadership styles, these gender effects are small (Eagly, Johannesen-Schmidt, & van Engen, 2003; Eagly & Johnson, 1990). Regarding the second idea, there is no solid evidence that any leadership advantages that women may enjoy are also translated into increased chances in gaining leadership positions. Although women now make up about half of Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

the labor force in America, they still hold only 4% of CEO positions (Catalyst, 2016). Even if women make much better leaders than men, there may be many reasons preventing such advantages from boosting their career and helping them gain promotions – sexism, discrimination, and patriarchy ranking among the most important of them (Eagly, 2007; Ellemers, 2014; Heilman, 2001; Powell, Butterfield, & Parent, 2002; Schein, 2001). In the current research, we try to answer the new and until now unanswered question how such claims about the inherent advantages of female leadership affect support for increasing more women in leadership and influence the likelihood that women are actually selected for leadership roles. Intuitively, the answer to that question is obvious; communicating the strengths of female leadership should increase the market value of potential female leaders and therefore increase the likelihood that they are selected. Yet we propose that conveying the advantages of female leadership lowers that likelihood – at least if this is done in the way that the media does, by making exaggerated and stereotypical statements about the leadership skills of women and combining them with the sensationalist claim that women are in fact taking over. In particular, we expect that such information reduces people’s support for affirmative action that is aimed at reducing gender inequality in leadership.

Support for Affirmative Action Affirmative action refers to any policy that seeks to increase the chances of groups that are underrepresented due to sexism, racism, etc. As a program, affirmative action has always been the subject of controversy. Critics may feel that influencing the hiring or selection procedure can violate procedural justice and prevents that the best candidates are selected (Kravitz & Platania, 1993; Parker, Baltes, & Christiansen, 1997; Veilleux & Tougas, 1989). Furthermore, affirmative action can devalue the achievements of minorities, because people infer that success is not exclusively due to competence but also at least partially due to receiving additional help (Heilman, Block, & Lucas, 1992; Major, Feinstein, & Crocker, 1994). Also, affirmative action can replace old wrongs with new wrongs and by making the less-advantaged members of the majority the principal losers (Austen-Smith & Wallerstein, 2006). Despite this, many people support affirmative action, because they consider that affirmative action helps in overcoming earlier injustice and thus can lead to a fairer distribution. In other words, the advantages of affirmative action are considered to outweigh the disadvantages (Crosby, Iyer, Clayton, & Downing, 2003; Leek, Saunders, & Charbonneau, 1996; Parker et al., 1997). Ó 2017 Hogrefe Publishing

We argue in the current manuscript that this equation changes, however, if the message is conveyed that women have superior leadership skills that make them likely to take over dominant positions. Specifically, people may infer that affirmative action is not or no longer necessary, for two reasons: Leadership Traits and Numerical Dominance. Leadership Traits First, if people believe that women are on average advantaged in exercising leadership, then affirmative action can be seen as superfluous for individual female job candidates. At least if people believe in some level of meritocracy and a just world, where the best receive promotions first, women should reach a higher position sooner or later, without any affirmative action. (Furnham, 2003; Lerner & Simmons, 1966). Such a position would lead to reduced support for affirmative action, for very similar reasons that a laissez-faire-liberalist view reduces support for redistribution; it unnecessarily intervenes without much benefit (e.g., Nozick, 1974; Rand, 1979). Psychologically, if people come to believe that women enjoy an advantage over men, then it unburdens them from the need to be sensitive toward the plight of women and instead frees them to rely on any sexist biases they may have (Devine, 1989; Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997; Monteith, 1993). Nevertheless, such an influence of claims about female leadership skills on attitudes toward affirmative action would be remarkable, as it would suggest that while men receive benefits from stereotypical associations with superior leadership skills (e.g., Powell et al., 2002; Schein, 1973, 1975, 2001), women suffer from positive associations. Numerical Dominance Second, in claiming that women have superior leadership skills, the media also often claim that as a result women are or will soon be taking over the majority of leadership positions (e.g., Bloomberg Business, 2012; Business Week, 2003; Economist, 2015). If people are convinced of such claims and come to believe that women will be at some point the numerically dominant group, relegating men to a minority position, then affirmative action will violate distributive justice concerns, because any affirmative action would further reduce the position of the new minority group – men (Rawls, 1971/2009; Tyler, 1994). From a more psychological perspective, the perception that women will eventually be the new dominant group can justify expressing any biases toward women, because such biases are not only allowed, they are even helpful to assist the new weaker group of men (Crandall & Eshleman, 2003). In summary, we predict that the common idea in contemporary popular media that women have an edge over men in exercising leadership in postindustrial society Social Psychology (2017), 48(1), 28–39

can undermine support for affirmative action and we propose two mechanisms that can contribute to it. The current manuscript aims to test this prediction and also determine which of these two mechanisms (or neither, or both) drives that effect. Our prediction connects to an existing literature showing that seemingly positive information about women can have negative effects on women’s position in society (Glick & Fiske, 1996; Glick et al., 2000). For example, benevolent sexist stereotypes that stress women’s warmth and communion can bar women from leadership positions, which are seen as requiring stereotypically masculine cold and dominant traits (Eagly & Karau, 2002; Heilman, 2001). Our prediction is different, however, because rather than showing that positive information can hurt women by stressing that they are ill suited, we aim to show that positive information can also hurt women by expressing that they are well suited for leadership (and thus do not require affirmative action). Overview of Studies Experiment 1 employs an actual article taken from the media discussing these stereotypical female strengths in leadership and tests whether reading that article reduces participants’ support for affirmative action to increase women in leadership positions. Experiment 2 seeks to replicate this effect but uses a more controlled manipulation. Experiment 3 aims to test whether the leadership traits or the numerical-dominance explanation (or both) best explains the effect. In particular, it is possible that thoughts about women’s supposedly superior skills will also lead to beliefs that they will take over the dominant position from men. Therefore we test the two mediators in that sequential order. Finally, Experiment 4 seeks to show that this negative effect of reading about female strengths on support for affirmative action also decreases the likelihood that a female candidate is selected for a leadership position in a hiring situation. Throughout these studies, we determined our sample size a priori, we did not look at any of the results prior to collecting all data, we did not exclude any data, we report all manipulations, and we report all measures. In running these studies, which were conducted using online services, we adhered to the ethical guidelines for Academic Requesters listed at http://www.wearedynamo.org.1

Experiment 1 Participants in the Female Dominance condition first read a portion of an actual article taken from The Atlantic (2012a), 1

J. Lammers & A. Gast, Future Female Leadership

one of America’s leading public opinion magazines, that conveys the idea that stereotypical female traits provide women with an edge over men in postindustrial society. We expected that this would reduce support for affirmative action to achieve gender equality in leadership positions, compared to participants in the Control condition, who read a neutral article from the same journal.

Method Participants and Design One hundred US American Mturk users (47 women, 53 men, mean age 33.44 years) participated in this research in return for $0.30. Only workers with an approval rate of 90% or higher were allowed to participate. Average completion time was 2 min and 24 s, yielding an average hourly payment of $7.50. Participants were randomly assigned to one of two experimental conditions (Female Dominance, Control). Given a lack of prior information on effect size, we set sample size a priori to 50 per cell. This provides us with enough power (1 β = 0.80) to detect a medium effect of d = .50 (Faul, Erdfelder, Lang, & Buchner, 2007). Manipulation After providing informed consent, participants were first asked to read a short text taken from a 2012 issue of the journal The Atlantic. Specifically, participants in the Female Dominance condition read a short (93 words) abstract of an article entitled “The End of Men” (Atlantic, 2012a). The abstract noted that women had made great advances in reaching equality, in particular in holding leadership positions. Consistent with the idea that women have better leadership traits than men and therefore may take over as a group, it posed the rhetoric questions “What if equality isn’t the end point? What if modern, postindustrial society is simply better suited to women?” In the Control condition, participants instead read an equally short abstract of a different article from the same issue of the Atlantic, entitled “No Refills” (Atlantic, 2012b) that concerned the difficulty of marketing pharmaceuticals. Affirmative Action Next, all participants completed a four-item scale measuring their support for affirmative action in dealing with gender inequality in leadership. Given that our predictions focused in particular on affirmative action to increase the proportion of women in leadership positions, we used a novel measure that focused on leadership in particular. Specifically, participants were asked to indicate their

Our only deviation from these guidelines is that these studies were conducted under a blanket Institutional Review Board (IRB) protocol, precluding workers from directly contacting an ethical board. Instead, all feedback or critique by workers was shared with a senior lab member, providing at least some alternative social control against abuse.

Social Psychology (2017), 48(1), 28–39

Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

opinion on the social issue whether more should be done to help women advance to leadership positions or not, by indicating their (dis)agreement with four items: “Government should force employers to use sex quotas, to ensure that more female managers are appointed,” “Future developments will by themselves ensure a balanced distribution of male and female leaders” (recoded), “The US Government needs to stimulate employers to hire more women,” and “In the long run, gender inequalities in corporate boards will balance themselves out” (recoded). All items were administered in randomized order and completed between 1 (= Strongly disagree) and 7 (= Strongly agree) with Neither agree nor disagree in the middle (4). The four items were combined in a single scale with sufficient internal reliability (Cronbach’s α = 0.73). Higher scores indicate stronger support for affirmative action. Finally, we collected demographics and thanked participants for their participation.

Results An analysis of variance (ANOVA) on support for affirmative action found a medium-sized effect of experimental condition, F(1, 98) = 11.03, p = .001, η2p = .10. As expected, participants in the Female Dominance condition were less supportive of affirmative action (M = 3.11, SD = 1.05) than participants in the Control condition (M = 3.82, SD = 1.09), 95% CI [ 1.13; 0.29], d = 0.66. Although not central to our hypotheses, we also checked whether this effect was moderated by gender. Adding gender as a factor showed that the interaction between condition and gender was marginally significant, F(1, 96) = 2.95, p = .089, η2p = .03. Separate t-tests for both genders suggested that the effect of condition was stronger for men, t(51) = 3.63, p = .001, d = 1.00, than for women, t(45) = 0.94, p = .35, d = 0.28. Given that we did not predict this nonsignificant interaction, we are careful in interpretation before first establishing its robustness in the next studies.

Experiment 2 Experiment 1 showed that a text about the advantages women allegedly enjoy over men, taken from an actual American public opinion magazine, significantly reduced support for affirmative action against gender inequality. Experiment 2 sought to replicate this while establishing more experimental control. First, to avoid building on the one specific article from The Atlantic that we used in Experiment 1, participants in the Female Dominance condition read a text that was an amalgam of elements of Ó 2017 Hogrefe Publishing

different articles on that same topic. Second, we added a third condition, in which participants read that women and men do not differ much in their leadership abilities (Gender Equality condition). We expected that participants in the Female Dominance condition would show less support for affirmative action, compared to participants in the Control condition and participants in the Gender Equality condition. This rules out that merely reading about gender reduces support for affirmative action.

Method Participants and Design One hundred fifty US American Mturk users (43 women, 107 men, mean age 33.25 years) participated in this research in return for $0.30. Only workers with an approval rate of 90% or higher who did not participate in Experiment 1 were allowed to participate. Average completion time was 4 min and 16 s, yielding an average hourly payment of $4.23. Participants were randomly assigned to one of three experimental conditions (Female Dominance, Gender Equality, Control). Given the size of the effect found in Experiment 1 (f = .33), this provides us with more than enough power (1 β = 0.95) to detect a similar-sized effect (Faul et al., 2007). Manipulation After providing informed consent, participants first read a short text about alleged recent research about leadership. In the Female Dominance condition, participants read a text that mirrored the most common claims about the rise of women, taken from three other articles on that topic (Bloomberg Business, 2012; Business Week, 2003; Economist, 2015). Specifically, the text had the heading “Future Female Leadership Strengths” and repeated the gender stereotype that women are dispositionally better equipped for leadership, because they are better in focusing on the long term, in consensus building, and in collaborating than men. Furthermore, consistent with the common theme that appears in such articles, it added the argument that these skills are a “perfect match for future business demands,” that “10 years from now the demand for female management skills will be stronger than ever,” and that it “will help women outshine men and help them gain more leadership positions.” In the Gender Equality condition, the text instead stated that “women are still underrepresented in leadership,” that “female leadership does not differ from the quality of male leadership,” and that “Female managers are equally suited to meet the demands of the business world of today as their male competitors.” In the Control condition, participants instead read a neutral text about the effect of personality on leadership and no claims were made about gender or its relation with Social Psychology (2017), 48(1), 28–39

leadership. All participants answered two items to ensure that they read the text. Affirmative Action Next, all participants completed the same four-item scale measuring support for affirmative action to achieve gender equality in leadership positions, as used in Experiment 1 (α = 0.67). Higher scores indicate stronger support. Finally, we collected demographics and thanked participants for their participation.

Results An analysis of variance (ANOVA) on support for affirmative action found a small to medium effect of experimental condition, F(2, 147) = 4.98, p = .008, η2 p = .06. A planned contrast testing the Female Dominance condition against the other two conditions was significant, t(147) = 2.84, p = .005. Participants in the Female Dominance condition were less supportive of affirmative action (M = 3.22, SD = 0.98) than participants in the Gender Equality condition (M = 3.89, SD = 1.21), 95% CI [ 1.10; 0.25], d = 0.61, and participants in the Control condition (M = 3.60, SD = 1.01), 95% CI [ 0.80; 0.04], d = 0.38. The latter two conditions did not differ significantly, p = .17, 95% CI [ 0.72; 0.13], d = 0.26. Unlike in Experiment 1, adding gender did not yield a Condition Gender interaction, F(2, 144) = 0.79, p = .45, η2p = .01.

Experiment 3 Experiments 1 and 2 showed that communicating the advantages that women enjoy over men in exercising leadership reduces support for affirmative action against gender inequality. Experiment 3 aims to test why this is the case, by comparing the two mechanisms that may drive the effect and that we identified in the Introduction. Specifically, we tested a first proposed mechanism that communicating superior leadership skills of women directly undermines support for affirmative action, because it adds to candidates who are already advantaged. A second proposed mechanism, which potentially follows from the first, is that women as a group will take over the dominant positions from men. Theoretically, it could also be that both inferences mediate the effect (or none). In particular, logically the second mediator may follow the first mediator, sequentially. Experiment 3 differentiates between these

J. Lammers & A. Gast, Future Female Leadership

two explanations, by using a multiple mediation design. Finally, because in previous studies our samples only included US Americans, we here include a broader sample that we contacted over the Prolific Academic platform.

Method Participants and Design One hundred five Prolific Academic users (46 women, 59 men, mean age 29.72 years; 61% European, 17% Indian, 11% US American, 11% other) participated in this research in return for £0.50. Only adult workers with an approval rate of 90% or higher were allowed to participate. Participants took on average 5 min and 8 s, yielding an average hourly payment of £5.85. Participants were randomly assigned to one of two experimental conditions (Female Dominance, Equality). We used the same sample size per cell as in Experiments 1 and 2. Measures The procedure and measure were the same as in Experiment 2, including the same measure of support for affirmative action (α = .50), except for the addition of the following two mediators: We used two items to measure participants’ belief that women have Superior Leadership Traits than men: “A female leadership style is more effective in business than a male leadership style” and “Female management skills are more important in business than male management skills” (Pearson r = .63, p < .001, SpearmanBrown = .77, α = .77). To measure participants’ belief that women will take over the Dominant Group Position, we used two further items: “There will soon be more women than men in management” and “Men will soon become a dying race in business management” (Pearson r = .53, p < .001, Spearman-Brown = .69, α = .68). All items were administered in randomized order and completed between 1 (= Strongly disagree) and 7 (= Strongly agree).2 Table 1 shows a correlation table for the independent, mediating, and dependent variables.

Results Affirmative Action An ANOVA on support for affirmative action found a small to medium effect of experimental condition, F(1, 103) = 7.23, p = .008, η2 p = .07. Participants in the Female Dominance condition were less supportive of

We also administered an additional item of beliefs about leadership traits, “Men are more likely to have the personality that is needed to be a great leader,” and one additional item of group dominance beliefs “Men will continue to hold more than half of the positions in business management.” but dropped these to increase alpha levels. Inclusion of these items leads to highly similar results in the mediation analysis.

Social Psychology (2017), 48(1), 28–39

Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

Table 1. Correlation table for the independent, mediating, and dependent variables in Study 3 Condition Traits Condition (1 = Female Dominance, 0 = Equality) Leadership traits

.260** .260**

Dominant group position

.248**

.655**

Support for affirmative action

.256**

.173

Group

Aff. action

.248**

.256**

.655**

.173 .094

.094

Notes. Values indicate Pearson r. Associated p-values: *p < .05; **p < .01; ***p < .001.

affirmative action (M = 3.83, SD = 0.90) than participants in the Control condition (M = 4.34, SD = 1.05), 95% CI [ 0.89; 0.13], d = 0.53. As in Experiment 2 and unlike in Experiment 1, adding gender did not yield a Condition Gender interaction, F(1, 101) = 0.01, p = .92, η2p = .00. Beliefs About Leadership and Dominance An ANOVA on participants’ beliefs in women’s Superior Leadership Traits found a small to medium effect of experimental condition, F(1, 103) = 7.49, p = .007, η2p = .07. Participants in the Female Dominance condition believed stronger in the superiority of female leadership (M = 4.16, SD = 1.54) than participants in the Control condition (M = 3.42, SD = 1.21), 95% CI [0.20; 1.27], d = 0.53. An ANOVA on participants’ beliefs on women’s projected Dominant Group Position found a small to medium effect of experimental condition, F(1, 103) = 6.75, p = .011, η2p = .06. Participants in the Female Dominance condition believed stronger in women’s projected dominance (M = 3.55, SD = 1.31) than participants in the Control condition (M = 2.88, SD = 1.33), 95% CI [0.16; 1.18], d = 0.51. Mediation To test whether beliefs about women’s Superior Leader Traits, women’s Dominant Group Position, or both, explained the effect of condition on support for affirmative action, we used Model 6 of the Process macro (Hayes, 2013), using 10,000 bootstrap resamples, using unstandardized predictors. This allowed us to determine whether any of the two single mediators or the combination of the two better explains the direct effect of condition on support for affirmative action, by testing three corresponding contrast definitions. This showed that a model that included only belief in women’s Superior Leadership Traits was superior to a model that also included belief in women’s Dominant Group Position, 95% CI [0.08; 0.66], and also superior to a model that only included belief in women’s Dominant Group Position (and not belief in women’s Superior Leadership Traits), 95% CI [0.08; 0.79]. Running a simple mediation that included only participants’ belief in women’s Superior Leadership Traits as Ó 2017 Hogrefe Publishing

a mediator, using Model 4 of the Process macros (Hayes, 2013), showed a significant indirect effect through that mediator, 95% CI [0.03; 0.33], although this mediation was only partial and left a significant direct effect of condition, B = 0.65, SE = 0.19, p = .001.

Experiment 4 Experiment 4 sought to show that stressing the advantages of female leadership consequently also reduces the likelihood that women are preferentially hired for leadership positions. We therefore added a second dependent variable, in which participants made a mock hiring decision. We expected that participants in the Female Dominance condition would be less inclined to preferentially hire a female candidate (than those in the Gender Equality and Control conditions) and that this effect would be due to (i.e., statistically mediated by) reduced support for affirmative action

Method Participants and Design Three hundred two US American Mturk users (137 women, 165 men, mean age 35.05 years) participated in this research in return for $0.50. Only workers with an approval rate of 90% or higher who did not participate in Experiment 1 or 2 were allowed to participate. Average completion time was 3 min and 44 s, yielding an average hourly payment of $8.04. The design was the same as in Experiment 2, with one additional dependent measure. Participants were randomly assigned to one of three experimental conditions (Female Dominance, Equality, Control). We doubled sample size a priori compared to previous studies, to 100 per cell, because we added a dichotomous dependent variable. Assuming a small to medium effect (Cohen’s ω = .2) this provides us with enough power (1 β = 0.90) to detect this effect (Faul et al., 2007). Measures After receiving the message, participants first completed the same four-item measure of support for affirmative action in achieving gender equality in leadership (α = 0.76) as in previous studies. Next, participants were told that “we are interested in your perceptions of possible leaders.” Participants were presented with six “leadership candidates” for a vacancy for a position as “high-ranking civil servant.” All six candidates were represented by photographs and except for that were only identified with letters A–F. Three candidates Social Psychology (2017), 48(1), 28–39

were male and three were female. We coded whether participants picked a female or male candidate. To ensure that the people depicted in the stimuli were comparable in actual power but also unknown to participants, all pictures were of white, middle-aged Members of Parliament of the Swedish Riksdagen. The selection of the six stimuli was based on a pilot test in which participants (N = 204) rated pictures of 26 MPs on their suitability as a leader. We picked the six candidates that were closest on average ratings (overall difference, p > .90). Furthermore, given findings that smiles (Horiuchi, Komatsu, & Nakaya, 2012) and the wearing of glasses (Fleischmann, Lammers, Stoker, & Garretsen, 2016) can influence participants’ selection of candidates, we ensured that all pictures showed a similar expression (mild smile) and that within both genders, two candidates wore glasses and one did not. Also, to avoid any gender role congruency effects, no information was given about the job requirements, except that the vacancy concerned a high-ranking position.

Results Affirmative Action An ANOVA on support for affirmative action found a small to medium effect of experimental condition, F(2, 299) = 12.15, p < .0001, η2p = .08. As in Experiment 2, a planned contrast testing the Female Dominance condition against the other two conditions was highly significant, t(299) = 4.61, p < .0001. As in Experiment 2, participants in the Female Dominance condition showed less support for affirmative action (M = 3.14, SD = 0.99) than participants in the Equality condition (M = 3.95, SD = 1.27), 95% CI [ 1.13; 0.48], d = 0.70, and less support for affirmative action than participants in the Control condition (M = 3.66, SD = 1.22), 95% CI [ 0.84; 0.19], d = 0.46. As in Experiment 2, participants in the latter two conditions did not differ significantly, p = .08, 95% CI [ 0.61; +0.03], d = 0.23. Similar to Experiments 2 and 3, adding gender did not yield a Condition Gender interaction, F(2, 296) = 0.81, p = .45, η2p = .01. Hiring Decision Participants in the Female Dominance condition were less likely to recommend a female candidate for the vacancy (61.0% female, 39.0% male), compared to participants in the Equality (73.0% female, 27.0% male) and Control conditions (71.6% female, 28.4% male), w2(1) = 3.95, p = .047, π = 0.11. Adding gender did not yield a Condition Gender interaction, w2(1) = 1.09, p = .30. In other words, although participants were more likely to select a female than a male candidate across conditions, this tendency was reduced after participants read about stereotypical female strengths. Social Psychology (2017), 48(1), 28–39

J. Lammers & A. Gast, Future Female Leadership

Mediation We then tested whether this latter effect of condition on decision to hire a woman was mediated by differences in support for affirmative action, by creating two dummy variables to code the three conditions and test mediation for each using the Process procedure with bootstrapping, using unstandardized predictors, and drawing 10,000 resamples (Hayes, 2013). Following instructions by Hayes and Preacher (2014) for analyzing the effects of a multicategorical independent variable, we used an analysis of covariance (ANCOVA) to test for the effect of the one contrast while controlling for the other, and vice versa. In both cases, we found significant support for an indirect effect through support for affirmative action. Specifically, when testing the contrast between the Female Dominance and Equality condition, we found a significant indirect effect through support for affirmative action, 95% CI [0.30; 0.94], after which the direct effect of condition was no longer significant, B = 0.09, SE = 0.33, p = .80. Similarly, testing the contrast between the Female Dominance and Control condition, we also found a significant indirect effect through support for affirmative action, 95% CI [0.14; 0.66], after which the direct effect of condition was no longer significant, B = 0.21, SE = 0.32, p = .50.

General Discussion In recent years, a slew of articles in the mainstream media and an emerging popular scientific literature have claimed that in postindustrial society women are dispositionally better equipped for exercising leadership than men, that the era of male overrepresentation in leadership will therefore inevitably end, and that consequently women are destined to take over leadership in society, business, and industry from men. Four studies show that such messages can be harmful to women. Experiment 1 presented participants with a portion of an actual article that conveys this idea, taken from The Atlantic (2012a). Indeed, merely reading the 93-word abstract of that article had a medium-sized negative effect on support for affirmative action, compared to participants who instead read a different article from the same journal. From a practical perspective, it is quite surprising that an actual text, printed in a widely read, high-quality journal that is read by hundreds of thousands of readers, has such negative effects on support for affirmative action. In the remaining studies we changed the manipulation to sacrifice some ecological validity for increased experimental control. Rather than using an existing text, we exposed participants to a text about the dispositional advantages of female leadership that was an amalgam of such articles. Experiment 2 showed that this reduced Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

support for affirmative action, compared to a control condition and compared to a condition in which participants read that women and men do not differ much in their leadership skills and that women are still underrepresented in leadership positions. Experiment 3 tested between two mechanisms driving this effect: belief in female leadership traits as being superior over male traits, and belief that women are taking over as the numerically most dominant group in management. Results showed that belief in female leadership traits was the critical mechanism driving this effect. The degree to which participants believed that women would as a group take over the dominant position in leadership from men did not mediate the decrease in support for affirmative action. In other words, reading such texts undermines support for affirmative action because readers consider there to be no need to help female candidates if they are truly superior leaders, but this is independent of whether readers believe that women will take over the numerical dominant position from men or not. Finally, Experiment 4 showed that reading about the rise of women reduced participants’ tendency to preferentially hire women for a high-ranking position. This effect was mediated by their reduced support for affirmative action.

Implications The current result shows that inflated claims about the inherent dispositional advantages of female leadership and their implications for overcoming gender inequality can actually reduce support for affirmative action and reduce the likelihood that women are selected for leadership roles. When people read that women have certain personality traits that provide them with an advantage over men, then this reduces people’s support for affirmative action, presumably because they consider it superfluous if women already have an advantage. We believe that the effect we identified here may have large effects on actual real life hiring decisions, for three reasons. First, these claims about dispositional advantages of female leadership appear to be particularly convincing, because they match common gender stereotypes of women being warmer and more communally oriented (Fiske, Cuddy, Glick, & Xu, 2002). Second, these claims are also well disguised, because they appear to be genuinely female-friendly, given that they point to stereotypical female strengths, rather than weaknesses. Third, these claims are often made in popular media that is directed at readers who hold leadership positions in society, such as Bloomberg Business (2012), Business Week (2003), and The Economist (2015). In other words, the people who are most likely to decide who to hire and recruit for Ó 2017 Hogrefe Publishing

future leadership positions are likely most directly and strongly influenced by these claims. These results do not mean that the media should stop discussing the strengths of female leadership. Indeed, sound empirical research has firmly demonstrated a small gender effect, meaning that women outperform men slightly in some aspects of leadership performance (Eagly & Carli, 2003; Eagly et al., 2003). It is important to share such findings with a wider audience as it can defeat common stereotypes that equate good leadership with being male (Schein, 1973, 1975). Yet it is important that the media do not exaggerate such findings. These observed gender effects are small and the reality is that although women are making progress and although women have advantages over men in exercising leadership, they are nowhere near equality in attaining leadership positions and there is no indication at all that the problem of female underrepresentation will solve itself. Any advantages that women enjoy, relative to men, are more than offset by the large disadvantages that they still face. For example, compared to men, women need to be more highly qualified to be offered similar leadership roles (Biernat & Kobrynowicz, 1997; Foschi, 2000; Koch, D’Mello, & Sackett, 2015) and women with an MBA are less likely to adopt a leadership position than men with an MBA (Wilton, 2007). Women need to perform better to obtain similar promotions than men (Baxter & Wright, 2000; Elliott & Smith, 2004; Inesi & Cable, 2015) and are also paid less for similarly responsible positions than men (Aláez-Aller, Longás-García, & Ullibarri-Arce, 2011; Castilla, 2015; O’Neill & O’Reilly, 2010). Even though the stereotypical association between leaders and masculinity has decreased over the past decades, people still robustly associate leadership with being male (Haines, Deaux, & Lofaro 2016; Koenig, Eagly, Mitchell, & Ristikari, 2011). As a result, female underrepresentation in leadership is nowhere near to being overcome – let alone to reverse. Although women now make up 45% of the labor force of America’s 500 largest companies, they still hold only 4.0% of the CEO positions of those same companies (Catalyst, 2016). To summarize, even though women are equally or even slightly better skilled to fill leadership positions, compared to men, they are nonetheless still largely underrepresented in such positions. Besides structural disadvantages to women, this may be due to two mechanisms both related to the stereotypical trait attributions. First, women suffer from sexism that traditionally associates leadership with being male (Eagly, 2007; Ellemers, 2014; Heilman, 2001; Powell et al., 2002; Schein, 2001). Second, newly emerging associations between women and (modern) leadership skills also work against women, because they can be used as arguments against affirmative action or other type of Social Psychology (2017), 48(1), 28–39

support that – if we look at the actual situation – may still be needed.

Relation to Literature We are not the first to show that seemingly positive or benevolent stereotypes about women can nonetheless reinforce women’s subordinate position in society, in relation to men (Glick & Fiske, 1996; Glick et al., 2000). By stressing women’s warmth, communion, and other elevated traits, benevolent sexist stereotypes keep women from leadership positions, because these are seen as requiring more base traits, such as competitiveness and dominance, which are stereotypically attributed to men (Eagly & Karau, 2002; Heilman, 2001). Such benevolent stereotypes are particularly effective in maintaining gender equality, because they hide their message in praise and admiration of women – thus easily fending off detection or resistance against them (Becker & Swim, 2011; Becker & Wright, 2011; Jost & Kay, 2005; Kay & Jost, 2003). The current findings are different from (and add to) the above-mentioned literature, because rather than showing that positive stereotypes can hurt women by conveying the message that they are ill suited for leadership, we show that positive stereotypes can even hurt women by conveying that women are particularly well suited for leadership. Paradoxically, stressing the value of women as a group can undermine the motivation to install more women in leadership positions. These findings also connect to a wider literature showing that advances toward solving inequality can generate psychological processes that paradoxically undermine or hinder that same advancement (Ellemers, 2014; Ellemers, Rink, Derks, & Ryan, 2012). For example, even though they are nowadays more likely to be appointed to leadership positions than in the past, women are particularly often appointed in times of crises. As a result of being more often appointed in such so-called glass cliff positions, women are more likely to fail than men – ironically reinforcing sexist stereotypes (Ryan & Haslam, 2005, 2007). As another example, women in powerful positions can be particularly likely to criticize women in more subordinate positions, to fit in a male culture, and to avoid devaluing their own past struggle. As a result, women in leadership positions may contribute to a female unfriendly culture that helps to perpetuate inequality (Derks, Ellemers, Van Laar, & De Groot, 2011; Derks, Van Laar, Ellemers, & De Groot, 2011; Ellemers, 2014; Ellemers & Barreto, 2008). Similar examples are found outside gender relations. For example, increased contact with advantaged group members may lead to a decreased support for efforts to reduce inequality among members of disadvantaged groups (Dixon, Durrheim, & Tredoux, 2007). Much similar to that Social Psychology (2017), 48(1), 28–39

J. Lammers & A. Gast, Future Female Leadership

literature, the current effects show that a phenomenon that may seem to help overcoming the gender gap, in fact backfires. Stressing the dispositional advantages of female leadership runs the risk of undermining support for affirmative action and the likelihood that women are selected for leadership positions. Current and past results underscore the necessity of gaining a detailed understanding of such ironic processes that hinder progress.

Limitations One limitation to our research is that we exclusively focused on support for increasing the proportion of women in leadership roles, thus leaving open the question whether these results generalize to hiring decisions for lowerranked, non-leadership positions. We believe it is important to focus on leadership, however, because practically this is where the greatest gains can be made. As we mentioned in the Introduction, women have made great advances throughout the entire work force, but remain tragically underrepresented at the top. Furthermore, increasing the proportion of women at the top can also help other women across the organization, as they female leaders can serve as role models (Asgari, Dasgupta, & Stout, 2012; Latu, Mast, Lammers, & Bombari, 2013; Lockwood, 2006). Another limitation of our research is that the experimental manipulation that we used in our studies combined various elements of information – participants read that women are different from men and lead better than them in various ways and are therefore on the rise in management and will eventually take over as the dominant gender group holding leadership positions. This choice was deliberate and flowed from our goal to make the applied point how the messages that appear in popular media about female strengths (that use a similar mixture of arguments) affect leadership selection outcomes. A downside of this combination is that it is unclear what exactly drives the effect. Experiment 3 at least partially resolved this issue by showing that the belief that women have better leadership traits than men was critical, while the belief that they may eventually take over was less critical. Nonetheless, we welcome research that can help to tease apart these elements even better. Our claim that messages about female leadership strengths can hurt women appears to be undermined by the fact that in Experiment 4 still 61% of participants who read about these stereotypical female strengths picked a female candidate (compared to 73% and 72% in the other two conditions). Although careful pilot testing showed that our stimuli are seen as equally suited for a leadership position (on a Likert scale), it is well possible that the relatively overall high percentage may be due to a subtle stimulus effect. In addition, it is possible that participants Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

experienced some pressure to pick a female candidate (dichotomous choice). Clearly, it is difficult to interpret absolute values and it is more informative to consider differences between conditions. A further limitation of Experiment 4 is that participants were asked to base their judgment on a rather minimal amount of information – receiving only a picture of the six candidates. Nonetheless, we do note that other research has demonstrated that such seemingly naïve candidate preference measures nonetheless accurately reflect actual selection outcomes, such as in political elections (Antonakis & Dalgas, 2009; Todorov, Mandisodza, Goren, & Hall, 2005). A final concern with our studies was the exclusive reliance on online studies, conducted using the Amazon Mechanical Turk (Experiments 1, 2, and 4) and the Prolific Academic (Experiment 3) platforms. On the one hand, it has been noted that MTurk and other online platforms offer a more representative sample than typical college student samples and also exhibit good internal and external validity (Berinsky, Huber, & Lenz, 2012; Buhrmester, Kwang, & Gosling, 2011; Mason & Suri, 2011). On the other hand, these online platforms offer their own challenges, such as reduced naivety (Chandler, Mueller, & Paolacci, 2013; Chandler, Pe’er, Paolacci, Mueller, & Ratliff, 2015; Stewart, Ungemach, & Harris, 2015). To reduce that concern, we ensured that participants could not take part in more than one of the Mturk studies that we conducted. Another concern with the use of online services pertains to the ethical treatment of workers. Although evidence on the effect of the amount of compensation on the quality of the obtained responses is mixed (Buhrmester et al., 2011; Litman, Robinson, & Rosenzweig, 2014) ensuring sufficient compensation is important by itself. Across our studies, we aimed to pay participants on average at least $6 or £5 per hour, and although we succeeded in three of the four studies (average hourly payment: $6.92 and £5.85, across all participants), completing Experiment 2 unexpectedly took longer, yielding lower average payment. Additionally, although running any incentivized study creates a power dynamic between requester and participant, we sought to minimize this by explaining the goal of the research prior to participation, by avoiding false feedback, by using automatic exclusion criteria, and by approving all submissions, independent of their quality (Fort, Adda, & Cohen, 2011; Gleibs, 2016).

Conclusion An ever-recurring theme in many public opinion magazines and in a sprawling popular scientific literature holds that women have better people skills than men and that – given Ó 2017 Hogrefe Publishing

the central importance of such people skills in postindustrial society – women are destined to take over men’s dominant role in management and leadership. We show that such prophecies may be self-defeating. They undermine people’s support for affirmative action and reduce the likelihood that they preferentially pick women for leadership positions. Research on the advantages of female leadership should not be interpreted as meaning that the problem of inequality will solve itself. Rather, it should inspire people to use these advantages and make a change by hiring more women today.

References Aláez-Aller, R., Longás-García, J. C., & Ullibarri-Arce, M. (2011). Visualising gender wage differences in the European Union. Gender, Work & Organization, 18, e49–e87. doi: 10.1111/j.14680432.2009.00469.x Antonakis, J., & Dalgas, O. (2009). Predicting elections: Child’s play!. Science, 323, 1183–1183. doi: 10.1126/science.1167748 Asgari, S., Dasgupta, N., & Stout, J. G. (2012). When do counterstereotypic ingroup members inspire versus deflate? The effect of successful professional women on young women’s leadership self-concept. Personality and Social Psychology Bulletin, 38, 370–383. doi: 10.1177/0146167211431968 Atlantic. (2012a). The end of men. Boston, MD: Atlantic Media. Retrieved from http://www.theatlantic.com/magazine/archive/ 2010/07/the-end-of-men/308135/ Atlantic. (2012b). No refills. Boston, MD: Atlantic Media. Retrieved from http://www.theatlantic.com/magazine/archive/2010/07/ no-refills/308133/ Austen-Smith, D., & Wallerstein, M. (2006). Redistribution and affirmative action. Journal of Public Economics, 90, 1789–1823. doi: 10.1016/j.jpubeco.2006.05.005 Baxter, J., & Wright, E. O. (2000). The glass ceiling hypothesis: A comparative study of the United States, Sweden, and Australia. Gender & Society, 14, 275–294. doi: 10.1177/ 089124300014002004 Becker, J. C., & Swim, J. K. (2011). Seeing the unseen attention to daily encounters with sexism as way to reduce sexist beliefs. Psychology of Women Quarterly, 35, 227–242. doi: 10.1177/ 0361684310397509 Becker, J. C., & Wright, S. C. (2011). Yet another dark side of chivalry: Benevolent sexism undermines and hostile sexism motivates collective action for social change. Journal of Personality and Social Psychology, 101, 62–77. doi: 10.1037/a0022615 Berinsky, A. J., Huber, G. A., & Lenz, G. S. (2012). Evaluating online labor markets for experimental research: Amazon.com’s Mechanical Turk. Political Analysis, 20(3), 351–368. doi: 10.1093/polana/mpr057/-/DC1 Biernat, M., & Kobrynowicz, D. (1997). Gender- and race-based standards of competence: Lower minimum standards but higher ability standards for devalued groups. Journal of Personality and Social Psychology, 72, 544–557. doi: 10.1037/ 0022-3514.72.3.544 Bloomberg Business. (2012). Behind every great woman. New York, NY: Bloomberg. Retrieved from http://www.bloomberg.com/ bw/magazine/behind-every-great-woman-01042012.html#p1 Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon's Mechanical Turk: A New Source of Inexpensive, Yet HighQuality, Data? Perspectives on Psychological Science, 6(1), 3–5. doi: 10.1177/1745691610393980

Social Psychology (2017), 48(1), 28–39

Business Week. (2003). The new gender gap, Print edition, May 26, 2003. New York, NY: Bloomberg. Castilla, E. J. (2015). Accounting for the gap: A firm study manipulating organizational accountability and transparency in pay decisions. Organization Science, 26, 311–333. doi: 10.1287/orsc.2014.0950 Catalyst. (2016). Pyramid: Women in S&P 500 companies. New York, NY: Catalyst. Retrieved from http://www.catalyst. org/knowledge/women-sp-500-companies Chandler, J., Mueller, P., & Paolacci, G. (2013). Nonnaïveté among Amazon Mechanical Turk workers: Consequences and solutions for behavioral researchers. Behavior Research Methods, 46, 112–130. doi: 10.3758/s13428-013-0365-7 Chandler, J., Pe’er, E., Paolacci, G., Mueller, P., & Ratliff, K. A. (2015). Non-naïve participants can reduce effect sizes. Psychological Science, 26, 1131–1139. Crandall, C. S., & Eshleman, A. (2003). A justification-suppression model of the expression and experience of prejudice. Psychological Bulletin, 129, 414–446. doi: 10.1037/0033-2909.129.3.414 Crosby, F. J., Iyer, A., Clayton, S., & Downing, R. A. (2003). Affirmative action: Psychological data and the policy debates. American Psychologist, 58, 93–115. doi: 10.1037/0003-066X.58.2.93 Derks, B., Ellemers, N., Van Laar, C., & De Groot, K. (2011). Do sexist organizational cultures create the Queen Bee? British Journal of Social Psychology, 50, 519–535. doi: 10.1348/ 014466610X525280 Derks, B., Van Laar, C., Ellemers, N., & De Groot, K. (2011). Gender-bias primes elicit queen-bee responses among senior policewomen. Psychological Science, 22, 1243–1249. doi: 10.1177/0956797611417258 Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5–18. doi: 10.1037/0022-3514.56.1.5 Dixon, J., Durrheim, K., & Tredoux, C. (2007). Intergroup contact and attitudes towards the principle and practice of racial equality. Psychological Science, 18, 867–872. doi: 10.1111/ j.1467-9280.2007.01993.x Dovidio, J. F., Kawakami, K., Johnson, C., Johnson, B., & Howard, A. (1997). The nature of prejudice: Automatic and controlled processes. Journal of Experimental Social Psychology, 33, 510–540. doi: 10.1006/jesp.1997.1331 Eagly, A. H. (2007). Female leadership advantage and disadvantage: Resolving the contradictions. Psychology of Women Quarterly, 31, 1–12. doi: 10.1111/j.1471-6402.2007.00326.x/full Eagly, A. H., & Carli, L. (2003). The female leadership advantage: An evaluation of the evidence. The Leadership Quarterly, 14, 807–834. doi: 10.1016/j.leaqua.2003.09.004 Eagly, A. H., Johannesen-Schmidt, M. C., & van Engen, M. L. (2003). Transformational, transactional, and laissez-faire leadership styles: A meta-analysis comparing women and men. Psychological Bulletin, 129, 569–591. doi: 10.1037/0033-2909.129.4.569 Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 108, 233–256. Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders. Psychological Review, 109, 573–598. doi: 10.1037/0033-295X.109.3.573 Economist. (2015, May). The weaker sex, 2015. Ellemers, N. (2014). Women at work: How organizational features impact career development. Policy Insights from the Behavioral and Brain Sciences, 1, 46–54. doi: 10.1177/2372732214549327 Ellemers, N., & Barreto, M. (2008). Putting your own down: How members of disadvantaged groups perpetuate or exacerbate their disadvantage. In A. Brief (Ed.), Diversity at work (pp. 202–261). Cambridge, UK: Cambridge University Press. Ellemers, N., Rink, F., Derks, B., & Ryan, M. K. (2012). Women in high places – When and why promoting women into top

Social Psychology (2017), 48(1), 28–39

J. Lammers & A. Gast, Future Female Leadership

positions can harm them individually or as a group (and how to prevent this). In A. P. Brief & B. M. Staw (Eds.), Research in organizational behavior: An annual series of analytical essays and critical reviews (Vol. 32, pp. 163–187). New York, NY: Elsevier. Elliott, J. R., & Smith, R. A. (2004). Race, gender, and workplace power. American Sociological Review, 69, 365–386. doi: 10.1177/000312240406900303 Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39, 175–191. Fiske, S., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878–902. doi: 10.1037/0022-3514.82.6.878 Fleischmann, A., Lammers, J., Stoker, J. I., & Garretsen, H. (2016). Wearing glasses increases political election success Unpublished manuscript, University of Cologne. Fort, K., Adda, G., & Cohen, K. B. (2011). Amazon mechanical turk: Gold mine or coal mine? Computational Linguistics, 37, 413–420. doi: 10.1162/COLI_a_00057 Foschi, M. (2000). Double standards for competence: Theory and research. Annual Review of Sociology, 26, 21–42. Furnham, A. (2003). Belief in a just world: Research progress over the past decade. Personality and Individual Differences, 34, 795–817. Gleibs, I. H. (2016). Are all “research fields” equal? Rethinking practice for the use of data from crowd-sourcing market places. Behavior Research Methods, 8, 1210. doi: 10.3758/ s13428-016-0789-y Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: differentiating hostile and benevolent sexism. Journal of personality and social psychology, 70, 491. Glick, P., Fiske, S. T., Mladinic, A., Saiz, J. L., Abrams, D., Masser, B., . . . Annetje, B. (2000). Beyond prejudice as simple antipathy: Hostile and benevolent sexism across cultures. Journal of Personality and Social Psychology, 79, 763–775. doi: 10.1037/ 0022-3514.79.5.763 Haines, E. L., Deaux, K., & Lofaro, N. (2016). The times they are achanging. . . or are they not? A comparison of gender stereotypes, 1983–2014. Psychology of Women Quarterly. doi: 10.1177/0361684316634081 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press. Hayes, A. F., & Preacher, K. J. (2014). Statistical mediation analysis with a multicategorical independent variable. British Journal of Mathematical and Statistical Psychology, 67, 451–470. doi: 10.1111/bmsp.12028 Heilman, M. E. (2001). Description and prescription: How gender stereotypes prevent women’s ascent up the organizational ladder. Journal of Social Issues, 57, 657–674. doi: 10.1111/ 0022-4537.00234 Heilman, M. E., Block, C. J., & Lucas, J. A. (1992). Presumed incompetent? Stigmatization and affirmative action efforts. Journal of Applied Psychology, 77, 536–544. doi: 10.1037/00219010.77.4.536 Horiuchi, Y., Komatsu, T., & Nakaya, F. (2012). Should candidates smile to win elections? An application of automated face recognition technology. Political Psychology, 33, 925–933. doi: 10.1111/j.1467-9221.2012.00917.x Hyde, J. S. (2005). The gender similarities hypothesis. American Psychologist, 60, 581–592. doi: 10.1037/0003-066X.60.6.581 Hymowitz, K. S. (2012). Manning up: How the rise of women has turned men into boys. New York, NY: Basic Books.

Ó 2017 Hogrefe Publishing

J. Lammers & A. Gast, Future Female Leadership

Inesi, M. E., & Cable, D. M. (2015). When accomplishments come back to haunt you: The negative effect of competence signals on women’s performance evaluations. Personnel Psychology, 68, 615–657. doi: 10.1111/peps.12083 Jost, J. T., & Kay, A. C. (2005). Exposure to benevolent sexism and complementary gender stereotypes: Consequences for specific and diffuse forms of system justification. Journal of Personality and Social Psychology, 88, 498–509. Kay, A. C., & Jost, J. T. (2003). Complementary justice: Effects of “poor but happy” and “poor but honest” stereotype exemplars on system justification and implicit activation of the justice motive. Journal of Personality and Social Psychology, 85, 823–837. Koch, A. J., D’Mello, S. D., & Sackett, P. R. (2015). A meta-analysis of gender stereotypes and bias in experimental simulations of employment decision making. Journal of Applied Psychology, 100, 128–161. doi: 10.1037/a0036734 Koenig, A. M., Eagly, A. H., Mitchell, A. A., & Ristikari, T. (2011). Are leader stereotypes masculine? A meta-analysis of three research paradigms. Psychological Bulletin, 137, 616. doi: 10.1037/a0023557 Kravitz, D. A., & Platania, J. (1993). Attitudes and beliefs about affirmative action: Effects of target sex and ethnicity. Journal of Applied Psychology, 78, 928–938. doi: 10.1037/0021-9010.78.6.928 Latu, I. M., Mast, M. S., Lammers, J., & Bombari, D. (2013). Successful female leaders empower women’s behavior in leadership tasks. Journal of Experimental Social Psychology, 49, 444–448. doi: 10.1016/j.jesp.2013.01.003 Lerner, M., & Simmons, C. H. (1966). Observer’s Reaction to the ‘Innocent Victim: Compassion or rejection? Journal of Personality and Social Psychology, 4, 203–210. doi: 10.1037/h0023562 Litman, L., Robinson, J., & Rosenzweig, C. (2014). The relationship between motivation, monetary compensation, and data quality among US- and India-based workers on Mechanical Turk. Behavior Research Methods, 47, 519–528. doi: 10.3758/ s13428-014-0483-x Leek, J. D., Saunders, D. M., & Charbonneau, M. (1996). Affirmative action programs: An organizational justice perspective. Journal of Organizational Behavior, 17, 79–89. Lockwood, P. (2006). “Someone like me can be successful”: Do college students need same-gender role models? Psychology of Women Quarterly, 30, 36–46. doi: 10.1111/j.1471-6402.2006. 00260.x Major, B., Feinstein, J., & Crocker, J. (1994). Attributional ambiguity of affirmative action. Basic and Applied Social Psychology, 15, 113–141. doi: 10.1080/01973533.1994.9646075 Mason, W., & Suri, S. (2011). Conducting behavioral research on Amazon’s Mechanical Turk. Behavior Research Methods, 44(1), 1–23. doi: 10.3758/s13428-011-0124-6 Monteith, M. J. (1993). Self-regulation of prejudiced responses: Implications for progress in prejudice-reduction efforts. Journal of Personality and Social Psychology, 65, 469–485. doi: 10.1037/0022-3514.65.3.469 Mundy, L. (2012). The richer sex: How the new majority of female breadwinners is transforming sex, love and family. New York, NY: Simon & Schuster. Nozick, R. (1974). Anarchy, state, and utopia. New York, NY: Basis Books. O’Neill, O. A., & O’Reilly, C. A. (2010). Careers as tournaments: The impact of sex and gendered organizational culture preferences on MBAs’ income attainment. Journal of Organizational Behavior, 31, 856–876. Parker, C. P., Baltes, B. B., & Christiansen, N. D. (1997). Support for affirmative action, justice perceptions, and work attitudes: A study of gender and racial–ethnic group differences. Journal of Applied Psychology, 82, 376–389. doi: 10.1037/0021-9010. 82.3.376

Ó 2017 Hogrefe Publishing

Powell, G. N., Butterfield, D. A., & Parent, J. D. (2002). Gender and managerial stereotypes: Have the times changed? Journal of Management, 28, 177–193. doi: 10.1177/014920630202800203 Rand, A. (1979). The Ayn Rand Letter: Volumes I-IV, 1971–1976. Palo Alto, CA: Palo Alto Book Service. Rawls, J. (2009). A theory of justice. Harvard University Press. Rosin, H. (2012). The end of men: And the rise of women. New York: Riverhead. Ryan, M. K., & Haslam, S. A. (2005). The glass cliff: Evidence that women are over-represented in precarious leadership positions. British Journal of Management, 16, 81–90. doi: 10.1111/ j.1467-8551.2005.00433.x Ryan, M. K., & Haslam, S. A. (2007). The glass cliff: Exploring the dynamics surrounding women’s appointment to precarious leadership positions. Academy of Management Review, 32, 549–572. doi: 10.2307/20159315 Schein, V. E. (1973). The relationship between sex role stereotypes and requisite management characteristics. Journal of Applied Psychology, 57, 95–100. doi: 10.1037/h0037128 Schein, V. E. (1975). Relationships between sex role stereotypes and requisite management characteristics among female managers. Journal of Applied Psychology, 60, 340–344. doi: 10.1037/h0076637 Schein, V. E. (2001). A global look at psychological barriers to women’s progress in management. Journal of Social Issues, 57, 675–688. doi: 10.1111/0022-4537.00235 Stewart, N., Ungemach, C., & Harris, A. (2015). The average laboratory samples a population of 7,300 Amazon Mechanical Turk workers. Judgment and Decision Making, 10, 479–491. Telegraph. (2015). Hillary Clinton is not alone these days. Are women finally taking over the world? Retrieved from http://www. telegraph.co.uk/news/worldnews/hillary-clinton/11531542/ Hillary-Clinton-is-not-alone-these-days.-Are-women-finallytaking-over-the-world.html Todorov, A., Mandisodza, A. N., Goren, A., & Hall, C. C. (2005). Inferences of competence from faces predict election outcomes. Science, 308, 1623–1626. doi: 10.1126/science. 1113075 Tyler, T. R. (1994). Psychological models of the justice motive: Antecedents of distributive and procedural justice. Journal of Personality and Social Psychology, 67, 850–863. doi: 10.1037/0022-3514.67.5.850 Vecchio, R. P. (2002). Leadership and gender advantage. The Leadership Quarterly, 13, 643–671. doi: 10.1016/S1048-9843 (02)00156-X Veilleux, F., & Tougas, F. (1989). Male acceptance of affirmative action programs for women: The results of altruistic or egoistic motives? International Journal of Psychology, 24, 485–496. doi: 10.1080/00207594.1989.10600060 Wilton, N. (2007). Does a management degree do the business for women in the graduate labour market? Higher Education Quarterly, 61, 520–538. Received June 6, 2016 Revision received September 14, 2016 Accepted September 14, 2016 Published online March 15, 2017 Joris Lammers University of Cologne Social Cognition Center Cologne Richard-Strauss-Str. 2 50931 Cologne Germany joris.lammers@uni-koeln.de

Social Psychology (2017), 48(1), 28–39

Research Report

Stepping in the Shoes of Leaders of Populist Right-Wing Parties Promoting Anti-Immigrant Views in Times of Economic Prosperity Jolanda Jetten, Rachel Ryan, and Frank Mols School of Psychology, University of Queensland, Brisbane, Australia

Abstract: What narrative is deemed most compelling to justify anti-immigrant sentiments when a country’s economy is not a cause for concern? We predicted that flourishing economies constrain the viability of realistic threat arguments. We found support for this prediction in an experiment in which participants were asked to take on the role of speechwriter for a leader with an anti-immigrant message (N = 75). As predicted, a greater percentage of realistic threat arguments and fewer symbolic threat arguments were generated in a condition in which the economy was expected to decline than when it was expected to grow or a baseline condition. Perhaps more interesting, in the economic growth condition, the percentage realistic entitlements and symbolic threat arguments generated were higher than when the economy was declining. We conclude that threat narratives to provide a legitimizing discourse for anti-immigrant sentiments are tailored to the economic context. Keywords: opposition to immigration, economic decline, economic growth, leadership, realistic entitlement

Anti-immigrant narratives often focus on immigrants as an economic threat as well as a symbolic and existential threat (Riek, Mania, & Gaertner, 2006). Even though realistic conflict and symbolic threat narratives are not incompatible and often go hand in hand, it is often the case that one of the two narratives dominates at a particular moment in time. For example, in the Netherlands, the populist rightwing leader Geert Wilders has focused on both; at times presenting immigrants and refugees as a realistic threat (i.e., immigrants are portrayed as a burden to ordinary hardworking Dutch taxpayers), and at other times representing them as a symbolic threat – depicting Islam as an existential threat to national values and as an incompatible alien influence (e.g., the “Islamization” of Dutch culture; see González, Verkuyten, Weesie, & Poppe, 2008). In the present research, we argue that to understand the type of argument people might use to argue their anti-immigration case is shaped by the broader context in which they have to pitch their arguments. In particular, we need to take account of the state of the economy. Because successful leaders are “crafty identity entrepreneurs” who position themselves strategically (Haslam, Reicher, & Platow, 2011; Mols & Jetten, 2015, in press), we propose that the way populist leaders hope to garner support for their message is tailored to the context. While realistic threat narratives may be compelling in times of economic Social Psychology (2017), 48(1), 40–46 DOI: 10.1027/1864-9335/a000289

downturn, because they speak to the fears and anxieties that dominate public discourse at that particular moment in time, this discourse is less powerful (and is therefore less likely to be used) in times of economic prosperity. How then do anti-immigration advocates garner support for their message in times of economic prosperity? We propose that in such contexts, narratives that focus on immigrants as other types of threats will dominate – immigrants as a threat to culture (symbolic threat) or as a threat to that what is seen as rightly owned by the autochthonous population (labeled here as realistic entitlement).

Economic Downturn and Anti-Immigration Attitudes There is considerable evidence for the notion that economic downturn results in an increase in negative attitudes toward outsiders (Quillian, 1995; Riek et al., 2006). For example, across 17 European countries, McLaren (2003) found that perceived economic struggle was related to an increase of perceived realistic threat. Similar findings have been reported in studies conducted in North America: perceived competition with other groups in society was associated with negative attitudes toward immigration (Esses, Dovidio, Jackson, & Armstrong, 2001; Esses, Jackson, & Armstrong, 1998). It has also Ó 2017 Hogrefe Publishing

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

been found that populist parties are most successful in times of economic struggle, presumably because such contexts provide an ideal platform to create fear-based narratives (Arzheimer & Carter, 2006; Betz, 1993; Jackman & Volpert, 1996; Vanneman & Pettigrew, 1972). Attempts to understand these effects have so far focused on classic realistic conflict theorizing (Dollard, 1938; Sherif, Harvey, White, Hood, & Sherif, 1961) and relative deprivation theorizing (Walker & Smith, 2002). As a result of both realistic conflict and relative deprivation, people experience frustration and dissatisfaction and this leads to “lashing out” at minorities – in particular those minorities who appear to be competing for the same resources, or are perceived to be the cause of the relative deprivation (see also frustration-aggression reasoning and scapegoat theory, Allport, 1954). While this process is well understood and rather intuitive, it does not help us to explain why populist parties often flourish in times of economic growth (see Betz, 1993; Guimond & Dambrun, 2002). Given that it is easier to build a case that immigration should be curbed in times of economic downturn, the question arises what types of arguments leaders or anti-immigrant advocates generate to convince others of their anti-immigrant agenda when the economy is not a cause for concern. One possibility is that, in such contexts, anti-immigrant advocates will resort to narratives that emphasize symbolic threat. Symbolic threat stems from a perceived conflict in beliefs and cultural values, rather than from competition over resources (González et al., 2008; Kinder & Sears, 1981; Riek et al., 2006). There is already evidence that the power of symbolic threat versus realistic threat perceptions to predict antiimmigrant attitudes depends on the broader national and cultural context. For example, in Israel, Bizman and Yinon (2001) found that more negative attitudes toward Russian immigrants were explained by realistic threat not by symbolic threats. However, in the Netherlands symbolic threats were not only more prominent than realistic threats, symbolic threat perceptions were also predictive of prejudice against Muslim immigrants, whereas realistic threat perceptions were not (González et al., 2008; Sniderman & Hagendoorn, 2007). These findings are particularly interesting given the time the data were collected – at a time when the economy in the Netherlands was flourishing. Taking this argument one step further, we also explored whether realistic threat narratives in times of economic prosperity differ from those put forward in times of economic decline. Whereas the latter are likely to focus on immigrants competing over scarce resources such as

Ó 2017 Hogrefe Publishing

jobs, education, and housing, the former should be concerned with (a) fear that prosperity that was secured might be lost if resources are shared with immigrants, and (b) perceived injustice of having to share collective resources that were obtained in the past with others who are new to the country. Here, we label this threat as realistic entitlement and define it as specific form of realistic threat located in notions of autochthony (see Martinovic & Verkuyten, 2013) – as fear that “our” resources, and resources that “we are entitled to” might be lost to newcomers.

The Present Research Does the economic performance in a country or society shape the type of argument generated to present a case that immigrants are a problem? There are a number of ways in which this question can be examined. For example, one can examine speeches of populist leaders and assess the type of anti-immigrant argument they evoke as a function of whether the economy is prosperous or declining. Even though such an analytic strategy has merit (and has been used in previous research, see Mols & Jetten, 2015, in press), there are some downsides to this as well. While such an analysis would allow for a rich and detailed examination, a focus on a number of small and specific cases (each with their own idiosyncratic features) may also limit the extent to which findings can be generalized to other contexts. To examine this question, we therefore decided to take a different approach. We conducted an experiment in which we asked participants to “stand in the shoes” of a speechwriter for a leader who promotes an anti-immigrant agenda. The participant’s job would be to generate the most compelling arguments for their leader that immigration should be curbed. In this way, we can use an experimental design to examine this question. Summarizing our predictions, we expect rather straightforwardly that in times of economic decline, realistic threat narratives should be seen as most compelling and they would therefore be mentioned more often than in times of economic growth. Perhaps more interesting, given that there are reality constraints to realistic threat narratives when the economy is prosperous compared to declining (Riek et al., 2006), we predicted that there would be a search for a different way to make a case for the anti-immigrant agenda. In times of economic growth, we examined the extent to which we would find more narratives focusing on (a) symbolic threat, and (b) realistic entitlement (compared to times of economic decline). To determine whether economic performance leads to a rise or fall of specific threat narratives, a control

Social Psychology (2017), 48(1), 40–46

condition was included where no information was provided about the nation’s economic performance.1

Method Participants Seventy-five undergraduate students were recruited at a large Australian University. The sample consisted of 58 females and 17 males with ages ranging from 17 to 36 (M = 19.45, SD = 3.58). Only Australian citizens were eligible to participate and participants received course credit.

Design The experiment utilized a one-way between-participants’ design and participants were randomly allocated to one of three conditions: economic decline (n = 25), economic growth (n = 25), or a baseline control condition (n = 25).

Materials and Measures Manipulation of Economic Performance Participants were asked to read fictional articles from the newspaper, “The Australian” (for a similar procedure, see Mols & Jetten, 2015, see Appendices A and B). In a third baseline condition, no information was given about economic performance. However, to ensure that national identity was equally salient in this baseline condition, participants were asked to list as many Australian states and territories as they could think of and to name two big Australian rivers. Manipulation Check To ensure participants comprehended the newspaper, once they had finished reading they were asked to recall

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

information from the article and were also asked to write down the main message of the article. Speech-Writing Task Participants were then instructed as follows: “For the next exercise, we would like you to step into the shoes of a speechwriter for an anti-immigrant advocate. In the space provided below, please spend 5 minutes writing a speech highlighting why immigrants should not be allowed into Australia.” Two independent coders who were blind to conditions assessed the statements generated in the speech-writing task. For each participant, they coded arguments as realistic threat, symbolic threat, realistic entitlement, or “other” (e.g., see Table 1). All statements categorized as “other” were irrelevant to the topic or summary statements. Two statements from one participant could not be interpreted due to poor handwriting, and thus were disregarded (< 1%). The total number of statements was 341, with 68.90% coded as realistic threat, 11.95% as symbolic threat, 7.95% as realistic entitlement, and 11.20% as “other.” Participants wrote a total of 2–12 arguments. Coders disagreed on the coding of 61 statements (17.89%) and differences were resolved through discussion. There were no differences between conditions in the number of “other” statements or the total number of arguments per participant (Fs < 1.17).

Results2 Manipulation Check All participants recalled correctly two critical pieces of information from the relevant articles, and all participants correctly recalled the main message of the article they were given.

Other measures were also included to examine more generally whether economic performance (perhaps in combination with the argument generation task) would affect the extent to which participants opposed immigration. Six items were included to measure opposition to immigration. Three items examined realistic threat (e.g., “Immigrants abuse the system of social benefits”) and three items measured symbolic threat (e.g., “The cultural practices of immigrants threaten the Australian way of life”). All responses were recorded on a 7-point scale (1 = Strongly disagree to 7 = Strongly agree). Items were averaged separately as a realistic threat (α = .93) and a symbolic threat measure (α = .92) and, given the high reliability, also combined into a general 6-item opposition to immigration measure (α = .92). We also measured the strength of national identification with four items adapted from Doosje, Ellemers, and Spears (1995). An example item is “Being Australian is an important part of how I see myself.” All responses were recorded on a 7-point scale (1 = Strongly disagree to 7 = Strongly agree, α = .87). We found that condition did not affect opposition to immigration due to symbolic threat perceptions, F(2, 72) = .36, p = .360, η2 = .010, overall mean M = 2.84, SD = 1.63, or for realistic threat perceptions, F(2, 72) = .18, p = .833, η2 = .01, overall mean M = 3.22, SD = 1.49. The effect on the combination of these two scales was not significant either, F(2, 72) = .15, p = .857, η2 = .00, overall mean M = 3.03, SD = 1.47. This suggests that people requested to craft a persuasive anti-immigration message will tailor their message to the economic context at the time, and are able to not be influenced themselves by how well they map arguments to the situation at hand. We found no evidence that our economic performance manipulation affected identification as Australian, F(2, 72) = 1.03, p = .362, η2 = .03 (overall identification, M = 5.57, SD = 1.26).

Social Psychology (2017), 48(1), 40–46

Ó 2017 Hogrefe Publishing

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

Table 1. Examples of arguments coded as realistic threat, symbolic threat, realistic entitlement, and “other” Realistic Threat

“Immigrants place a strain on the job market, make it harder to get jobs” “More money is being spent on immigrants: cost of health care, refugee camps, benefits given for lack of employment” “We can no longer let these outsiders take our own people’s jobs”

Symbolic Threat

“Immigrants pollute our core values with their own beliefs” “Most immigrants do not try to assimilate to the Western culture” “The foreigners will not accept the ‘Aussie’ way of life” “They (immigrants) want to change our values and ruin what has made Australia so great”

Realistic Entitlement

“Increased returns in the housing market should be made available to Australian citizens, not overseas immigrants” “Australia’s economic boom should be used for Australians so they can prosper” “We made this country fair and square – do you want them just to take it away from you?” “If we want Australian’s to have a chance in this market they should be given first priority”

Other

“Immigrants should not be allowed into Australia” “Immigrants have a negative impact on Australia”

Realistic Threat

Symbolic Threat We found that the percentage of symbolic threat arguments only marginally significantly differed by condition, F(2, 71) = 2.78, p = .069, η2 = .072. The mean percentage of symbolic threat generated in the economic growth condition (M = 14.64, SD = 17.95) was marginally

Ó 2017 Hogrefe Publishing

Realistic threat arguments

In line with our prediction, analysis of variance (ANOVA) on the percentage of realistic threat arguments generated showed a significant effect for condition, F(2, 71) = 8.67, p < .001, η2 = .194. Participants in the economic decline condition were most likely to generate realistic threat arguments (M = 83.60, SD = 17.50) compared to those in the economic growth condition (M = 53.98, SD = 30.08, p < .001). The baseline condition fell in between the economic decline and economic growth condition (M = 69.12, SD = 26.19), with significantly fewer realistic treat arguments generated in the baseline than in the economic decline condition (p = .046), and significantly more realistic treat arguments in the baseline than in the economic growth condition (p = .037, see Figure 1). This finding provides support for our prediction that realistic threat arguments “fit” economic decline contexts. Even though realistic threat arguments were still the dominant response in times of economic growth, relatively speaking, they were used less than in times of economic decline, presumably because they become less plausible in such contexts. To explore to what extent participants would turn to symbolic threat, or realistic entitlement arguments in economic growth conditions, three further analyses of variance (ANOVAs) were conducted.

80 70 60 50 40 30 20 10 0 Economic decline

Economic growth

Baseline

Figure 1. Mean percentage of realistic threat arguments as a function of economic performance.

significantly higher than in the economic decline condition (M = 4.93, SD = 10.31, p = .066, see Figure 2). Analysis also revealed that the mean percentage or symbolic threat arguments in the baseline condition (M = 16.29, SD = 24.21) was significantly higher than in the economic decline condition, p = .032, but somewhat contrary to our prediction, not significantly different from the economic growth condition, p = .752.

Realistic Entitlement Analyses of the realistic entitlement arguments generated did produce a significant condition effect, F(2, 71) = 7.09, p = .002, η2 = .165 (see Figure 3). Participants in the economic growth condition were most likely to generate realistic entitlement arguments (M = 17.73, SD = 26.68) – more so than those in the economic decline condition (M = 0.80, SD = 4.00, p = .001), and more so than those in the baseline condition (M = 5.52, SD = 9.25, p = .009). The economic decline and baseline condition did not differ from each other, p = .336.

Social Psychology (2017), 48(1), 40–46

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

Realistic entitlement arguments

Symbolic threat arguments

14 12 10 8 6 4 2 0

16 14 12 10 8 6 4 2 0

Economic decline

Economic growth

Baseline

Economic decline

Economic growth

Baseline

Figure 2. Mean percentage of symbolic threat arguments as a function of economic performance.

Figure 3. Mean percentage of realistic entitlement arguments as a function of economic performance.

Discussion

when it comes to thinking of compelling arguments why immigrants should not be allowed to share in societies’ wealth. Perhaps, our results show evidence of crafty entrepreneurship whereby new narratives are developed about why immigrants pose a threat. That is, compared to the decline or baseline condition, our analysis showed that participants resorted to another type of threat argument when the economy was expected to grow in the future: entitlement threat. Ironically, like realistic threat narratives, entitlement threat arguments were also concerned with themes relating to threat to resources. However, rather than arguing that immigrants bring unwanted competition, the reasoning focuses on the unfairness of having to share “our” resources with new arrivals, and the greater entitlement of “real” Australians (nativism or autochthony; see Martinovic & Verkuyten, 2013) to resources. Realistic entitlement arguments reflect concerns about future relative deprivation by those that are relatively gratified – a fear that resources will be lost to those that did not build the country (see also Mols & Jetten, 2014).

In our experiment, we asked our participants to generate arguments that they saw as compelling to stop immigration. We showed that participants intuitively understood that developing a compelling anti-immigration argument requires acting like a chameleon: adjusting the message and appearance to that context. Perhaps not surprisingly, we found that participants generated most realistic threat arguments when the economic climate made such realistic threat arguments most plausible – when the economy was perceived to be in decline. Even though the link between economic downturn and realistic threat is well established in the literature, this finding is unique as it is to our knowledge the first to reveal experimentally that economic decline encourages the generation of realistic threat narratives. However, the more interesting effect to emerge from this experiment is that, relatively speaking, these realistic threat arguments lose their appeal in times of economic prosperity – participants taking on the role of speechwriters were least likely to use realistic threat arguments when they expected economic prosperity (less so compared to an economic growth and a baseline condition). We therefore examined the extent to which when the economy is expected to boom, narratives emerge arguing that immigrants should be excluded because they pose a threat to identity and culture (i.e., symbolic threat) or that it would be unfair to have to share resources that were acquired in the past by Australian citizens (i.e., realistic entitlement). Somewhat unexpectedly, we did not find that symbolic threat arguments were particularly sought out when the economy was growing – there was no difference in the number of symbolic threat arguments generated in the economic growth condition and the baseline condition. It is unclear why we did not find more support for this prediction. Even though we can only speculate about this, one reason may be that in times of economic prosperity, symbolic threat arguments may not be the first port of call Social Psychology (2017), 48(1), 40–46

Limitations and Future Research Note too that in all conditions, many more realistic threat arguments were generated than other types of threat arguments. The prevalence of realistic threat arguments may be specific to the Australian context and participants’ overwhelming focus on realistic threat arguments may also explain why we found no difference between the percentage of symbolic threat arguments generated in the baseline and in the economic growth condition. Future research should therefore examine support for our hypotheses in contexts where symbolic threat narratives have historically been more prominent (e.g., The Netherlands). In future research, it would also be important to examine whether effects are stronger or weaker for particular groups of immigrants. For example, is it the case that realistic threat arguments are particularly used in times of economic Ó 2017 Hogrefe Publishing

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

downturn when considering immigrants who are less educated and skilled? And, are symbolic threat arguments used more in relation to highly skilled immigrants when the economy is prosperous? Future research should also focus on further examining whether the selection of particular arguments indeed reflected strategic considerations about what type of argument would best fit the context (as we suggest here). One way to assess this would be by examining the effectiveness of the arguments – whether attitude change among those exposed to arguments differs as a function of the economic performance condition.

In Sum For the first time, we showed experimentally that the perceived economic climate changes how people form arguments to convince others that immigration is a problem for society. By placing participants in the shoes of a speechwriter for an anti-immigrant leader we showed that such “speechwriters” use elements of the economic reality they find themselves in to create a compelling antiimmigrant argument. In line with a growing body of social identity literature on leadership and social influence (Haslam et al., 2011; Reicher & Hopkins, 2001), these findings fit with the notion that those who seek to influence others are “crafty identity entrepreneurs” who enhance the persuasiveness of their message by tailoring their messages strategically, thereby taking into account their knowledge of the needs and wants of their audience, and the overarching political and economic context. Acknowledgments This research was supported by the Australian Research Council’s Discovery Project funding scheme (DP120100053).

References Allport, G. W. (1954). The nature of prejudice. Reading, MA: Addison-Wesley. Arzheimer, K., & Carter, E. (2006). Political opportunity structures and right-wing extremist party success. European Journal of Political Research, 45, 419–443. doi: 10.1111/j.1475-6765. 2006.00304.x Betz, H. G. (1993). The new politics of resentment: Radical rightwing populist parties in Western Europe. Comparative Politics, 25, 413–427. doi: 10.2307/422034 Bizman, A., & Yinon, Y. (2001). Intergroup and interpersonal threats as determinants of prejudice: The moderating role of in-group identification. Basic and Applied Social Psychology, 23, 191–196. doi: 10.1207/153248301750433669 Dollard, J. (1938). Hostility and fear in social life. Social Forces, 17, 15–26. doi: 10.2307/2571143

Ó 2017 Hogrefe Publishing

Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psychology, 31(5), 410–436. doi: 10.1006/jesp.1995.1018 Esses, V. M., Dovidio, J. F., Jackson, L. M., & Armstrong, T. L. (2001). The immigration dilemma: The role of perceived group competition, ethnic prejudice, and national identity. Journal of Social Issues, 57, 389–412. doi: 10.1111/0022-4537.00220 Esses, V. M., Jackson, L. M., & Armstrong, T. L. (1998). Intergroup competition and attitudes toward immigrants and immigration: An instrumental model of group conflict. Journal of Social Issues, 54, 699–724. doi: 10.1111/0022-4537. 911998091 González, K. V., Verkuyten, M., Weesie, J., & Poppe, E. (2008). Prejudice towards Muslims in The Netherlands: Testing integrated threat theory. The British Journal of Social Psychology, 47, 667–685. doi: 10.1348/014466608X284443 Guimond, S., & Dambrun, M. (2002). When prosperity breeds intergroup hostility: The effects of relative deprivation and relative gratification on prejudice. Personality and Social Psychology Bulletin, 28, 900–912. doi: 10.1177/ 01467202028007004 Haslam, S. A., Reicher, S. D., & Platow, M. J. (2011). The new psychology of leadership: Identity, influence and power. New York, NY: Psychology Press. Jackman, R. W., & Volpert, K. (1996). Conditions favouring parties of the extreme right in Western Europe. British Journal of Political Science, 26, 501–521. doi: 10.1017/S0007123400007584 Kinder, D. R., & Sears, D. O. (1981). Prejudice and politics: Symbolic racism versus racial threats to the good life. Journal of Personality and Social Psychology, 40, 414–431. doi: 10.1037/ 0022-3514.40.3.414 Martinovic, B., & Verkuyten, M. (2013). “We were here first, so we determine the rules of the game”: Autochthony and prejudice towards out-groups. European Journal of Social Psychology, 43, 637–647. doi: 10.1002/ejsp.1980 McLaren, L. M. (2003). Anti-immigrant prejudice in Europe: Contact, threat perception, and preferences for the exclusion of migrants. Social Forces, 81, 909–936. doi: 10.1353/sof. 2003.0038 Mols, F., & Jetten, J. (2014). No guts, no glory: How framing the collective past paves the way for anti-immigrant sentiments. International Journal of International Relations, 43, 74–86. doi: 10.1016/j.ijintrel.2014.08.014 Mols, F., & Jetten, J. (2015). Explaining the appeal of populist right-wing parties in times of economic prosperity. Political Psychology, 37, 275–292. doi: 10.1111/pops.12258 Mols, F., & Jetten, J. (in press). The wealth paradox: Economic prosperity and the hardening of attitudes. Cambridge University Press. Quillian, L. (1995). Prejudice as a response to perceived group threat: Population composition and anti-immigrant and racial prejudice in Europe. American Sociological Review, 60, 586–611. doi: 10.2307/2096296 Reicher, S., & Hopkins, N. (2001). Self and nation. London, UK: Sage. Riek, B. M., Mania, E. W., & Gaertner, S. L. (2006). Intergroup threat and outgroup attitudes: A meta-analytic review. Personality and Social Psychology Review, 10, 336–353. doi: 10.1207/s15327957pspr1004_4 Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., & Sherif, C. W. (1961). Intergroup conflict and cooperation: The Robbers Cave experiment. (Vol. 10). Norman, OK: University Book Exchange. Sniderman, P. M., & Hagendoorn, L. (2007). When ways of life collide: Multiculturalism and its discontents in the Netherlands. Princeton, NJ: Princeton University Press.

Social Psychology (2017), 48(1), 40–46

Vanneman, R. D., & Pettigrew, T. F. (1972). Race and relative deprivation in the urban United States. Race, 13, 461–486. doi: 10.1177/030639687201300404 Walker, I., & Smith, H. J. (2002). Relative deprivation: Specification, development, and integration. Cambridge, NY: Cambridge University Press. Received June 2, 2016 Revision received July 18, 2016 Accepted August 4, 2016 Published online March 15, 2017 Jolanda Jetten School of Psychology University of Queensland Brisbane, QLD 4072 Australia j.jetten@psy.uq.edu.au

Appendix A Stable Economic Growth for Foreseeable Future, Says Queensland Study Australia’s economic situation seems likely to improve significantly in the coming years, according to a report published by economists and business experts at Brisbane’s University of Queensland. The report claims that, owing to Australia’s prosperity during the resources boom, economic growth is likely to continue strongly into the foreseeable future. Dr Jack Dalton, from the University of Queensland’s School of Economics, says that the benefits of the resource boom will be felt by many Australians and in many areas of life. “The fortunate position we find ourselves in on the global market is something that many of us will be taking advantage of in coming years,” he said. “For instance, one place we’re definitely going to see an improvement is the housing market. People beginning to invest in property within the next five years will see significantly large returns on their investments as property values continue to rise steeply.” Dr Dalton also believes consumer spending is set to increase, which will have a positive impact on small business across the country. “With an increase in overall economic markers, we’re also going to see people putting more back into the economy, which means small business will be thriving all over Australia.” The report also states that unemployment will also continue to fall, and that we are likely to see a long-lasting budget surplus. Stable returns in superannuation is also a likelihood.

Social Psychology (2017), 48(1), 40–46

J. Jetten et al., Economic Prosperity and Anti-Immigrant Sentiments

The report also stresses that Australia’s future economic growth is likely to be stable and long-lasting. “We have been very lucky during the GFC, and have managed to escape any significant financial downturn,” says Dr Dalton. “From now on, we’re really going to feeling seeing the benefits of our economic position.” “Basically, it’s good news all round.”

Appendix B Economic Downturn for Foreseeable Future, Says Queensland Study Australia’s economic situation seems likely to decline significantly in the coming years, according to a report published by economists and business experts at Brisbane’s University of Queensland. The report claims that, despite to Australia’s prosperity during the resources boom, economic growth is likely to experience decline in the foreseeable future. Dr Jack Dalton, from the University of Queensland’s School of Economics, says that the negative impacts of the Global Financial Crisis will be felt by many Australians and in many areas of life. “The unfortunate position we are going to find ourselves in on the global market is something that impact many of us in coming years,” he said. “For instance, one place we’re definitely going to see significant problems is the housing market. People beginning to invest in property within the next five years will see significantly diminished returns on their investments as property values continue to fall steeply.” Dr Dalton also believes consumer spending is set to decrease, which will have a negative impact on small business across the country. “With a decrease in overall economic markers, we’re also going to see people putting less back into the economy, which means small business will be struggling all over Australia.” The report also states that unemployment will also continue to rise, and that we are likely to see a long-lasting budget deficit. Unpredictability in superannuation is also a likelihood. The report also stresses that Australia’s future economic decline is likely to be relatively unstable and long-lasting. “We may have been lucky during the GFC, but the research shows that we’re really going to see a significant economic downturn,” says Dr Dalton. “From now on, we’re really going to start feeling the problems of our economic position.” “Basically, it’s bad news all round.”

Ó 2017 Hogrefe Publishing

Research Report

Glucose Increases Risky Behavior and Attitudes in People Low in Self-Control A Pilot Study Michaela Pfundmair, Eva Lermer, and Dieter Frey Department of Psychology, Ludwig-Maximilians-University of Munich, Germany

Abstract: People low in self-control have a strong proclivity toward risk-taking. Risk-taking behavior provides an opportunity to obtain some form of reward. Glucose, on the other hand, seems to facilitate reward and goal-directed behavior. In a pilot study executed in the laboratory, we investigated whether consuming a glucose drink would increase risky behavior and attitudes in people low in self-control. Our findings revealed that a dose of glucose compared to placebo increased risk-taking on a behavioral and cognitive level in participants low in selfcontrol but not in participants high in self-control. The findings may shed some light on the psychological underpinnings of glucose: By showing glucose’s association with high-risk behavior, they support the assumption of glucose driving a goal-directed motivation. Keywords: risk-taking, self-control, glucose, reward

In decisions to buy shares, considerations about touring ski destinations, or even thinking about taking an umbrella when leaving the house, human beings are constantly faced with risky decisions and behavior. Risktaking behaviors involve some potential for danger or harm and, at the same time, provide an opportunity to receive some form of reward (Leigh, 1999). Risky behavior is also related to the psychological dimension of sensation seeking: High sensation seekers tend to estimate risks as lower (Zuckerman, 2007). Some personality traits have shown to be more prone to risky behavior and attitudes than others, inter alia, selfcontrol, the capacity to bring own responses into line with ideals, values, morals, social expectations, and long-term goals (Baumeister, Vohs, & Tice, 2007). Self-control empowers a person to restrain or override responses (Baumeister et al., 2007). Plausibly, people low in selfcontrol – both when measured at trait level (Zuckerman & Kuhlman, 2000) and when state fluctuations are taken into account (Freeman & Muraven, 2010) – show increased risk-taking. This relationship can also be observed in everyday live: More impulsive (and thus less self-regulated) individuals show more drug and alcohol use (Wills, Sandy, & Yaeger, 2002), more dangerous driving (Vavrik, 1997), and a higher affinity for gambling (Martins, Tavares, da Silva Lobo, Galetti, & Gentil, 2004). Ó 2017 Hogrefe Publishing

As risk itself, also glucose, a carbohydrate used as cellular source of energy, is associated with reward: Recent literature suggests that a dose of glucose increases activation in the ventral striatum and anterior cingulate cortex, regions that enhance reward and goal-directed behavior (Chambers, Bridge, & Jones, 2009). Consistently, glucose consumption has been observed to increase reward- and goal-related behaviors, for example, control in dogs (Miller, Pattison, DeWall, Rayburn-Reeves, & Zentall, 2010) and working memory in human beings (Owen, Scholey, Finnegan, Hu, & Sünram-Lea, 2012). That is, brain regions responding to glucose, in particular the dopaminergic pathways within the striatum, seem to mediate emotional and behavioral reactions associated with rewarding stimuli (Berridge & Robinson, 1998; Kelley et al., 2002). It seems plausible that this glucose-induced reward motivation could also drive a preference for risks that per definition provide some form of reward (Leigh, 1999). A dose of glucose should therefore increase susceptibility to risk. Particularly people low in self-control who dispositionally have a strong proclivity toward risk-taking per se might be pushed by glucose to give in to their natural impulse. Consequently, we predicted that consuming a glucose (vs. placebo) drink would increase both behavioral and cognitive risk-taking in participants low in self-control. Social Psychology (2017), 48(1), 47–53 DOI: 10.1027/1864-9335/a000290

To investigate this hypothesis, participants drank lemonade sweetened with sugar or a substitute, filled out a measure of self-control, and completed a behavioral task as well as a cognitive measure of risk-taking.

Method Participants and Design One hundred twelve German undergraduates (74 female, 37 male, and 1 who did not specify gender; Mage = 25.05, SDage = 6.18) participated in this study. Participants were not allowed to take part if they reported to be diabetics. Participants were randomly assigned to a 2 (Beverage: glucose vs. placebo) between-subjects design; self-control served as continuous moderator variable.

Procedure Participants arrived at the laboratory for a study that ostensibly consisted of two independent experiments investigating a beverage’s appearance and taste, and responsiveness in playful situations. After completing informed consent, participants consumed either a glucose or a placebo beverage without any information about the drink. Subsequently, they answered questions about mood, self-control, and liking for the beverage. After this, they completed the behavioral risk measure and filled out a questionnaire consisting of the cognitive risk measure and mood.

Materials Beverage Participants drank 14 ounces of Kool-Aid or Sprite lemonade sweetened with either sugar (glucose drink) or a sugar substitute (placebo condition). We used Kool-Aid lemonade and Kool-Aid lemonade sweetened with Stevia in the first half of data collection. However, as Kool-Aid, which is not sold in Germany, is experienced as particularly unusual by German drinkers, we changed the beverage into Sprite and Sprite Zero in the second half of data collection. Both glucose drinks contained approximately 140 calories, both placebo drinks approximately 0 calory. Mood (Premeasure) Participants’ implicit affect was assessed by the Implicit Positive and Negative Affect Test (IPANAT; Quirin, Kazén, & Kuhl, 2009). They were instructed to rate how well six artificial words (e.g., “VIKES”), purportedly from an artificial language, express six different mood adjectives Social Psychology (2017), 48(1), 47–53

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

on 1 (= does not fit at allI) to 4 (= fits very well) response scales. We calculated the mean score of the positive (e.g., “pleased”; α = .66) and negative mood adjectives (e.g., “helpless”; α = .76). Self-Control Participants filled out a measure of dispositional self-control capacity (SCS; Bertrams & Dickhäuser, 2009). The 13 items (e.g., “I am good at resisting temptation”) answered on 1 (= not at all) to 5 (= very much) response scales were combined into a mean score (α = .81). Liking for the Beverage Among filler measures intended to bolster the cover story on taste and appearance of the drink, participants completed two measures of liking for the drink (“How likely would you drink the beverage again?”; “How well does the beverage score compared to your favorite beverage?”; r = .76) and one sweetness evaluation (“How sweet did you experience the beverage?”) on 1 (= not at all) to 5 (= very much) response scales. Behavioral Measure of Risk Participants’ level of behavioral risk-taking was assessed using the Balloon Analog Risk Task (BART; Lejuez et al., 2003). The BART is a computer program designed to simulate an actual risk-taking scenario. On a computer screen, participants are presented a simulated balloon, a button labeled Press this button to pump up the balloon, a button labeled Press to collect $$$, and a box labeled Total earned. Participants earned five virtual Cents each time they pumped the balloon; as in previous studies (e.g., Hiemer & Abele, 2012), the total task contained 20 trials. Each trial started with a deflated balloon in the center of the screen. Participants could inflate the balloon by clicking on the corresponding button. With each click, the balloon increased in size by approximately 0.3 cm. Participants could click on the Press to collect $$$ button at any point in the trial which transferred the money earned during the current trial to their permanent bank (the box labeled Total earned). Participants were also informed that the balloon may explode at any point in the trial. With the increase of pumps, each subsequent pump risked more money. A trial ended either when the participant pressed the collect button or when the balloon exploded. If a balloon exploded, all money from the current trial was lost. The number of pumps required to burst each of the 20 balloons was random but constant across participants. Following the recommendation of Lejuez et al. (2003), the number of pumps per balloon that did not burst was used as an indicator of willingness to take risks (ICC = .61; M = 14.78, SD = 11.51). Seven participants did not complete the BART because of technical difficulties. Ó 2017 Hogrefe Publishing

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

Cognitive Measure of Risk Participants’ propensity for risk-taking was assessed using the Sensation Seeking Scale form V (SSS-V; Zuckerman, 1994). The SSS-V consists of 40 dichotomous forced-choice items, each belonging to one of four subscales: thrill and adventure seeking (e.g., “A sensible person avoids activities that are dangerous – I sometimes like to do things that are a little frightening”; α = .62), experience seeking (e.g., “I like to explore a strange city or section of town by myself, even if it means getting lost – I prefer a guide when I am in a place I don’t know well”; α = .53), disinhibition (e.g., “I like ‘wild’ uninhibited parties – I prefer quiet parties with good conversation”; α = .67), and boredom susceptibility (e.g., “I get bored seeing the same old faces – I like the comfortable familiarity of everyday friends”; α = .53). The subscales were obtained by summing up the matching answers in terms of points. Due to their low reliabilities, we performed a factor analysis on the items of the sensation seeking scale which did not support four but 16 separate factors. At least, however, nine of 10 items of the disinhibition subscale loaded highly onto one single factor. Therefore, we decided to use only this subscale for the following analyses. Mood (Postmeasure) Participants’ affect was again assessed by the IPANAT. Reliabilities were acceptable for the mean score of the positive (α = .74) and negative mood adjectives (α = .83).

Results Preliminary Analyses To ensure equality between the Kool-Aid and Sprite beverage, we first investigated differences in initial mood and beverage evaluations using t-tests. Participants who drank Kool-Aid neither differed from participants who drank Sprite in their positive, t(110) = 1.24, p = .217, or negative mood, t(110) = 0.83, p = .410, nor in liking for the drink, t(110) = 0.65, p = .517, or the sweetness evaluation, t(110) = 0.43, p = .672. Thus, we combined the two kinds of drinks together. To moreover ensure equality between the glucose and placebo condition, we calculated some other t-tests. Participants who drank a glucose-laden drink neither differed from participants who drank the placebo in their initial positive, t(110) = 0.99, p = .324, or negative mood, t(110) = 0.71, p = .482, nor in liking for the drink, 1

t(110) = 0.57, p = .567, or the sweetness evaluation, t(110) = 0.28, p = .784. Furthermore, they did not differ in their level of self-control, t(110) = 0.66, p = .511. Behavioral Risk-Taking To investigate the prediction that risky behavior would be increased in participants low in self-control when drinking a glucose-laden drink, we calculated a mixed-effects model (random intercept, no nested structure) entering beverage (effect coded as +1 = glucose and 1 = placebo), self-control (mean centered), the 20 BART trials (coded from 0 to 19) and their interaction terms as independent variables, and the number of balloon pumps (log transformed) as dependent variable; the procedural difference (change of beverage) was modeled as additional factor (effect coded as +1 = Sprite and 1 = Kool-Aid). There was a significant main effect of trial, see Table 1, and, notably, a significant interaction between beverage and self-control, see Table 2, τ2 = .30, residual = .44. This interaction showed that participants low in self-control chose more pumps in the glucose than in the placebo condition, whereas participants high in self-control did not differ in their number of pumps between the glucose and placebo condition, see Figure 1A. No other significant effects emerged. Cognitive Risk-Taking To investigate the predicted relationship between glucose and sensation seeking in participants low in self-control, we conducted a multiple moderation analysis: We entered beverage (effect coded as +1 = glucose and 1 = placebo) and self-control (mean centered) into a regression model using Hayes’ (2013) Process tool; the procedural difference was modeled as additional factor (effect coded as +1 = Sprite and 1 = Kool-Aid). For disinhibition as dependent variable, there was a significant main effect of self-control which was qualified by a significant interaction between beverage and self-control, see Table 3. This interaction showed that participants low in selfcontrol indicated more disinhibition in the glucose than in the placebo condition, whereas participants high in self-control did not differ in their disinhibition ratings between the glucose and placebo condition, see Figure 1B. No other significant effects emerged.1 (See Electronic Supplementary Material, ESM 1 for analyses on the other sensation seeking subscales.)

We also investigated the predicted relationship between glucose and risk-taking in participants low in self-control using another statistical approach: For behavioral risk-taking, we entered beverage (effect coded as +1 = glucose and 1 = placebo), self-control (standardized) and their interaction term as independent variables into a linear regression; the mean number of balloon pumps served as dependent variable. As in the current analysis, we observed a significant interaction between beverage and self-control, p = .041. For cognitive risk-taking, we used the same linear regression, entering disinhibition as dependent variable. Again, a significant interaction between beverage and self-control emerged, p = .022. Although the current analyses differed from these in several aspects (treatment of the self-control and BART variables, inclusion of the procedural difference as additional factor), the result pattern remained stable.

Ó 2017 Hogrefe Publishing

Social Psychology (2017), 48(1), 47–53

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

Table 1. Means and standard deviations (in parentheses) for balloon pumps on each trial, averaged across participants Trial number

Mean 17.00 12.48 14.39 16.13 11.10 11.63 13.23 11.81 12.08 12.33 13.53 12.41 13.14 11.97 14.40 11.81 13.03 13.01 12.47 14.05 pumps (16.65) (9.44) (11.44) (13.86) (6.06) (7.42) (9.69) (8.15) (9.25) (6.93) (13.58) (8.57) (8.61) (6.93) (9.84) (6.36) (9.61) (7.32) (8.44) (9.21)

Table 2. Results for the mixed-effects model with beverage, self-control, trial, the procedural difference, and all interaction terms as predictors of behavioral risk-taking SE

Beverage

Estimate 0.22

.19

1.15

.251

Self-Control

0.11

.20

0.54

.587

Trial

0.01

.004

1,041

3.04

.002

Procedural difference

0.19

.17

1.11

.271

Beverage Self-Control

0.67

.33

2.02

.047

Beverage Trial

0.01

.01

1,041

1.10

.270

Beverage Procedural Difference

0.13

.26

0.52

.605

Self-Control Trial

0.01

.01

1,041

0.72

.469

Self-Control Procedural Difference

0.06

.27

0.21

.838

Trial Procedural Difference

0.01

.01

1,041

1.26

.208

Beverage Self-Control Trial

0.01

.01

1,041

0.56

.574

Beverage Self-Control Procedural Difference

0.29

.42

0.69

.490

Beverage Trial Procedural Difference

0.00

.01

1,041

0.03

.974

Self-Control Trial Procedural Difference

0.00

.01

1,041

0.44

.662

Beverage Self-Control Trial Procedural Difference

0.00

.02

1,041

0.11

.911

Mood (Postmeasure) For positive mood, the same moderation analysis revealed neither a significant main effect of self-control, b = 0.02, SE = 0.06, t(103) = 0.26, p = .795, beverage, b = 0.02, SE = 0.04, t(103) = 0.61, p = .540, or the procedural difference, b = 0.01, SE = 0.04, t(103) = 0.36, p = .718, nor a Significant Beverage Self-Control, b = 0.02, SE = 0.06, t(103) = 0.29, p = .771, Beverage Procedural Difference, b = 0.02, SE = 0.04, t(103) = 0.42, p = .673, Self-Control Procedural Difference, b = 0.04, SE = 0.06, t(103) = 0.66, p = .510, or Beverage Self-Control Procedural Difference interaction, b = 0.05, SE = 0.06, t(103) = 0.81, p = .419. Similarly, no main effects of self-control, b = 0.11, SE = 0.07, t(103) = 1.53, p = .129, beverage, b = 0.02, SE = 0.04, t(103) = 0.38, p = .703, or the procedural difference, b = 0.07, SE = 0.04, t(103) = 1.48, p = .141, nor a Significant Beverage Self-Control, b = 0.02, SE = 0.07, t(103) = 0.26, p = .794, Beverage Procedural Difference, b = 0.07, SE = 0.04, t(103) = 1.59, p = .116, Self-Control Procedural Difference, b = 0.05, SE = 0.07, t(103) = 0.68, p = .498, or Beverage SelfControl Procedural Difference interaction, b = 0.03, SE = 0.07, t(103) = 0.44, p = .662, emerged for negative mood. The pilot study’s script and datafile are accessible under Electronic Supplementary Materials (ESM 2 and 3). Social Psychology (2017), 48(1), 47–53

Discussion Our findings revealed that glucose increases risk-taking in people low in self-control, not only on a cognitive but also on a behavioral level. This effect did not emerge in people high in self-control and could not be explained by mood. The results may shed some light on the effect of glucose in human behavior. Previous research has reasoned that glucose may be the physical basis of self-control (Gailliot et al., 2007). A growing body of literature, however, is sceptical of links between glucose and self-control (e.g., Kazén, Kuhl, & Leicht, 2015; Lange & Eggert, 2014). Recent work has suggested that the role of glucose in self-control is one of allocation and not of depletion (Beedie & Lane, 2012). Other work has argued that glucose is an input to decision-making systems (Kurzban, 2010). Empirical studies have supported this assumption: Specifically, glucose has shown to work through the activation of brain regions that are involved in reward (Chambers et al., 2009). It appears to increase a central drive or motivation (Carter, Jeukendrup, Mann, & Jones, 2004) by unconsciously suppressing fatigue signals (Pottier, Bouckaert, Gilis, Roels, & Derave, 2010). In our pilot study, the boosted susceptibility to reward-related behavior through a dose of glucose might have been reflected in increased risk-taking. These effects only emerged among people low in self-control. According to the dual system perspective Ó 2017 Hogrefe Publishing

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

(A)

(B) 7.5 7

Disinhibition

Mean pumps (20 rounds)

18 16 14 12 10

6.5 6 5.5 5 4.5 4

Low self-control

High self-control Placebo

Low self-control

High self-control

Glucose

Figure 1. The interaction of Beverage Self-Control influences risky behavior (A) and attitudes (B).

Table 3. Results for the multiple regression analysis with beverage, self-control, the procedural difference, and all interaction terms as predictors of cognitive risk-taking SE

Beverage

Estimate 0.32

.23

104

1.40

.165

Self-Control

1.20

.38

104

3.18

.002

Procedural difference

0.03

.23

104

0.14

.893

Beverage Self-Control

0.90

.38

104

2.36

.020

Beverage Procedural Difference

0.07

.23

104

0.32

.746

Self-control Procedural Difference

0.57

.38

104

1.50

.135

Beverage Self-Control Procedural Difference

0.53

.38

104

1.41

.161

(Hofmann, Friese, & Strack, 2009), certain boundary conditions can shift the degree of potential activation in favor of an impulsive or reflective reaction. In our pilot study, glucose might have pushed people low in self-control who have a strong proclivity toward risk-taking per se (e.g., Zuckerman & Kuhlman, 2000) to give in to their natural impulse. However, the individual difference also allows for another interpretation: Recently, it has been proposed that socially oriented people conserve more energy as they construe relationships as opportunities to conserve resources (Beckes & Coan, 2011; see also Fitzsimons, Finkel, & vanDellen, 2015). Socially avoidant individuals, on the other hand, who normally have lower self-regulatory resources (e.g., Fuendeling, 1998) have been shown to devote higher levels of glucose to their bloodstream and consume more glucose with the expectation of increased personal effort (Ein-Dor et al., 2015). That is, highly attachment avoidant (and thus lower self-regulated) people are physiologically adapted to a relatively independent way of life. Accordingly, it is possible that people low in self-control in our study have chosen more extreme investments under glucose to manage the situation individually. Ă&#x201C; 2017 Hogrefe Publishing

A large body of research, mainly stemming from research on prospect theory (Kahneman & Tversky, 1979), has shown the importance of considering gain versus loss frames in risk-taking. The increased susceptibility to risky behavior by consuming glucose has been observed in a gain frame in our pilot study: When playing the BART, participants were only able to win but not to lose. As glucose seems to enhance reward-directed behavior (Chambers et al., 2009), the increase of risky behavior in the gainframed paradigm of BART was expectable. On the other hand, glucose should not increase but even decrease risktaking in loss frames as those situations might not be able to provide any form of reward. Therefore, future research contrasting gain and loss frames by using different risktaking paradigms would move us toward a better understanding of glucoseâ&#x20AC;&#x2122;s proposed mechanism. Besides the behavioral measure of risk, we observed the predicted pattern on a cognitive risk-taking measure, namely the sensation seeking subscale of disinhibition. This preference for ignoring social constraints seems to be particularly associated with risk-taking: Disinhibition significantly contributes to patterns of risky behavior Social Psychology (2017), 48(1), 47â&#x20AC;&#x201C;53

(Greene, Krcmar, Walters, Rubin, & Hale, 2000). It should, however, be noted that the other sensation seeking subscales were excluded from our main analyses as they yielded relatively low reliabilities and a factor analysis did not support the assumed factor structure. Thus, considering one of these subscales as “the” focal scale has to be treated with some caution. In our interdisciplinary pilot study, we combined a wellvalidated behavioral measure with a self-report questionnaire and observed consistent effects. It should, however, be taken into account that our sample size was rather small for the present design. Thus, a replication of this effect in future research is crucial and could give more confidence in our conclusions. Glucose’s boosting role in risk-taking could explain behaviors from risky salary negotiations to dangerous driving after an intake of sugar. Moreover, it offers an important piece of the puzzle to understand the monosaccharide’s psychological underpinnings.

Acknowledgments The authors wish to thank Alexander Bauer and Helmut Küchenhoff for statistical consulting at STABLAB. The authors declare that there are no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Electronic Supplementary Material The electronic supplementary material is available with the online version of the article at http://dx.doi.org/10.1027/ 1864-9335/a000290 ESM 1. Text (.doc). Additional analyses. ESM 2. Text (.doc). Questionnaires used in the study. ESM 3. Data (.xls). Raw data of the participants.

References Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control. Current Directions in Psychological Science, 16, 351–355. doi: 10.1111/j.1467-8721.2007.00534.x Beckes, L., & Coan, J. A. (2011). Social baseline theory: The role of social proximity in emotion and economy of action. Social and Personality Psychology Compass, 5, 976–988. doi: 10.1111/ j.1751-9004.2011.00400.x Beedie, C. J., & Lane, A. M. (2012). The role of glucose in selfcontrol another look at the evidence and an alternative conceptualization. Personality and Social Psychology Review, 16, 143–153. doi: 10.1177/1088868311419817 Berridge, K. C., & Robinson, T. E. (1998). What is the role of dopamine in reward: Hedonic impact, reward learning, or

Social Psychology (2017), 48(1), 47–53

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

incentive salience? Brain Research Reviews, 28, 309–369. doi: 10.1016/S0165-0173(98)00019-8 Bertrams, A., & Dickhäuser, O. (2009). Messung dispositioneller Selbstkontroll-Kapazität [Measuring dispositional self-control capacity]. Diagnostica, 55, 2–10. doi: 10.1026/ 0012-1924.55.1.2 Carter, J. M., Jeukendrup, A. E., Mann, C. H., & Jones, D. A. (2004). The effect of glucose infusion on glucose kinetics during a 1-h time trial. Medicine and Science in Sports and Exercise, 36, 1543–1550. doi: 10.1249/01.MSS.0000139892.69410.D8 Chambers, E. S., Bridge, M. W., & Jones, D. A. (2009). Carbohydrate sensing in the human mouth: Effects on exercise performance and brain activity. The Journal of Physiology, 587, 1779–1794. doi: 10.1113/jphysiol.2008.164285 Ein-Dor, T., Coan, J. A., Reizer, A., Gross, E. B., Dahan, D., Wegener, M. A., . . . Zohar, A. H. (2015). Sugarcoated isolation: Evidence that social avoidance is linked to higher basal glucose levels and higher consumption of glucose. Frontiers in Psychology, 6, 492. doi: 10.3389/fpsyg.2015.00492 Fitzsimons, G. M., Finkel, E. J., & vanDellen, M. R. (2015). Transactive goal dynamics. Psychological Review, 122, 648–673. doi: 10.1037/a0039654 Freeman, N., & Muraven, M. (2010). Self-control depletion leads to increased risk taking. Social Psychological and Personality Science, 1, 175–181. Fuendeling, J. M. (1998). Affect regulation as a stylistic process within adult attachment. Journal of Social and Personal Relationships, 15, 291–322. doi: 10.1177/0265407598153001 Gailliot, M. T., Baumeister, R. F., DeWall, C. N., Maner, J. K., Plant, A., Tice, D. M., . . . Schmeichel, B. J. (2007). Self-control relies on glucose as a limited energy source: Willpower is more than a metaphor. Journal of Personality and Social Psychology, 92, 325–336. doi: 10.1037/0022-3514.92.2.325 Greene, K., Krcmar, M., Walters, L. H., Rubin, D. L., & Hale, L. (2000). Targeting adolescent risk-taking behaviors: The contributions of egocentrism and sensation-seeking. Journal of Adolescence, 23, 439–461. doi: 10.1006/jado. 2000.0330 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press. Hiemer, J., & Abele, A. E. (2012). High power = Motivation? Low power = Situation? The impact of power, power stability and power motivation on risk-taking. Personality and Individual Differences, 53, 486–490. doi: 10.1016/j.paid.2012.04.008 Hofmann, W., Friese, M., & Strack, F. (2009). Impulse and selfcontrol from a dual-systems perspective. Perspectives on Psychological Science, 4, 162–176. doi: 10.1111/j.1745-6924. 2009.01116.x Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263–291. doi: 10.2307/ 1914185 Kazén, M., Kuhl, J., & Leicht, E. M. (2015). When the going gets tough. . .: Self-motivation is associated with invigoration and fun. Psychological Research, 79, 1064–1076. doi: 10.1007/ s00426-014-0631-z Kelley, A. E., Bakshi, V. P., Haber, S. N., Steininger, T. L., Will, M. J., & Zhang, M. (2002). Opioid modulation of taste hedonics within the ventral striatum. Physiology & Behavior, 76, 365–377. doi: 10.1016/S0031-9384(02)00751-5 Kurzban, R. (2010). Does the brain consume additional glucose during self-control tasks? Evolutionary Psychology, 8, 244–259. doi: 10.1177/147470491000800208 Lange, F., & Eggert, F. (2014). Sweet delusion. Glucose drinks fail to counteract ego depletion. Appetite, 75, 54–63. doi: 10.1016/ j.appet.2013.12.020

Ó 2017 Hogrefe Publishing

M. Pfundmair et al., Risk-Taking, Self-Control, And Glucose

Leigh, B. C. (1999). Peril, chance, and adventure: Concepts of risk, alcohol use and risky behavior in young adults. Addiction, 94, 371–383. doi: 10.1046/j.1360-0443.1999.9433717.x Lejuez, C. W., Aklin, W. M., Jones, H. A., Richards, J. B., Strong, D. R., Kahler, C. W., & Read, J. P. (2003). The Balloon Analogue Risk Task (BART) differentiates smokers and nonsmokers. Experimental and Clinical Psychopharmacology, 11, 26–33. doi: 10.1037/1064-1297.11.1.26 Martins, S. S., Tavares, H., da Silva Lobo, D. S., Galetti, A. M., & Gentil, V. (2004). Pathological gambling, gender, and risk-taking behaviors. Addictive Behaviors, 29, 1231–1235. doi: 10.1016/ j.addbeh.2004.03.023 Miller, H. C., Pattison, K. F., DeWall, C. N., Rayburn-Reeves, R., & Zentall, T. R. (2010). Self-control without a “self”? Common self-control processes in humans and dogs. Psychological Science, 21, 534–538. doi: 10.1177/0956797610364968 Owen, L., Scholey, A. B., Finnegan, Y., Hu, H., & Sünram-Lea, S. I. (2012). The effect of glucose dose and fasting interval on cognitive function: A double-blind, placebo-controlled, six-way crossover study. Psychopharmacology, 220, 577–589. doi: 10.1007/s00213-011-2510-2 Pottier, A., Bouckaert, J., Gilis, W., Roels, T., & Derave, W. (2010). Mouth rinse but not ingestion of a carbohydrate solution improves 1-h cycle time trial performance. Scandinavian Journal of Medicine & Science in Sports, 20, 105–111. doi: 10.1111/j.1600-0838.2008.00868.x Quirin, M., Kazén, M., & Kuhl, J. (2009). When nonsense sounds happy or helpless: The Implicit Positive and Negative Affect Test (IPANAT). Journal of Personality and Social Psychology, 97, 500–516. doi: 10.1037/a0016063

Ó 2017 Hogrefe Publishing

Vavrik, J. (1997). Personality and risk-taking: A brief report on adolescent male drivers. Journal of Adolescence, 20, 461–465. doi: 10.1006/jado.1997.0100 Wills, T. A., Sandy, J. M., & Yaeger, A. M. (2002). Moderators of the relation between substance use level and problems: Test of a self-regulation model in middle adolescence. Journal of Abnormal Psychology, 111, 3–21. doi: 10.1037/0021-843X. 111.1.3 Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation seeking. Cambridge, UK: Cambridge University Press. Zuckerman, M. (2007). Sensation seeking and risky behavior. Washington, DC: American Psychological Association. Zuckerman, M., & Kuhlman, D. M. (2000). Personality and risktaking: Common biosocial factors. Journal of Personality, 68, 999–1029. doi: 10.1111/1467-6494.00124 Received January 9, 2016 Revision received September 5, 2016 Accepted September 5, 2016 Published online March 15, 2017 Michaela Pfundmair Department of Psychology Ludwig-Maximilians-University of Munich Leopoldstr. 13 80802 Munich Germany michaela.pfundmair@psy.lmu.de

Social Psychology (2017), 48(1), 47–53

Replication

Projection of Visceral Needs Satisfaction of the Need and Similarity of the Target Person as Moderators Janet N. Ahn,1 Gabriele Oettingen,2,3 and Peter M. Gollwitzer2,4 1

Psychology Department, William Paterson University, Wayne, NJ, USA

Psychology Department, New York University, NY, USA

Psychology Department, University of Hamburg, Germany

Psychology Department, University of Konstanz, Germany Abstract: People’s immediate needs influence how they perceive other’s needs – people project their needs onto others. Two studies sought to replicate and extend previous research by identifying moderators of need projection: need satisfaction and a global sense of similarity. Projection increased when a strong need to quench one’s thirst was induced and not satisfied (Study 1). Projection of the need to quench one’s thirst increased when existing thirst and the global sense of similarity with the target person were high, but again only when the need was not satisfied (Study 2). Visceral needs influence judgments about other people’s needs when the individual is experiencing the need in a given moment; lingering effects are not observed once that need is satisfied. Keywords: projection, visceral needs, need satisfaction, global similarity

Experiencing visceral needs (e.g., hunger pains) or psychological needs (e.g., loneliness) orient people to thoughts and behaviors related to fulfilling or satisfying those needs (Aarts, Dijksterhuis, & De Vries, 2001; Ariely & Loewenstein, 2006; Atkinson & McClelland, 1948; Balcetis & Dunning, 2010; Kappes, Schwörer, & Oettingen, 2012; Loewenstein, 1996; Przybylski, Murayama, DeHaan, & Gladwell, 2013; Risen & Critcher, 2011; Veltkamp, Aarts, & Custers, 2008). Needs even affect how we interpret other people’s internal states (O’Brien & Ellsworth, 2012; Van Boven & Loewenstein, 2003). Van Boven and Loewenstein (2003) observed that people’s assumptions of how others feel in need-related situations are based on how they think they themselves would feel, given that they were in those situations. However, there is limited empirical research identifying the specific conditions that influence people’s views of others’ visceral needs. O’Brien and Ellsworth (2012) demonstrated that need projection is curbed when individuals perceive target persons as dissimilar on a specific dimension (e.g., political orientation). We want to build on this research because we think that a person’s sensitivity to others’ needs has important interpersonal implications. For instance, need projection can be a mechanism that could either elicit or hinder empathy. Viewing others as possessing the same needs could elicit empathic feelings toward them (Hodges & Wegner, 1997). On the other hand, Social Psychology (2017), 48(1), 54–59 DOI: 10.1027/1864-9335/a000286

viewing others as not possessing the same needs might undermine need projection and thus empathy (as suggested by Nordgren, McDonnell, & Loewenstein, 2011; O’Brien & Ellsworth, 2012). The present research sought to conceptually replicate and extend past relevant findings in three ways. First, we tested whether the satisfaction of a current need would curb projection. As needs are thought to arise as a result of deprivation, they should stimulate activities and behaviors that reduce or eliminate the deprivation through satisfaction (e.g., Hull, 1943, 1952; Spence, 1956; for reviews, see e.g., Gollwitzer, Kappes, & Oettingen, 2011; Kenrick, Griskevicius, Neuberg, & Schaller, 2010; McClelland, Atkinson, & Clark, 1949; Reeve, 2008). Thus, once a need is met, that need state as well as the satiated state should no longer be activated and therefore not projected. Second, we investigated whether having a global sense of similarity to a target person would moderate the projection of needs just as having a specific sense of similarity has been previously shown to be a moderator. O’Brien and Ellsworth (2012) noted that being similar or dissimilar in terms of political orientation can be “affectively charged and can activate strong ingroup/outgroup membership or ally versus enemy dynamics.” We tested whether having a general sense of similarity with another person, defined as “an idiosyncratic and subjective sense that one is similar Ó 2017 Hogrefe Publishing

J. N. Ahn et al., Projection of Visceral Needs

to a target group/person” (Ames 2004a, 2004b), would also moderate need projection. And finally, we tested whether people project their needs onto real people as they have been shown to project their needs onto hypothetical target persons. Prior research showed that people project their needs onto fictitious characters in hypothetical situations (e.g., hikers lost in the woods). If more individuating information about target persons is provided, projection might be hampered (or enhanced) because people can use such information to infer either more similarity or dissimilarity with the person, and thereby project more or less. Thus, we examined whether individuals would also project their needs onto real target persons.

Study 1: Need Satisfaction as Moderator Method Participants and Design One hundred twenty-six participants (104 females; Mage = 19.26, SDage = 1.19) from a large university in the US completed a paper-pencil study alone in an experimental room to fulfill a partial requirement for an introductory psychology course. The study utilized a 2 (Thirst: weak vs. strong) 2 (Water provided: yes vs. no) factorial design. Procedure Adapted from prior research (e.g., Aarts et al., 2001; Kappes et al., 2012), participants were given a plate containing two differently flavored jelly beans and two pieces of chocolate. Half the participants were asked to eat three salty crackers after each of the two food samples supposedly to cleanse their palates. For these participants, the resultant consumption of six crackers was thought to create a strong need to quench their thirst, as eating salty crackers can be expected to make people thirsty. The other half of participants never consumed any crackers – thus these participants were not made thirsty.1 Then, half of the participants who were made thirsty and half of those who were not made thirsty were given a cup of water to drink (8 ounces): “You may drink some water now before answering the rest of the questionnaire. Please drink all of the cup’s content.” The other half of participants (of those who were made thirsty and those who were not made thirsty) were instead told: “You may drink some 1

water after answering the rest of the questionnaire. You may drink all of the cup’s content after you are done.” As a manipulation check, participants indicated their thirst level on an answer scale ranging from 1 (= not at all) to 7 (= extremely).2 As the dependent variable, we assessed the projection of being thirsty (see Gollwitzer et al., 2011) by asking participants two items: “How badly do you think other students [on campus] want to drink water right now?” and “How committed do you think other students [on campus] are to quenching their thirst right now?” using a 1 (= not at all) to 7 (= extremely) scale (α = .79).

Results and Discussion We conducted a 2 (Thirst: weak vs. strong) 2 (Water Provided: yes vs. no) ANCOVA on participants’ assumptions regarding other students’ thirst, adjusting for participants’ baseline thirst level and found the predicted interaction effect, F(1, 121) = 6.25, p = .014, η2p = .05, observed power = .70. As shown in Figure 1, participants in the strong thirst/water not provided condition (M = 4.84, SD = 1.45) indicated that other students were thirstier than participants in all other conditions – the strong thirst/water provided condition (M = 3.65, SD = 1.23), t(60) = 3.50, p = .001, d = .89; the weak thirst/water provided condition (M = 3.55, SD = .95), t(62) = 4.23, p < .001, d = 1.07; and the weak thirst/water not provided condition (M = 3.66, SD = .93), t(60) = 3.81, p < .001, d = .98. There were no differences among these three latter conditions, ps > .74. As noted in the Electronic Supplementary Material, ESM 1, the same effect was observed without controlling for participants’ baseline thirst level.

Study 2: Need Satisfaction and Perceived Similarity as Conjoint Moderators Method Participants and Design Forty-two participants (18 females and one specified as “other”; Mage = 33.35, SDage = 16.06) were invited to a paper-pencil study at a major park in a large city. Participants were randomly assigned to either the need unsatisfied condition (n = 18) or the need satisfied condition (n = 24).

As a baseline, we measured participants’ thirst level (as well as various other states such as hunger and fatigue) before any experimental manipulations to ensure there were no preexisting differences. A 2 (Thirst: weak vs. strong) 2 (Water provided: yes vs. no) ANOVA on the experienced thirst level confirmed that our manipulation of thirst was effective (see full report of analysis in ESM).

Ó 2017 Hogrefe Publishing

Social Psychology (2017), 48(1), 54–59

J. N. Ahn et al., Projection of Visceral Needs

Weak Need Strength

Strong Need Strength

Table 1. Study 2: Means, standard deviations, and correlations among key variables for need unsatisfied and satisfied conditions Mean

Need unsatisfied condition

Inferred Thirst

Amount target person would pay more ($) Existing thirst level

1.56

0.88

1.27

0.47

Target similarity

3.04

1.81

.61**

.07 .11

Need satisfied condition

Amount target person would pay more ($) Existing thirst level

1.13

0.33

1.14

0.34

Target similarity

3.83

2.12

.02

.28 .37

Notes. The dollar amount the target person would pay for water was participants’ indication of how thirsty they perceived the target person to be. Existing thirst level was the participants’ indication of his/her commitment to quench his/her thirst. Similarity was the participants’ indication of the target person’s similarity to themselves.

2 Water Provided

No Water Provided

Need Satisfaction

Figure 1. Other students on campus: Inferred thirst as a function of induced need (weak vs. strong) and need satisfaction (water provided vs. not provided) (Study 1).

Procedure Participants’ thirst was not induced (as was done in Study 1) but instead measured indirectly by the following question: “Knowing that the price of an 8 oz bottle of water is approximately $1.00, how much more than the regular price of water would you pay to quench your thirst right now?” using a 1 (= $1.00 more) to 7 (= $4.00 more) scale in increasing increments of $0.50. At this point, the experimenter chose a target person who was in close proximity. A target person was anyone who was passing by a bench that was 10 feet in front of the water station. Participants first indicated how similar they viewed that person to themselves: “In general, how similar do you think that person is to you?” using a 1 (= not at all) to 7 (= extremely) scale. Then, participants in the need satisfied condition received an 8 oz bottle of water to drink and were told: “You may drink some water now before answering the rest of the survey. Please drink at least half of its content.” The participants in the need unsatisfied condition were given an 8 oz bottle of water and told: “You may drink some water after answering the rest of the survey. You may drink its content after you are done.” We assessed the projection of being thirsty indirectly by asking participants about the target person’s wanting to satisfy his/her thirst: “Knowing that the price of an 8 oz 3

bottle of water is approximately $1.00, how much more than the regular price of water would that person pay to quench his/her thirst right now?” using a 1 (= $1.00 more) to 7 (= $4.00 more) scale in increasing increments of $0.50 (see Table 1 for means and standard deviations for the respective conditions).

Results and Discussion Projecting Onto Other People at the Park as a Function of Condition Because of our small sample size and concerns of violation of the assumptions of regression models, we applied a robust estimator3 on our analysis and found a significant threeway interaction among these variables, b = .65, SE = .27, Wald X2 = 5.98, p = .01, observed power = .18. To clarify this three-way interaction effect, we analyzed the two-way interactions of participants’ existing thirst and perceived similarity separately for the respective conditions. A visual representation of the raw data is provided in ESM 1. Need Unsatisfied Condition We conducted a generalized linear model (GLM) analysis and entered participants’ existing thirst level, the perceived similarity of the target person, and the interaction term to predict the target person’s thirst in the need unsatisfied condition. We observed a significant interaction effect on the assumed amount of money the target person would be willing to pay for an 8 oz bottle of water, b = .77, SE = .18, Wald X2 = 18.81, p < .001. As illustrated on the

A robust estimator is an estimation technique designed to circumvent some limitations of traditional parametric methods, making it insensitive to small departures from the idealized assumptions, which have been used to optimize the algorithm. Thus, when regression models are vulnerable to outliers and not particularly robust to suspicions of heteroscedasticity (as is true for small sample sizes), a robust model is more appropriate (Cohen, Cohen, West, & Aiken, 2003).

Social Psychology (2017), 48(1), 54–59

Ó 2017 Hogrefe Publishing

J. N. Ahn et al., Projection of Visceral Needs

(A)

Need Satisfied Condition

(B)

Need Unsatisfied Condition

Strong Need Strength

Weak Need Strength

2.5

Inferred Thirst

Weak Need Strength

1.5 1 0.5

1.5 1 0.5 0

0 Low -0.5

Strong Need Strength

Figure 2. Another person at the park: Inferred thirst as a function of existing need (weak or strong) and the perceived similarity of the target person (low or high) by need satisfaction condition (Study 2).

Low

High

Perceived Similarity of Target Person

-0.5

Perceived Similarity of Target Person

right side of Figure 2, participants who had strong existing thirst were more likely to think that the target person was also thirsty the more similar they perceived that person to be, b = .47, Wald X2 = 17.68, p = .001. There was an analogous pattern for those who had weak existing thirst – they were less likely to think the target person was thirsty the more similar they perceived that person to themselves, b = .17, Wald X2 = 17.84, p = .001. Need Satisfied Condition We applied the same analysis to predict projection in the need satisfied condition and did not observe any interaction effect, b = .12, SE = .20, Wald X2 = .38, p < .53 (see Figure 2A). This suggests that need strength and perceived similarity no longer affected projection for participants who had quenched their thirst.

Discussion Van Boven and Loewenstein (2003) demonstrated that people project their visceral need states (hunger and thirst) onto others, biased in the direction of their currently deprived state. Going beyond these findings, we considered whether the projection of a given need would be effectively diminished once the need is satisfied. In Study 1, we observed that when participants were made thirsty by consuming salty crackers but could not quench their thirst, they projected their thirst onto other students. On the other hand, those who were made thirsty but quenched their thirst did not project. Moreover, participants who were never made thirsty in the first place (did not consume any crackers) did not project at all, whether they were provided with water or not. These findings demonstrate that it is important to have the active experience of the need in the moment to elicit projection. Ó 2017 Hogrefe Publishing

High

O’Brien and Ellsworth (2012) demonstrated that having a specified sense of dissimilarity from a hypothetical target person (in their study, in terms of political orientation) hampers projection. In Study 2, a field study, we tested whether having a global sense of similarity (or dissimilarity) with the target person would moderate projection effects just as it has been shown for a specific sense of similarity/dissimilarity. Additionally, we tested whether need projection would also occur for real people and not just for hypothetical target persons. We observed that when people’s needs were satisfied, they did not project onto the target person whether or not they had experienced a strong (or weak) existing need or perceived the person to be similar (or not). That is, the strength of people’s needs and the global sense of similarity to the target person do not seem to matter when needs are satisfied. In contrast, people who did not drink water (did not have their need satisfied) and had a strong existing need to drink water did project their need onto the target person the more similar in general they perceived themselves to that person. These findings suggest that need projection occurs for real people (not just hypothetical persons) when the existing strength of the need and the global similarity of the target are high, but these findings are only obtained when the need is not satisfied. Interestingly, an analogous pattern was observed for people who did not drink water and had a weak existing need – they projected that the target person also had a weak need to the degree they perceived themselves as similar. This latter finding emphasizes the point that when people do not experience the need at all, they believe others do not experience that need either. In sum, our findings of Studies 1 and 2 imply that unsatisfied needs in our everyday lives lead to inferring that similar others possess those same needs. Importantly, visceral needs influence judgments only when someone is actually experiencing that need in the moment but do not have lingering effects once that need is satisfied. However, Social Psychology (2017), 48(1), 54–59

there are limitations of the present studies. First, both studies did not have adequate power (especially Study 2), and thus future studies should opt for much larger sample sizes. Second, because perceived similarity was defined in the most general sense, it is difficult to know which aspects of similarity are most influential. Future research might want to disentangle the specific aspects of similarity/ dissimilarity that most effectively moderate need projection. Our findings are important because in everyday life, having certain needs may lead to inappropriate assumptions about other people’s needs with unfortunate consequences for interpersonal sensitivity and interpersonal interactions. For instance, parents and doctors may gauge their young children’s or patients’ hunger, thirst, fatigue, and bodily temperature based on how hungry, thirsty, tired, and warm/cold they themselves feel at the given moment. As a consequence, they may lose sight of their children’s and patients’ actual needs. Acknowledgments We thank the Motivation Lab at NYU for helpful comments on an earlier version of this article. The help of Brittan Badenhop with collecting and coding the data is gratefully acknowledged. We also appreciate the help of Erin P. Hennes and Sean P. Lane on data analysis. Electronic Supplementary Material The electronic supplementary material is available with the online version of the article at http://dx.doi.org/10.1027/ 1864-9335/a000286. ESM 1. Text and Figure (PDF). Additional analyses (manipulation check and an additional measure of hunger) for Study 1 are provided. And, a visual representation of the raw data for Study 2 is also provided.

References Aarts, H., Dijksterhuis, A., & De Vries, P. (2001). On the psychology of drinking: Being thirsty and perceptually ready. British Journal of Psychology, 92, 631–642. doi: 10.1348/ 000712601162383 Ames, D. R. (2004a). Inside the mind reader’s tool kit: Projection and stereotyping in mental state inference. Journal of Personality and Social Psychology, 87, 340–353. doi: 10.1037/ 0022-3514.87.3.340 Ames, D. R. (2004b). Strategies for social inference: A similarity contingency model of projection and stereotyping in attribute prevalence estimates. Journal of Personality and Social Psychology, 87, 573–585. doi: 10.1037/0022-3514. 87.5.573 Ariely, D., & Loewenstein, G. (2006). The heat of the moment: The effect of sexual arousal on sexual decision making. Journal

Social Psychology (2017), 48(1), 54–59

J. N. Ahn et al., Projection of Visceral Needs

of Behavioral Decision Making, 19, 87–98. doi: 10.1002/ bdm.501 Atkinson, J. W., & McClelland, D. C. (1948). The projective expression of needs. II: The effect of different intensities of the hunger drive on thematic apperception. Journal of Experimental Psychology, 38, 643–658. doi: 10.1037/h0061442 Balcetis, E., & Dunning, D. (2010). Wishful seeing: More desired objects are seen as closer. Psychological Science, 21, 147–152. doi: 10.1177/0956797609356283 Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum. doi: 10.4324/ 9780203774441 Gollwitzer, P. M., Kappes, H. B., & Oettingen, G. (2011). Needs and incentives as sources of goals. In H. Aarts & A. Elliot (Eds.), Goal-directed behavior (pp. 115–149). New York, NY: Psychology Press. doi: 10.4324/9780203869666 Hodges, S. D., & Wegner, D. M. (1997). Automatic and controlled empathy. In W. J. Ickes (Ed.), Empathic accuracy (pp. 311–339). New York, NY: Guilford. Hull, C. L. (1943). The problem of intervening variables in molar behavior theory. Psychological Review, 50, 273–291. doi: 10.1037/h0057518 Hull, C. L. (1952). A behavior system: An introduction to behavior theory concerning the individual organism. New Haven, CT: Yale University Press. Kappes, H. B., Schwörer, B., & Oettingen, G. (2012). Needs instigate positive fantasies of idealized futures. European Journal of Social Psychology, 42, 299–307. doi: 10.1002/ ejsp.1854 Kenrick, D. T., Griskevicius, V., Neuberg, S. L., & Schaller, M. (2010). Renovating the pyramid of needs: Contemporary extensions build upon ancient foundations. Perspectives on Psychological Science, 5, 292–314. doi: 10.1177/ 1745691610369469 Loewenstein, G. (1996). Out of control: Visceral influences on behavior. Organizational Behavior and Human Decision Processes, 65, 272–292. doi: 10.1006/obhd.1996.0028 McClelland, D. C., Atkinson, J. W., & Clark, R. A. (1949). The projective expression of needs: III. The effect of egoinvolvement, success, and failure on perception. Journal of Psychology: Interdisciplinary and Applied, 27, 311–330. doi: 1290585496 Nordgren, L., McDonnell, M., & Loewenstein, G. (2011). What constitutes torture? Psychological impediments to an objective evaluation of interrogation tactics. Psychological Science, 22, 689–694. doi: 10.1177/0956797611405679 O’Brien, E., & Ellsworth, P. C. (2012). More than skin deep: Visceral states are not projected onto dissimilar others. Psychological Science, 23, 391–396. doi: 10.1177/0956797611432179 Przybylski, A. K., Murayama, K., DeHaan, C. R., & Gladwell, V. (2013). Motivational, emotional, and behavioral correlates of fear of missing out. Computers in Human Behavior, 29, 1841–1848. doi: 10.1016/j.chb.2013.02.014 Reeve, J. (2008). Understanding motivation and emotion. Hoboken, NJ: Wiley. Risen, J. L., & Critcher, C. R. (2011). Visceral fit: While in a visceral state, associated states of the world seem more likely. Journal of Personality and Social Psychology, 100, 777–793. doi: 10.1037/a0022460 Spence, K. W. (1956). Behavior theory and conditioning. New Haven, CT: Yale University Press. doi: 10.1037/10029-000 Van Boven, L., & Loewenstein, G. (2003). Social projection of transient drive states. Personality and Social Psychology Bulletin, 29, 1159–1168. doi: 10.1177/0146167203254597

Ó 2017 Hogrefe Publishing

J. N. Ahn et al., Projection of Visceral Needs

Veltkamp, M., Aarts, H., & Custers, R. (2008). Perception in the service of goal pursuit: Motivation to attain goals enhances the perceived size of goal-instrumental objects. Social Cognition, 26, 720–736. doi: 10.1521/soco.2008.26.6.720 Received February 23, 2016 Revision received July 2, 2016 Accepted July 8, 2016 Published online March 15, 2017

Ó 2017 Hogrefe Publishing

Janet N. Ahn Department of Psychology William Paterson University of New Jersey 300 Pompton Road Wayne, NJ, 07470 USA ahnj9@wpunj.edu

Social Psychology (2017), 48(1), 54–59

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

Social Psychology (2017), 48(1)

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

Ó 2017 Hogrefe Publishing

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

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

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

www.hogrefe.com

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

How information concerning the social context affects motivation and behavior Topics covered include • Achievement motivation as a function of assimilation and differentiation needs • Differential gender and ethnic differences in math performance: a self-regulatory perspective • The proactive control of stereotype activation: implicit goals to not stereotype • Willful stereotype control: the impact of internal motivation to respond without prejudice on the regulation of activated stereotypes • Strategic control over the unhelpful effects of primed social categories and goals • Interactive self-regulation during mate searching: reciprocal romantic interest increases attention allocation to opposite-sex others

Kai Sassenberg / Jennifer Fehr (Editors)

Self-Control and Self-Regulation in Social Contexts (Series: Zeitschrift für Psychologie – Vol. 220) 2012, iv + 56. pp., large format US $49.00 / € 34.95 ISBN 978-0-88937-439-3 Over the years, a sophisticated scientific tool box full of theories and models facilitating our understanding of the motivational dynamics in humans has been developed that has recently also advanced our understanding of social phenomena by linking the intrapersonal motivational processes to characteristics of the

www.hogrefe.com

social context and its influences on behavior. Considering both, the intrapersonal processes and the social context of motivation, facilitates both - (a) the understanding of the sources and the influence of motivation as well as (b) the understanding of social phenomena.