Sjp 2016 75 issue 1

Volume 75 / Number 1 / 2016

Swiss Journal of Psychology

Editor-in-Chief GrĂŠgoire Zimmermann Associate Editors Thierry Lecerf Nicolas Rothen Christian StaerklĂŠ

Swiss Journal of Psychology Official Publication of the Swiss Psychological Society

Volume 75, Issue 1, January 2016 Editor-in-Chief GrĂŠgoire Zimmermann, Lausanne

Swiss Journal of Psychology © 2016 Hogrefe

Editor-in-Chief

Grégoire Zimmermann, Lausanne

Associate Editors

Thierry Lecerf, Geneva Nicolas Rothen, Bern Christian Staerklé, Lausanne

Assistants to the Editors

Tamara Herz, Zurich Grégory Mantzouranis, Lausanne, gregory.mantzouranis@chuv.ch

Advisory Board

Fabrizio Butera, Lausanne Franz Caspar, Bern Paolo Ghisletta, Geneva Alexander Grob, Basel Claude-Alain Hauert, Geneva Ulrich Hoffrage, Lausanne Lutz Jäncke, Zurich Martin Kleinmann, Zurich Petra Klumb, Fribourg Klaus Opwis, Basel Sonja Perren, Konstanz Diego Pizzagalli, Cambridge, MA Pasqualina Perrig-Chiello, Bern Jérôme Rossier, Lausanne Willibald Ruch, Zurich Klaus Scherer, Geneva Franziska Tschan, Neuchâtel

Publisher

Hogrefe AG, Postfach, Länggass-Strasse 76, CH-3000 Bern 9, Tel. +41 31 300 45 00, Fax +41 31 300 45 93 verlag@hogrefe.ch, www.hogrefe.com

Support

The Swiss Journal of Psychology is supported by a contribution from the Swiss Academy of Humanities and Social Sciences

Media Manager

Josef Nietlispach, Hogrefe AG, Postfach, Länggass-Strasse 76, CH-3000 Bern 9, Tel. +41 31 300 45 00, Fax +41 31 300 45 91, inserate@hogrefe.ch

Typesetting

Satzspiegel, Nörten-Hardenberg, Germany

Printing Office

AZ Druck und Datentechnik GmbH, Kempten, Germany

ISSN

ISSN-L 1421-0185, 1421-0185 (Print), 1662-0879 (Online)

Frequency

4 issues per year

Subscriptions

Annual subscription rates: Libraries / Institutions CHF 320.– / e 249.–.

Individuals CHF 127.– / e 94.– plus postage and handling. For members of the Swiss Psychological Society the subscription rate is included in their annual membership.* Single issue CHF 53.50 / e 40.– plus postage and handling

Indexing

Swiss Journal of Psychology is abstracted/indexed in Social Sciences Citation Index (SCIE), Social Scisearch, Current Contents/Social and Behavioral Sciences, Journal Citation Reports/Social Sciences Edition, PSYCLIT (Psychological Abstracts), PSYNDEX, PsycINFO, PsyJOURNALS. Europ. Reference List for the Humanities (ERIH), IBZ, IBR, and Scopus.

Impact Factor: 0.778 2014 Journal Citation Reports® Social Sciences Edition (Thomson Reuters, 2015)

Electronic Full Data

www.psyjournals.com *To become a member of the Swiss Psychological Society, please apply to the President, Sabine Sczesny, SGP/SSP, Institute für Psychologie, Universität Bern, Muesmattstrasse 45, CH-3000 Bern 9.

Swiss Journal of Psychology © 2016 Hogrefe

Contents Swiss Journal of Psychology (2016), 75 (1)

Original Communications

Short Research Note

© 2016 Hogrefe

The Look of Fear from the Eyes Varies with the Dynamic Sequence of Facial Actions Eva G. Krumhuber and Klaus R. Scherer

5

Social Asymmetries and Anonymity in Dyadic Computer-Mediated Communication: An Experimental Study of Gender Perception Jérôme Guegan, Pascal Moliner, and Laurent Milland

15

Self-Reported Compliance with Traffic Rules in a Sample of Iranian Preschoolers: Knowledge of Rules, Perception of Danger, Moral Judgment, and Self-Regulation Zahra Tabibi, Fatemeh Grayeli, and Mohammad Saeid Abdekhodaei

25

Effect of Political Orientation on Judgment of Agency, Competence, Morality, and Sociability: The French Presidential Election of 2012 Astrid Mignon, Patrick Mollaret, Odile Rohmer, and Céline Bagès

35

Do Ambiguous Normative Ingroup Members Increase Tolerance for Deviants?

Ana C. Leite, Isabel R. Pinto, and José M. Marques

47

Swiss Journal of Psychology (2016), 75 (1)

E. G. Krumhuber SwissJournal & K. R. of Scherer: Psychology Sequential (2016), © 2016 75Dynamics (1), Hogrefe 5–14

Original Communication

The Look of Fear from the Eyes Varies with the Dynamic Sequence of Facial Actions Eva G. Krumhuber1 and Klaus R. Scherer2 1

Department of Experimental Psychology, University College London, UK

2

Swiss Center for Affective Sciences, University of Geneva, Switzerland

Abstract. Most research on the ability to interpret expressions from the eyes has utilized static information. This research investigates whether the dynamic sequence of facial actions in the eye region influences the judgments of perceivers. Dynamic fear expressions involving the eye region and eyebrows were created which systematically differed in the sequential occurrence of facial actions. Participants rated the intensity of sequential fear expressions, either in addition to a simultaneous, full-blown expression (Experiment 1) or in combination with different levels of eye gaze (Experiment 2). The results showed that the degree of attributed emotion and the appraisal ratings differed as a function of the sequence of facial expressions of fear, with direct gaze resulting in stronger subjective responses. The findings challenge current notions surrounding the study of static facial displays from the eyes and suggest that emotion perception is a dynamic process shaped by the time course of the facial actions of an expression. Possible implications for the field of affective computing and clinical research are discussed.

Keywords: emotion, facial expression, dynamic, sequence, eye gaze

The eye region attracts particular attention in social interaction. People look more frequently and for longer durations at the eyes of a person than at any other facial area (Langton, Watt, & Bruce, 2000). This is not surprising given that the eyes provide a rich source of emotional and motivational information. Past research showed that young infants and adults can detect complex mental states and emotions solely from the eye region and the direction of gaze (e.g., Adams & Kleck, 2005; Back, Ropar, & Mitchell, 2007; Baron-Cohen & Cross, 1992; Graham & LaBar, 2007; Smith, Cottrell, Gosselin, & Schyns, 2005).

The ability to identify the psychological state of others from features of the upper face is an important skill and has been assessed in various ways, for example, using the Reading the Mind in the Eyes Test (Baron-Cohen, Jolliffe, Mortimore, & Robertson, 1997; Baron-Cohen, Wheelwright, Hill, Raste, & Plumb, 2001).

This test presents participants with photographs of a person’s eye region and asks them to identify, from a list of response options, what the person is feeling or thinking. Apart from specific criticisms of the Eyes Test as a measure of social intelligence (see Johnston, Miles, & McKinlay, 2008), most research to date has focused on static information from the eyes (e.g., Adolphs et al., 2005; Baron-Cohen, Campbell, Karmiloff-Smith, Grant, & Walker, 1995; Calvo & Fernández-Martin, 2013; Eisenbarth & Alpers, 2011; Hopkins et al., 2011; Joseph & Tanaka, 2003). The present research goes beyond that approach by exploring dynamic expressions of the eye region.

Growing evidence points to the importance of movement in © 2016 Hogrefe DOI 10.1024/1422-4917/a000166

emotion perception, with significant recognition benefits afforded by dynamic as compared to static stimuli (see Krumhuber, Kappas, & Manstead, 2013, for a review). Such an advantage is particularly evident in individuals with markedly impaired processing of facial expressions from the eyes, including gaze (i.e., people with autism spectrum disorder; see Back et al., 2007; Gepner, Deruelle, & Grynfeltt, 2001; Uono, Sato, & Toichi, 2009). Dynamic information may convey vital cues that enable greater specification of the target’s emotional state. Although the beneficial effect of moving displays is well established, little is known about the ways in which dynamic facial actions convey information over time as part of complex expression patterns. In this research, we investigate the role played by dynamic sequences of facial actions in the perception of emotion expression from the eyes.

Componential appraisal theories (CAT) provide a central framework for studying expressive patterns in a dynamic context (Ellsworth & Scherer, 2003; Scherer, 2001; Smith & Scott, 1997). According to CAT, facial expressions are determined by a set of evaluations or appraisal processes that reflect how we interpret particular dimensions of stimulus meaning (i.e., novelty, pleasantness, and goal conduciveness; Scherer, 1984). Depending on the outcome of the appraisal, different facial actions are elicited. Consequently, individual components contributing to expressive patterns carry meaning because they are intrinsically linked with the appraisal dimension. Moreover, as a result of the ongoing cognitive processes of appraisal (Scherer, 2009), the constituent facial actions may occur at different Swiss Journal of Psychology (2016), 75 (1), 5–14

6

time points. Facial expressions can therefore be seen as a micro-sequence of partial expressive patterns that reach the apex in a cumulative-sequential fashion (Scherer & Ellgring, 2007).

This assumption of sequential accumulation of appraisal-driven facial actions differs from classical viewpoints (i.e., basic emotion theory, BET; Ekman, 2003; Ekman & Friesen, 1982) that consider (genuine) expressions as full-blown patterns that wax and wane with the simultaneous onset of all facial actions.

There is evidence to support the sequential processing of emotion-relevant stimuli (Aue, Flykt, & Scherer, 2007; Delplanque et al., 2009; Lanctôt & Hess, 2007). By applying fine-grained analyses to the time course of expressions, Fiorentini, Schmidt, and Viviani (2012) as well as Krumhuber and Scherer (2011) showed that facial actions unfolded sequentially and converged toward an apex in an asynchronous manner. This sequential structure of emotionally expressive displays was also found by Pilowsky, Thornton, and Stokes (1986), With and Kaiser (2011), and most recently by Jack, Garrod, and Schyns (2014), indicating reliable and identifiable patterns of facial movements. Furthermore, sequences of facial actions have been found to be related to emotion judgments made by observers (Fiorentini et al., 2012; With & Kaiser, 2011). Specifically, facial expressions that were rated high on an emotion (i.e., enjoyment, surprise, or sadness) consisted of distinct sequences of facial actions. The emotional meaning of expressions therefore seems to derive from their dynamic organization, with individual facial actions becoming gradually integrated over time.

Although previous research has systematically studied the effects of sequential information on the recognition of emotioneliciting situations (Scherer, 1999), evidence regarding the perception of dynamic facial emotional expressions is still scarce.

Wehrle, Kaiser, Schmidt, and Scherer (2000) studied participants’ ability to identify emotions in facial expressions that unfolded simultaneously versus sequentially and found equivalent rates of recognition accuracy. Similarly, Malatesta, Raouzaiou, Karpouzis, and Kollias (2009) investigated the perception of hot anger and fear using an additive versus sequential approach in the temporal unfolding of synthesized expressions. However, in both studies, based on theoretical predictions (i.e., CAT), sequentiality was operationalized by only one facial expression sequence. Moreover, emotion perception was measured as the accuracy in selecting the correct emotion label for each facial expression. Clearly, this restricts the variability in perceived emotional meaning and does not allow for systematic exploration of the effects of dynamic sequences in general. As facial actions can combine in a number of ways, a large range of possible sequences may have an impact on the perceiver.

Present Research This research investigated whether there is a specific affective quality in different facial sequences which influences perceivSwiss Journal of Psychology (2016), 75 (1), 5–14

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

ers’ judgments. Specifically, we wanted to know whether the perceived intensity of the expressive display varies as a function of the dynamic sequence. To test this, we systematically manipulated the order of occurrence of facial actions in fear expressions. Fear is of particular interest, as it is not only assumed to be a universally recognizable, basic emotion (e.g., Ekman, 1989, 1992), but can also be detected from minimal cues in the upper face. Specifically the eye region, including the eyebrows, has been shown to be a crucial area for fear perception (i.e., Bombari et al., 2013; Ekman, 1979; Morris, deBonis, & Dolan, 2002; Vuilleumier, 2005; Whalen et al., 2004). In addition, eye gaze modulates the perceived meaning of fear displays, with stronger and more accurate ratings when gaze is averted from the observer (Adams & Kleck, 2003, 2005; Milders, Hietanen, Leppänen, & Braun, 2011; N’Diaye, Sander, & Vuilleumier, 2009; Sander, Grandjean, Kaiser, Wehrle, & Scherer, 2007).

By focusing on this partial facial area, we aimed to test whether relevant information for fear processing can be extracted from the eye region alone.

Fear in the upper face is signaled by eyebrow raising accompanied by brow furrowing and upper eyelid raising (Ekman, Friesen, & Hager, 2002). These facial actions carry emotionspecific information (Ekman, 1979) and also reflect certain appraisal dimensions (Scherer, 1984; Scherer & Ellgring, 2007).

For example, according to CAT, eyebrow raising is characteristic of an appraisal of novelty, brow furrowing reflects an appraisal of goal obstruction, and eyelid raising indicates power/control appraisal (Scherer & Ellgring, 2007). If these facial actions convey specific meaning, the order of their occurrence should play a significant role in how fear is perceived. As such, expressions that unfold in a sequence postulated by CAT (i.e., novelty → goal conduciveness → coping potential) should increase perceptions of fear and have an expressive value similar to that of full-blown expressions (see Wehrle et al., 2000) compared to other types of sequential orders. We therefore predicted that the degree of attributed emotion and the appraisal dimension in fear expressions would vary depending on the dynamic sequence.

We report on two experiments in which we systematically examined the effects of sequentiality on expression perception.

In Experiment 1, we focused on a wide range of facial sequences and investigated whether these differentially influence perceivers in addition to seeing a full-blown expression with simultaneous onset of facial actions. In all cases, the final peak display was the same and consisted of a prototypical fear expression. In Experiment 2, we used a subset of fear sequences from the first experiment that were maximally discriminative in their sequential occurrence and tested their effect in combination with different levels of eye gaze. Given that gaze direction has been shown to moderate the perceptual clarity of fearful expressions (e.g., Adams & Kleck, 2005; Milders et al., 2011; Sander et al., 2007) we predicted that eye gaze would have different implications for different facial sequences.

In both experiments, participants gave intensity ratings that © 2016 Hogrefe

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

7

Table 1. Means and standard errors for dependent measures as a function of sequential or simultaneous dynamics of facial action units (Experiment 1) and with different levels of eye gaze (Experiment 2) Dependent Measures Suddenness Experiment Stimulus 1 (N = 52)

2 (N = 41)

Anger

Fear

Sadness

Action Unit M

SE

M

SE

M

SE

M

SE

Sequential Seq. 1

1+2, 4, 5

5.50ab

0.17

2.81b

0.24

5.38a

0.19

3.06ab

0.27

Seq. 2

1+2, 5, 4

5.71a

0.17

3.06b

0.29

4.48b

0.26

2.44bc

0.21

Seq. 3

5, 1+2, 4

5.69a

0.17

2.73b

0.21

4.71b

0.28

2.83abc

0.24

Seq. 4

5, 4, 1+2

5.42ab

0.15

3.92a’

0.26

4.33b

0.25

2.44c

0.24

Seq. 5

4, 1+2, 5

5.17b

0.20

3.13b

0.24

5.35a

0.18

2.38c

0.26

Seq. 6

4, 5, 1+2

5.15b

0.20

4.58a’

0.25

4.60b

0.26

2.36c

0.23

Simultaneous

1+2+4+5

5.10b

0.19

1.79c

0.17

5.50a

0.20

3.23a

0.27

Seq. 1

1+2, 4, 5

5.39ab

0.17

2.93b

0.29

5.38a

0.21

2.87a

0.29

Seq. 3

5, 1+2, 4

5.44a

0.16

2.61b

0.27

4.89b

0.24

3.10a

0.28

Seq. 6

4, 5, 1+2

5.01b

0.13

4.26a

0.24

4.13c

0.27

2.38b

0.27

Seq. 1

1+2, 4, 5

5.20ab

0.12

2.92b

0.23

4.54a

0.18

2.72a

0.23

Seq. 3

5, 1+2, 4

5.41a

0.14

2.48c

0.25

4.60a

0.19

2.81a

0.25

Seq. 6

4, 5, 1+2

5.00b

0.13

4.03a

0.22

4.07b

0.23

2.28b

0.20

Direct Gaze

Averted Gaze

Note. All ratings were given on Likert scales ranging from 1 to 7, with higher scores indicating higher levels on the given dimension. Column means not sharing a common subscript differ significantly at p < .05, with the exception of the means labeled a’, which differ marginally p < .06.

targeted major appraisal criteria (i.e., suddenness, pleasantness, goal obstruction; see Scherer, 1999) and emotions (i.e., fear, anger, surprise, sadness), thereby allowing us to analyze complex emotion judgments that go beyond simple recognition accuracy. Studying these issues requires fine control of the time course and intensity of dynamic facial actions and gaze direction. We therefore used facial synthesis in the form of recently developed animation software called FACSGen (Krumhuber, Tamarit, Roesch, & Scherer, 2012; see N’Diaye et al., 2009, for a similar approach). FACSGen enables the systematic and highly controlled activation of facial actions based on the Facial Action Coding System (FACS; Ekman et al., 2002). Using this software, we were able to conduct the first exploratory analysis of the effects of dynamic fear expressions with direct and averted gaze, in which the sequence of facial actions was systematically varied.

simultaneous expression in which all facial actions started at the same time. The participants rated each dynamic stimulus with respect to intensity of expression of various cognitive and emotion dimensions.

Method Participants A total of 52 students (45 women, 7 men) from a university in Switzerland participated in exchange for course credit or 15 Swiss francs. Their mean age was 21.48 years (SD = 2.42), ranging from 19 to 33 years.

Stimulus Material

Experiment 1 Experiment 1 examined the impact of different forms of facial sequences on perception. This required the creation of dynamic fear expressions that represented every possible sequence in which the relevant facial actions could cumulatively combine.

These sequential fear expressions were shown in addition to a © 2016 Hogrefe

Two synthetic male faces were selected from the face database (Oosterhof & Todorov, 2008). Both faces had been randomly generated using the FaceGen 3.1 modeller and achieved similar ratings of trustworthiness around the midpoint of the 9-point scale (M = 4.86 & M = 4.90; see Oosterhof & Todorov, 2008).

For each face, dynamic fear expressions involving only the eye region and eyebrows were created using FACSGen 2.0 software (Krumhuber et al., 2012). FACSGen is built on top of Swiss Journal of Psychology (2016), 75 (1), 5–14

8

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

Figure 1. Sequential (a–c) and simultaneous (4) cumulation of facial actions over time resulting in the final peak expression. Action Units (AU): 1+2 (eyebrow raise), 4 (brow furrowing), 5 (eyelid raising).

FaceGen and allows the generation of facial expressions in the form of action units (AUs). The targeted fear expressions were based on prototypes defined by Ekman et al. (2002) and consisted of eyebrow raising (AU1+2), brow furrowing (AU4), and eyelid raising (AU5).

To construct dynamic sequences of fear expressions, we systematically varied the order of occurrence of these facial actions, thereby creating fear sequences that differed with respect to the pattern in which the respective AUs merged over time (see Figure 1). For example, based on CAT predictions, one fear sequence commenced with eyebrow raising (AU1+2), followed cumulatively by brow furrowing (AU4) and eyelid raising (AU5), while another sequence commenced with brow furrowing (AU4), followed cumulatively by eyelid raising (AU5) and eyebrow raising (AU1+2). There were six possible sequences of fear (see Table 1).

In each sequence, the first facial action started at the beginning of the sequence at a neutral position and reached its apex after 1.4 s. To ensure a smooth cumulative transition between the occurrence of facial actions, each consecutive AU started at half of the onset duration (i.e., after 700 ms) of the previous AU and reached its apex after 1.4 s. This way, we could vary the sequentiality of the onset of the facial actions while preserving the dynamic trajectory of the expression. The final peak expression was the same for all six sequences and showed AU1+2 at target intensities of 0.70 and 1.00, respectively; AU4 at a target intensity of 0.41; and AU5 at a target intensity of 0.80.

In addition to the six fear sequences, a simultaneous expression of fear was generated in which, based on BET predictions, all of the facial actions started at the same time (i.e., AU1+2+4+5 from the beginning of the sequence) and were animated simultaneously to evolve into the final pattern of fear.

The six sequential expressions and the simultaneous expression of fear were displayed for the same amount of time and lasted a total of 2.8 s.

To ensure sufficient stimulus variability, we embedded the seven target expressions in a set of seven distractor stimuli consisting of various upper facial actions such as eyelid drop (AU43), eye squinting (AU44), lid tightening (AU7), outer brow raising (AU2), eyebrow raising (AU1+2), eyelid raising (AU5), and brow furrowing (AU4). The distractor stimuli were also displayed in dynamic form from onset to apex phase and lasted a total of 2.1 s.1 The 28 dynamic stimuli resulting from the combination of two faces with seven target fear expressions and seven distractor expressions were edited in Adobe After Effects and displayed as movie clips (800 × 1032 pixels) on a black background.

Procedure and Design Using Eprime 2.0.8 presentation software (Psychology Software Tools, Inc.), participants were informed that they were

1 All target and distractor stimuli were subject to informal pretests in which we explored various durations and transition periods between the facial actions. A stimulus duration of 2.8 s was chosen for the sequential and simultaneous expressions that fell within the recommended 4-s timespan of natural expressions (Ekman & Friesen, 1982) and guaranteed comparable viewing times of both types of target expressions. Because distractor stimuli consisted mainly of one facial action, a slightly shorter duration was chosen to ensure sufficient plausibility in expression unfolding. Given that distractor stimuli were not subject to statistical analysis, we considered this approach preferable in terms of stimulus quality.

Swiss Journal of Psychology (2016), 75 (1), 5–14

© 2016 Hogrefe

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

going to see short video clips of an animated character showing various facial movements. They were further told that only the upper face, including the eyes and eyebrows, would be visible and that they would never see a full facial display of the person.

Their task was to rate the facial movement of the person with respect to several emotions and cognitive dimensions. Specifically, we wanted to know what the participants could tell about the experience the person had when making the facial movement.

Because computer-animated faces were used, we told the participants that the character’s movements would consist of a resynthesis of facial expressions that occur in real people in real situations. We also told them that there would be no right or wrong answers, but that we were interested in their spontaneous impressions.

A mixed factorial design was used, with expression (seven targets, seven distractors) as the within-subjects factor and encoder (Face 1, Face 2) as the between-subjects factor. Hence, each participant viewed 14 dynamic expressions with one of the two encoders. Stimulus presentation was randomized and preceded by a fixation cross that always appeared in the middle of the computer screen. To give participants sufficient time to view the peak expression, we held the final frame of all videos for an additional 1 s.

Dependent Variables For each stimulus, the participants successively rated how much the facial movement reflected (1) a sudden event experienced by the person (suddenness); (2) a pleasant experience of the person (pleasantness); and (3) an experience that is discrepant from what may have been expected by the person (discrepancy). On completion of the appraisal ratings, participants indicated how strongly each emotion – anger, fear, disgust, happiness, sadness, and surprise – was represented in the face. The six emotion categories were presented on the same screen, allowing participants to complete the ratings in their own order.

All questions were answered by clicking the mouse on the appropriate parts of a 7-point Likert scale, with response options ranging from 1 (not at all) to 7 (very).

Results and Discussion A multivariate analysis of variance with expression dynamics (Target Stimulus 1–7) as the within-subjects factor and encoder (1, 2) as the between-subjects factor was conducted on the intensity ratings of the three cognitive appraisal and six emotion dimensions. For all univariate analyses, a Greenhouse-Geisser adjustment to degrees of freedom was applied.

The multivariate main effect for encoder, F(9, 42) = 1.51, p = .176, ηp² = .24, was not significant, suggesting that the stimulus face made no difference to the ratings. As predicted, there © 2016 Hogrefe

9

was a significant multivariate main effect of expression dynamics, F(9, 297) = 14.48, p = .000, ηp² = .30. In univariate terms, it was significant only for ratings of suddenness, F(4.96, 247.97) = 2.44, p = .035, ηp² = .05, anger, F(5.06, 253.04) = 16.81, p = .000, ηp² = .25, fear, F(5.13, 256.43) = 5.11, p = .000, ηp² = .09, and sadness, F(5.60, 280.24) = 2.97, p = .009, ηp² = .06.

Pairwise comparisons showed that participant ratings distinguished significantly between the dynamic sequences of fear expressions. The means and standard errors of the main effects are shown in Table 1. Specifically, fear expressions were rated to be most sudden when eyebrow raising was followed or preceded by eyelid raising (Sequences 2 and 3). Similar levels of suddenness were attributed to Sequences 1 and 4, but ratings were significantly lower for expressions commencing with brow lowering (Sequences 5 and 6) and with simultaneous action.

In contrast, anger was most strongly attributed to fear expressions in which brow lowering was followed or preceded by eyelid raising. Specifically, Sequence 6 led to the highest ratings of anger, which differed only marginally from those of Sequence 4. For both sequences, perceived levels of anger were significantly higher than those of the other sequences, including the expression with simultaneous dynamics. The latter expression was judged as being least angry and differed significantly from all other sequences.

For ratings of fear, participants’ attributions were highest for expressions in which eyebrow raising was followed or preceded by brow lowering (Sequences 1 and 5). As predicted, similarly high levels of fear were ascribed to expressions with simultaneous dynamics, a result that is also in line with findings

by Wehrle et al. (2000). For all three expressions, judgments of fear were significantly higher than those of the other sequences.

Furthermore, sadness ratings varied significantly as a function of the dynamic occurrence of facial actions. Malatesta et al. (2009) reported similar effects for perceived sadness when showing expressions of fear. Here, Sequence 1 and the expressions with simultaneous dynamics were judged to be higher on sadness than the other sequences, although not all differences reached statistical significance.

When comparing participants’ ratings of the fear expressions between emotions/appraisal dimension, the attributed levels of suddenness were highest for Sequences 2 and 3 and significantly different from the emotion ratings (ps < .01). There was no significant difference in perceived intensity between anger and sadness for Sequences 2 and 3. Sequence 6 was judged to be equally angry, fearful, and sudden, with these all receiving higher ratings than sadness (ps < .001). This was also the case for Sequence 4, which had comparable ratings of anger and fear that were significantly different from ratings of suddenness and sadness (ps < .001). For Sequences 1 and 5, ratings of fear and suddenness were equally high and significantly different from those of anger and sadness (ps < .001). The expression with simultaneous dynamics evoked strong perceptions of fear, Swiss Journal of Psychology (2016), 75 (1), 5–14

10

with significant differences between all dependent variables (ps ≤ .05).

Together, these findings suggest that expression dynamics exert a powerful impact on the decoding of emotional expressions. Although the final image, a prototypical fear expression, was the same in all cases, the degree of perceived emotion and appraisal dimension differed significantly depending on the order in which the facial actions occurred. Specifically, the ratings of suddenness were highest for Sequences 2 and 3, whereas Sequences 1 and 5 evoked impressions of a more fearful/sudden experience that was comparable in ratings of fear to expressions with simultaneous dynamics. Anger judgments were most pronounced for Sequences 6 and 4, thereby making the fear expression appear equally angry as well as being fearful/sudden.

Experiment 2 Experiment 2 examined the role of expression dynamics using a limited set of maximally discriminative facial sequences and tested these effects in combination with different levels of eye gaze. Research has shown that gaze direction represents a critical cue for inferring what someone is thinking or feeling and interacts with the emotional information from the face. Specifically, fearful expressions are judged to be more intense with an averted gaze, whereas angry expressions appear to be stronger with direct gaze (Adams & Kleck, 2005; N’Diaye et al., 2009; Sander et al., 2007). Given that the results of Experiment 1 showed that the sequential nature of fear displays implied varying levels of perceived anger and fear, we predicted that gaze direction significantly moderates the intensity ratings of these facial sequences.

Method

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

used in Experiment 1, including the target intensity of each AU at the peak level. In addition to the expressions with direct gaze, two versions with averted gaze (leftward, rightward) were created for each fear sequence. Gaze aversion was manipulated in a static form and displayed at a target intensity of 0.50 to prevent attentional shifts which could leave insufficient resources for the analysis of dynamic facial sequences. The overall duration of all target stimuli was the same and lasted a total of 2.8 s.

To ensure sufficient stimulus variability, we embedded the nine target expressions in a set of nine distractor stimuli. Similar to those in Experiment 1, these consisted of eyelid drop (AU43), eye squinting (AU44), and lid tightening (AU7) and were displayed with direct gaze as well as averted gaze to the left and the right (target intensity of 0.50). The dynamic trajectory of all distractor stimuli from onset to apex phase was identical to that of the previous experiment, with an overall duration of 2.1 s.

The 36 dynamic stimuli resulting from the combination of two faces with nine target fear expressions and nine distractor expressions were edited in Adobe After Effects and displayed as movie clips (800 × 1032 pixels) on a black background.

A fully within-subjects design was used, with gaze (direct, leftward, rightward), expression (three targets, three distractors), and encoder (Face 1, Face 2) as within-subjects factors.

Hence, each participant viewed all 36 dynamic expressions.

The presentation of the stimuli was randomized and preceded by a fixation cross that always appeared in the middle of the computer screen. To allow participants sufficient time to view peak expression, we held the final frame of all videos for an additional 1 s.

Procedure and Dependent Variables The procedure and the dependent variables were identical to those used in Experiment 1.

Participants A total of 41 students (35 women, 6 men) from a university in Switzerland participated in exchange for course credit or 10 Swiss francs. Their mean age was 22.93 years (SD = 3.49), ranging from 18 to 38 years.

Stimulus Material and Design The two male faces used in Experiment 1 were used. These faces expressed the three fear sequences (of the six used in Experiment 1) that had been shown to be maximally discriminative in appearance: Sequence 1 (AU1+2, 4, 5), Sequence 3 (AU5, 1+2, 4), and Sequence 6 (AU4, 5, 1+2). The dynamic sequencing in the eye region and eyebrows was identical to that Swiss Journal of Psychology (2016), 75 (1), 5–14

Results and Discussion A multivariate analysis of variance with gaze (straight, averted), expression dynamics (Target Stimuli 1–3), and encoder (1, 2) as within-subjects factors was conducted on the intensity ratings of the three cognitive appraisal and six emotion dimensions. For all univariate analyses, a Greenhouse-Geisser adjustment to degrees of freedom was applied.

The multivariate main effect for encoder, F(9, 32) = 1.32, p = .266, ηp² = .27, was not significant, suggesting that the stimulus face made no difference to the ratings. There was a significant multivariate main effect of gaze, F(9, 32) = 3.08, p = .009, ηp² = .46. In univariate terms, this was significant only for ratings of fear, F(1, 40) = 8.13, p = .007, ηp² = .17. Pairwise comparisons showed that fear expressions with direct gaze were © 2016 Hogrefe

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

judged to be more fearful (M = 4.79) than those with averted gaze (M = 4.40). Contrary to predictions, no significant interaction between gaze and expression dynamics emerged, F(18, 23) = 1.58, p = .151, ηp² = .55.

As in Experiment 1, the multivariate main effect for expression dynamics, F(18.23) = 4.60, p = .000, ηp² = .78, was significant. In univariate terms, the effect again was only significant for ratings of suddenness, F(1.92, 76.67) = 5.60, p = .006, ηp² = .12, anger, F(1.69, 67.71) = 30.15, p = .000, ηp² = .43, fear, F(1.95, 77.84) = 12.36, p = .000, ηp² = .24, and sadness, F(1.90, 76.00) = 13.34, p = .000, ηp² = .25.

The overall pattern of results was similar to that of Experiment 1. Fear expressions commencing with eyebrow raising or eyelid raising (Sequences 1 and 3) were judged as being more sudden than those commencing with brow furrowing (Sequence 6).2 In contrast, ratings of anger were highest for fear expressions that started with brow furrowing (Sequence 6). Significantly less anger was ascribed to expressions commencing with eyebrow raising (Sequence 1) and eyelid raising (Sequence 3), with the latter expression judged as being least angry particularly in the context of averted gaze.

For ratings of fear, participants’ attributions were significantly higher for expressions commencing with eyebrow raising and eyelid raising (Sequences 1 and 3) than for those commencing with brow furrowing (Sequence 6). This was especially the case in the context of direct gaze, with Sequence 1 being judged as the most fearful of all sequential expressions.

The results were similar for sadness ratings. Greater levels of sadness were attributed to expressions commencing with eyebrow raising and eyelid raising (Sequences 1 and 3) as compared to expressions that started with brow furrowing (Sequence 6).

When the participants’ ratings of the fear expressions on the emotion and cognitive appraisal dimensions were compared, we found that the attributed levels of suddenness were significantly higher than the emotion ratings for Sequence 3 (ps < .01). There was no significant difference in perceived intensity between anger and sadness for Sequences 3 and 1. Due to strong perceptions of fear when Sequence 1 was combined with direct gaze, fear ratings were comparable to those of suddenness. Sequence 6 was judged to be equally angry and fearful, with ratings significantly different from those of suddenness and sadness (ps < .01).

Together, the findings replicate the pattern found in Experiment 1. Again, suddenness ratings were highest and significantly different from the emotion ratings for Sequence 3.

Whereas Sequence 1 evoked impressions of a more fearful/sudden experience, Sequence 6 was judged as being equally angry and fearful. Furthermore, the perceived intensity of fear depended on eye gaze, with direct gaze surprisingly leading to stronger responses. These findings suggest that the signal value

11

of expressions of fear varies reliably depending on the sequence of facial actions and on whether the processing of such sequential dynamics is facilitated by eye-to-eye contact.

General Discussion Evidence increasingly suggests that emotional expressions unfold in a sequential fashion with the constituent facial actions occurring at different points in time (e.g., Jack et al., 2014; Krumhuber & Scherer, 2011; With & Kaiser, 2011). As the sequence assumption gains empirical support in naturally expressed emotions, there is still much to learn about how this process relates to the decoding of sequential displays. The purpose of the present research was exploratory in nature and entailed systematically examining the effects of different dynamic sequences on the perception of emotion expression. In both experiments, prototypical displays of fear with different sequential orders of facial actions resulted in distinct perceptions of emotion (fear, anger, sadness) and appraisal ratings (suddenness).

Expressions that were presented in a sequence postulated

by CAT (Scherer, 1984), with eyelid raising (power/control appraisal) appearing last, were perceived to be most fearful. Fear ratings given to these sequential displays (i.e., Sequences 1 and 5) did not differ significantly from those given to simultaneous dynamics, suggesting that they were similar in their expressive value. These findings and those of others (Malatesta et al., 2009; Wehrle et al., 2000) challenge classical viewpoints that only consider typical fear displays to be dynamic expressions with simultaneous onset of all facial actions. The present data suggest that sequential displays that follow CAT predictions can provide similar clues to the perception of fear. When this sequence assumption was violated, attributions greatly differed, with fear expressions judged to be more sudden (i.e., Sequences 2 and 3) or angry (i.e., Sequences 4 and 6). Although this did not imply an entirely different emotion, it suggested that prototypical fear expressions may be a combination of two basic emotions (viz., anger and fear).

The findings have implications for understanding how various psychological states are derived from information contained in the eye region. Up to now, most work has been concerned with the perception of full-blown, static information from the eyes. That is, observers had to recognize what a person was feeling or thinking from a single position of the upper face at or very near the peak of the expression (e.g., Calvo & Fernández-Martin, 2013; Eisenbarth & Alpers, 2011; Hopkins et al., 2011; Joseph & Tanaka, 2003). Such an approach limits explanations of how emotions and mental states are understood, as facial displays communicate information not only at

2 The difference between Sequences 1 and 6 reached statistical significance (p < .05) when data were combined across gaze levels. © 2016 Hogrefe

Swiss Journal of Psychology (2016), 75 (1), 5–14

12

their apex, but throughout the course of the expression. The present research provides strong evidence that the attribution of meaning is not only shaped by the prototypical configuration of facial actions (in all cases it was recognizably a fear expression), but also by the dynamic sequence in which they occur.

This knowledge is relevant for gaining a better understanding of perceptual processes in healthy individuals, as well in those who have difficulties recognizing emotional and mental states from the face (see Baron-Cohen, Golan, & Ashwin, 2009; Hobson, 1986). People with autism spectrum disorder (ASD) show marked impairment in the use and interpretation of information conveyed by the eyes (Baron-Cohen et al., 1995; Joseph & Tanaka, 2003). Such deficiency is particularly evident in the detection of fear, which relies on various cues from the eye region (Adolphs et al., 2005; Pelphrey et al., 2002). By using static facial displays, researchers may underestimate people’s abilities to detect the emotional states, thoughts, and feelings of others (Moore, 2001). An interesting future step would be to explore whether people with ASD are sensitive to dynamic sequences of facial actions when judging emotional expressions. At present, the Eyes Test (Baron-Cohen et al., 1997, 2001), commonly used for determining mind-reading difficulties in ASD, only exists in a static form. Our test battery of dynamic facial sequences in combination with different levels of eye gaze could be used as a new measure of social intelligence. Its subtle nature makes it suitable for application in groups that exhibit slight degrees of autistic traits.

Besides allowing a more complex representation of facial displays for the purpose of assessment, these insights are also relevant for conceptualizing emotions in applied fields, such as that of affective computing. In order to equip virtual agents with emotional expressions that change in a believable and lifelike manner, computer models have recently taken to incorporating a sequential approach. Facial actions are activated at different time points and submerge into partial expressions, rather than full-blown prototypical patterns (Niewiadomski, Hyniewska, & Pelachaud, 2009a, 2009b; Paleari, Grizard, & Lisetti, 2007). To further develop these models, there is a need for data that go beyond the decoding of emotions at the apex and consider the perception of dynamic sequences of facial actions. The present findings provide some preliminary input with respect to the sequential nature of emotional expressions and their impact on the perceiver.

In this research, gaze direction was shown to influence ratings of fear, with stronger subjective responses for expressions with direct gaze. Surprisingly, this effect occurred independently of the sequential nature of the fear display. As such, eye gaze did not have a selective impact on various dynamic sequences, but instead seemed to generally enhance the perceived intensity of the expression. The data fail to replicate previous findings (see Adams & Kleck, 2005; N’Diaye et al., 2009; Sander et al., 2007) that yielded higher ratings of fear displays with averted rather than direct gaze. However, similar results have been reported by Bindemann, Burton, and Langton (2008), who also Swiss Journal of Psychology (2016), 75 (1), 5–14

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

found a significant effect of eye gaze for fearful-looking faces, with stronger fear ratings in the context of direct gaze.

A possible explanation is that eye-to-eye contact (direct gaze) facilitates emotion processing under specific circumstances, for example, when viewing complex expressions, such as sequential displays in which the meaning can only be derived over the course of the expression. This is because eye contact allows perceivers to map or simulate another person’s mental states and intentions (Niedenthal, Mermillod, Maringer, & Hess, 2010; Nummenmaa & Calder, 2009), which cannot be achieved when the eyes are averted. Supportive evidence comes from studies that have revealed a processing advantage for emotional faces with direct gaze (Graham & LaBar, 2007; Senju & Hasegawa, 2005; Wicker, Perrett, Baron-Cohen, & Decety, 2003). It is possible that perceivers benefited from eye contact and consequently gave stronger subjective responses for sequential fear expressions with direct gaze. This assumption is also plausible given the finding that individuals with ASD perform poorly when judging the emotions and thoughts of others, especially when the eyes are averted (e.g., Baron-Cohen et al., 1995). Direct gaze may therefore cause enhanced processing of social stimuli by focusing attention on the face.

In this research, the effects of sequential dynamics and gaze direction on perceivers’ judgments were explored through fear expressions. Additional research is clearly needed to examine the role of facial sequences in other emotional expressions. For example, preliminary evidence by Nishio, Koyama, and Nakamura (1998) suggests that smile expressions in which the mouth starts moving prior to the eyes are regarded as more pleasant than smile expressions in the reverse order. The sequential occurrence of facial actions might therefore influence judgments that go beyond the recognition or perceived intensity of emotions. It remains to be determined whether similar effects occur for ratings of expression sincerity or intentionality.

Moreover, the timing of the sequential occurrence of facial actions requires further investigation. The present experiments used a constant time period and transition interval between sequential facial actions, which may not be found in this exact form in naturally expressed emotions. Although this approach may pose a limitation to the present work in terms of ecological validity, it may act as an initial proof of concept in showing that observers are sensitive to varying sequential patterns of facial displays. Such sequences could be further studied in natural expressions to determine the boundary conditions in emotion perception. For example, how are different facial movements integrated over time, and what constitutes a diagnostic facial pattern? Future research will be necessary to determine the exact perceptual processes of dynamic facial sequences as they occur in everyday expressions. The present work is a first exploratory step in showing that the perceiver’s interpretation is not limited to drawing meaning from one expression, but that there is instead variation in the look of fear from the eyes due to the dynamic sequence of facial actions. © 2016 Hogrefe

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

Acknowledgments This research was supported by the Swiss National Science Foundation (FNRS 101411–100367). The authors would like to thank Flavie Martin for data collection.

References Adams, R. B., & Kleck, R. E. (2003). Perceived gaze direction and the processing of facial displays of emotion. Psychological Science, 14, 644–647. doi 10.1046/j.0956–7976.2003.psci_1479.x Adams, R. B., & Kleck, R. E. (2005). Effects of direct and averted gaze on the perception of facially communicated emotion. Emotion, 5, 3–11. doi 10.1037/1528-3542.5.1.3 Adolphs, R., Gosselin, F., Buchanan, T. W., Tranel, D., Schyns, P., & Damasio, A. R. (2005). A mechanism for impaired fear recognition after amygdala damage. Nature, 433(7021), 68–72. doi 10.1038/nature03086 Aue, T., Flykt, A., & Scherer, K. R. (2007). First evidence for differential and sequential efferent effects of stimulus relevance and goal conduciveness appraisal. Biological Psychology, 74, 347–357. doi 10.1016/j.biopsycho.2006.09.001 Back, E., Ropar, D., & Mitchell, P. (2007). Do the eyes have it? Inferring mental states from animated faces in autism. Child Development, 78, 397–411. doi 10.1111/j.1467-8624.2007.01005.x Baron-Cohen, S., Campbell, R., Karmiloff-Smith, A., Grant, J., & Walker, J. (1995). Are children with autism blind to the mentalistic significance of the eyes? British Journal of Developmental Psychology, 13, 379–398. doi 10.1111/j.2044-835X.1995.tb00687.x Baron-Cohen, S., & Cross, P. (1992). Reading the eyes: Evidence for the role of perception in the development of a theory of mind.

Mind and Language, 7, 172–186. doi 10.1111/j.14680017.1992.tb00203.x Baron-Cohen, S., Golan, O., & Ashwin, E. (2009). Can emotion recognition be taught to children with autism spectrum conditions?

Philosophical Transactions of the Philosophical Society B, 364, 3567–3574. doi 10.1098/rstb.2009.0191 Baron-Cohen, S., Jolliffe, T., Mortimore, C., & Robertson, M. (1997).

Another advanced test of theory of mind: Evidence from very high functioning adults with autism or Asperger syndrome. Journal of Child Psychology and Psychiatry, 38, 813–822. doi 10.1111/j.1469-7610.1997.tb01599.x Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I. (2001). The Reading the Mind in the Eyes Test revised version: A study with normal adults, and adults with Asperger syndrome or high-functioning autism. Journal of Child Psychology and Psychiatry, 42, 241–251. doi 10.1111/1469-7610.00715 Bindemann, M., Burton, A. M., & Langton, S. R. H. (2008). How do eye gaze and facial expression interact? Visual Cognition, 16, 708–733. doi 10.1080/13506280701269318 Bombari, D., Schmid, P. C., Schmid Mast, M., Birri, S., Mast, F. W., & Lobmaier, J. S. (2013). Emotion recognition: The role of featural and configural faceinformation. Quarterly Journal ofExperimental Psychology, 66, 2426–2442. doi 10.1080/17470218.2013.789065 Calvo, M. G., & Fernández-Martin, A. (2013). Can the eyes reveal a person’s emotions? Biasing role of the mouth expression. Motivation and Emotion, 37, 202–211. doi 10.1007/s11031-0129298-1 Delplanque, S., Grandjean, D., Chrea, C., Coppin, G., Aymard, L., Cayeux, I., . . . Scherer, K. R. (2009). Sequential unfolding of novelty and pleasantness appraisals of odors: Evidence from facial electromyography and autonomic reactions. Emotion, 9, 316–328. doi 10.1037/a0015369

© 2016 Hogrefe

13

Eisenbarth, H., & Alpers, G. W. (2011). Happy mouth and sad eyes: Scanning emotional facial expressions. Emotion, 11, 860–865. doi 10.1037/a0022758 Ekman, P. (1979). About brows: Emotional and conversational signals. In M. von Cranach, K. Foppa, W. Lepenies, & D. Ploog (Eds.), Human ethology: Claims and limits of a new discipline (pp. 169–202). Cambridge, UK: Cambridge University Press.

Ekman, P. (1989). The argument and evidence about universals in facial expressions of emotion. In H. Wagner & A. Manstead (Eds.), Handbook of social psychophysiology (pp. 143–164). Chichester, UK: Wiley.

Ekman, P. (1992). An argument for basic emotions. Cognition and Emotion, 6, 169–200. doi 10.1080/02699939208411068 Ekman, P. (2003). Darwin, deception and facial expression. Annals of the New York Academy of Sciences, 1000, 205–221. doi 10.1196/annals.1280.010 Ekman, P., & Friesen, W. V. (1982). Felt, false and miserable smiles.

Journal of Nonverbal Behavior, 6, 238–258. doi 10.1007/ BF00987191 Ekman, P., Friesen, W. V., & Hager, J. C. (2002). The Facial Action Coding System (2nd ed.). Salt Lake City, UT: Research Nexus eBook.

Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal processes in emotion. In R. J. Davidson, K. R. Scherer, & H. H. Goldsmith (Eds.), Handbook of affective sciences (pp. 572–595). New York: Oxford University Press.

Fiorentini, C., Schmidt, S., & Viviani, P. (2012). The identification of unfolding facial expressions. Perception, 41, 532–555. doi 10.1068/p7052 Gepner, B., Deruelle, C., & Grynfeltt, S. (2001). Motion and emotion: A novel approach to the study of face processing by young autistic children. Journal of Autism and Developmental Disorders, 31, 37–45. doi 10.1023/A:1005609629218 Graham, R., & LaBar, K. (2007). Garner interference reveals dependencies between emotional expression and gaze in face perception. Emotion, 7, 296–313. doi 10.1037/1528-3542.7.2.296 Hobson, R. P. (1986). The autistic child’s appraisal of expressions of emotion. Journal of Child Psychology and Psychiatry, 27, 321–342. doi 10.1111/j.1469-7610.1986.tb01836.x Hopkins, I. M., Gower, M. W., Perez, T. A., Smith, D. S., Amthor, F. R., Wimsatt, F. C., & Biasini, F. J. (2011). Avatar assistant: Improving social skills in students with an ASD through a computer-based intervention. Journal of Autism and Developmental Disorders, 41, 1543–1555. doi 10.1007/s10803-011-1179-z Jack, R. E., Garrod, O. G. B., & Schyns, P. G. (2014). Dynamic facial expressions of emotions transmit an evolving hierarchy of signals over time. Current Biology, 24, 187–192. doi 10.1016/j.cub.

2013.11.064 Johnston, L., Miles, L., & McKinlay, A. (2008). A critical review of the Eyes Test as a measure of social-cognitive impairment. Australian Journal of Psychology, 60, 135–141. doi 10.1080/ 00049530701449521 Joseph, R. M., & Tanaka, J. (2003). Holistic and part-based face recognition in children with autism. Journal of Child Psychology and Psychiatry, 44, 529–542. doi 10.1111/1469-7610.00142 Krumhuber, E. G., Kappas, A., & Manstead, A. S. R. (2013). Effects of dynamic aspects of facial expressions: A review. Emotion Review, 5, 41–46. doi 10.1177/1754073912451349 Krumhuber, E., & Scherer, K. R. (2011). Affect bursts: Dynamic patterns of facial expression. Emotion, 11, 825–841. doi 10.1037/a0023856 Krumhuber, E., Tamarit, L., Roesch, E. B., & Scherer, K. R. (2012).

FACSGen 2.0 animation software: Generating 3D FACS-valid facial expressions for emotion research. Emotion, 12, 351–363. doi 10.1037/a0026632 Lanctôt, N., & Hess, U. (2007). The timing of appraisals. Emotion, 7, 207–212. doi 10.1037/1528-3542.7.1.207 Langton, S. R. H, Watt, R. J., & Bruce, V. (2000). Do the eyes have it?

Swiss Journal of Psychology (2016), 75 (1), 5–14

14

Cues to the direction of social attention. Trends in Cognitive Sciences, 4(2), 50–59. doi 10.1016/S1364-6613(99)01436-9 Malatesta, L., Raouzaiou, A., Karpouzis, K., & Kollias, S. (2009). Towards modeling embodied conversational agent character profiles using appraisal theory predictions in expression synthesis.

Applied Intelligence, 30, 58–64. doi 10.1007/s10489-007-00769 Milders, M., Hietanen, J. K., Leppänen, J. M., & Braun, M. (2011).

Detection of emotional faces is modulated by the direction of eye gaze. Emotion, 11, 1456–1461. doi 10.1037/a0022901 Moore, D. G. (2001). Reassessing emotion recognition performance in people with mental retardation: A review. American Journal on Mental Retardation, 106, 481–502.

Morris, J. S., deBonis, M., & Dolan, R. J. (2002). Human amygdala responses to fearful eyes. NeuroImage, 17, 214–222. doi 10.1006/nimg.2002.1220 N’Diaye, K., Sander, D., & Vuilleumier, P. (2009). Self-relevance processing in the human amygdala: Gaze direction, facial expression and emotion intensity. Emotion, 9, 798–806. doi 10.1037/a0017845 Niedenthal, P. M., Mermillod, M., Maringer, M., & Hess, U. (2010).

The simulation of smiles (SIMS) model: Embodied simulation and the meaning of facial expression. Behavioral and Brain Sciences, 33, 417–433. doi 10.1017/S0140525X10000865 Niewiadomski, R., Hyniewska, S., & Pelachaud, C. (2009a). Evaluation of multimodal sequential expressions of emotions in ECA. In J. Cohn, A. Nijholt, & M. Pantic (Eds.), Proceedings of the 3rd International Conference on Affective Computing and Intelligent Interaction – ACII 2009 (pp. 635–641). Los Alamitos, CA: IEEE Computer Society Press.

Niewiadomski, R., Hyniewska, S., & Pelachaud, C. (2009b). Modeling emotional expressions as sequences of behaviors. In Z. Ruttkay, M. Kipp, A. Nijholt, & H. H. Vilhjálmsson (Eds.), Lecture Notes in Computer Science: Vol. 5773. Intelligent Virtual Agents: 9th International Conference, IVA 2009, Amsterdam, The Netherlands, September 14—16, 2009 Proceedings (pp. 316–322).

Berlin, Germany: Springer. doi 10.1007/978-3-642-04380-2_34 Nishio, S., Koyama, K., & Nakmura, T. (1998). Temporal differences in eye and mouth movements classifying facial expressions of smiles. In Proceedings of the 3rd International Conference on Face & Gesture Recognition – FG ’98 (pp. 206–213). Nara, Japan: IEEE Computer Society Press. doi 10.1109/AFGR.1998.670950 Nummenmaa, L., & Calder, A. J. (2009). Neural mechanisms of social attention. Trends in Cognitive Sciences, 13, 135–143. doi 10.1016/j.tics.2008.12.006 Oosterhof, N. N., & Todorov, A. (2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences of the USA, 105, 11087–11092. doi 10.1073/pnas.0805664105 Paleari, M., Grizard, A., & Lisetti, C. (2007). Adapting psychologically grounded facial emotional expressions to different anthropomorphic embodiment platforms. In D. Wilson & G. Sutcliffe (Eds.), Proceedings of the 20th Florida Artificial Intelligence Research Society Conference (FLAIRS 2007) (pp. 565–570). Menlo Park, CA: AAAI Press.

Pelphrey, K. A., Sasson, N. J., Reznick, S. J., Paul, G., Goldman, B. D., & Piven, J. (2002). Visual scanning of faces in autism. Journal of Autism and Developmental Disorders, 32, 249–261. doi 10.1023/A:1016374617369 Pilowsky, I., Thornton, M., & Stokes, B. (1986). Towards the quantification of facial expressions with the use of a mathematic model of the face. In H. Ellis, M. A. Jeeves, F. Newcombe, & A. Young (Eds.), Aspects of face processing (pp. 340–348). Dordrecht, The Netherlands: Martinus Nijhoff.

Swiss Journal of Psychology (2016), 75 (1), 5–14

E. G. Krumhuber & K. R. Scherer: Sequential Dynamics

Sander, D., Grandjean, D., Kaiser, S., Wehrle, T., & Scherer, K. R. (2007). Interaction effects of perceived gaze direction and dynamic facial expression: Evidence for appraisal theories of emotion. European Journal of Cognitive Psychology, 19, 470–480. doi 10.1080/09541440600757426 Scherer, K. R. (1984). On the nature and function of emotion: A component process approach. In K. R. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 293–317). Hillsdale, NJ: Erlbaum.

Scherer, K. R. (1999). On the sequential nature of appraisal processes: Indirect evidence from a recognition task. Cognition and Emotion, 13, 763–793. doi 10.1080/026999399379078 Scherer, K. R. (2001). Appraisal considered as a process of multilevel sequential checking. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion: Theory, methods, research (pp. 92–120). New York: Oxford University Press.

Scherer, K. R. (2009). The dynamic architecture of emotion: Evidence for the component process model. Cognition and Emotion, 23, 1307–1351. doi 10.1080/02699930902928969 Scherer, K. R., & Ellgring, H. (2007). Are facial expressions of emotion produced by categorical affect programs or dynamically driven by appraisal? Emotion, 7, 113–130. doi 10.1037/15283542.7.1.113 Senju, A., & Hasegawa, T. (2005). Direct gaze captures visuospatial attention. Visual Cognition, 12, 127–144. doi 10.1080/ 13506280444000157 Smith, M. L., Cottrell, G., Gosselin, F., & Schyns, P. G. (2005). Transmitting and decoding facial expressions of emotions. Psychological Science, 16, 184–189. doi 10.1111/j.0956-7976.2005.00801.x Smith, C. A., & Scott, H. S. (1997). A componential approach to the meaning of facial expressions. In J. A. Russell & J. M. FernàndezDols (Eds.), The psychology of facial expression (pp. 229–254).

Cambridge, UK: Cambridge University Press.

Uono, S., Sato, W., & Toichi, M. (2009). Dynamic fearful expressions enhance gaze-triggered attention orienting in high and low anxiety individuals. Social Behavior and Personality, 37, 1313–1326. doi 10.2224/sbp.2009.37.10.1313 Vuilleumier, P. (2005). Staring fear in the face. Nature, 433, 22–23. doi 10.1038/433022a Wehrle, T., Kaiser, S., Schmidt, S., & Scherer, K. R. (2000). Studying the dynamics of emotional expression using synthesized facial muscle movements. Journal of Personality and Social Psychology, 78, 105–119. doi 10.1037/0022-3514.78.1.105 Whalen P. J., Kagan J., Cook, R. G., Davis, F. C., Kim, H., Polis, S., . . .

Johnstone T. (2004). Human amygdala responsivity to masked fearful eye whites. Science, 306(5704), 2061. doi 10.1126/science.1103617 Wicker, B., Perrett, D. I., Baron-Cohen, S., & Decety, J. (2003). Being the target of another’s emotion: A PET study. Neuropsychologia, 41, 139–146. doi 10.1016/S0028-3932(02)00144-6 With, S., & Kaiser, S. (2011). Sequential patterning of facial actions in the production and perception of emotional expressions.

Swiss Journal of Psychology, 70, 241–252. doi 10.1024/14210185/a000062

Eva Krumhuber Department of Experimental Psychology University College London 26 Bedford Way, WC1H 0AP London United Kingdom e.krumhuber@ucl.ac.uk

© 2016 Hogrefe

J. Guegan Sw issJournal et al.: Gender of Psychology Perception (2016), in ©Dyadic 2016 75 (1), Hogrefe 15–23 CM C

Original Communication

Social Asymmetries and Anonymity in Dyadic Computer-Mediated Communication An Experimental Study of Gender Perception Jérôme Guegan1, Pascal Moliner2, and Laurent Milland3 1

LATI, EA 4469, Université Paris Descartes, France

2

Epsylon, EA 4556, Université Montpellier 3, France CeRCA, UMR-CNRS 6234, Université de Poitiers, France

3

Abstract. This research concerns the influence of dominant/dominated asymmetries in the perceptual structure of social categories in computer-mediated communication (CMC). We explore the connection between the cognitive effects described by the social identity model of deindividuation effects (SIDE) and the impact on perceptions of the asymmetrical status that characterizes gender groups. The study analyzes the gender perception of two interlocutors in a communication context of anonymity or visibility. The experiment was based on a standardized protocol designed to control the various declarative contents expressed during the communication. Eighty-eight participants were placed in a CMC situation with a confederate of the opposite sex. We manipulated visibility and anonymity through interlocutors’ names (provided vs. withheld) and the presence or absence of a webcam in the communication situation. In accordance with our predictions, the results suggest that gender asymmetry can modulate the cognitive effects of anonymity. The results and prospects of this research are discussed in light of work on the dominant/dominated groups and the SIDE model.

Keywords: computer-mediated communication, social category, gender, stereotypes

Situations in which two people communicate over the Internet have become part of our everyday social life. The settings for online communication – be they chatrooms, forums, dating sites, or social networks – have sparked the interest of many researchers in the field of computer-mediated communication (CMC).

Interactions between men and women, or between members of dominant/dominated groups, have been the subject of several studies in the field of CMC. For example, among the earlier conceptions of CMC (known as cues-filtered-out perspective, e.g., Kiesler, Siegel, & McGuire, 1984; Sproull & Kiesler, 1986), the equalization hypothesis (Dubrovsky, Kiesler, & Sethna, 1991) assumes that the impact of gender stereotypes and group boundaries should be significantly diminished during anonymous CMC. Indeed, most CMC is text-based (Lee, 2007), and a written text is less likely to reveal cues than, for example, a person’s physical appearance. However, people engaging in online interactions often reveal their sex during online interactions (Wolf, 2000) and gender identities can also be inferred through various cues such as writing style or communication content (e.g., Williams & Mendelsohn, 2008). The equalization hypothesis has received very little empirical validation, and several studies have invalidated it (e.g., Herring, © 2016 Hogrefe DOI 10.1024/1422-4917/a000167

1993; Postmes & Spears, 2002). Indeed, even if the interlocutors are anonymous (i.e., they cannot be identified by name or see each other), the mere presence of gender cues can influence perceptions and the formation of mutual impressions. In this research, we analyzed the effect of these cues in anonymous (nominal and visual anonymity) dyadic CMC between a man and a woman. In particular, we investigated the effect of gender cues on gender-stereotypical perception in CMC, taking into account group homogeneity/heterogeneity as a function of the social asymmetry that characterizes gender categories.

Gender Perceptions in CMC Gender stereotypes refer to different social roles and have been the subject of many studies (e.g., Bem, 1985; Eagly, 1987; Lorenzi-Cioldi, 1994). Parsons and Bales (1955) drew a distinction between instrumental roles (traditionally associated with men) and expressive roles (traditionally associated with women). Instrumental roles are associated with the control of the environment, competence, and social utility, while expressive roles are more related to interpersonal relationships and conSwiss Journal of Psychology (2016), 75 (1), 15–23

16

tact with others. Thus, women are portrayed as gentle and attentive to others, while men are more likely to be seen as competitive and dynamic. This type of stereotypical perception, defined through male and female profiles, has been observed in several CMC studies (e.g., Matheson, 1991). According to Fischer (2011), like any other area of social interaction, the Internet is “sex-typed” and full of stereotyping and sexism. Moreover, according to the gender-in-context approach (Deaux & Major, 1987), gender differences are highly context-dependent.

Thus, knowing that a partner is a man or a woman can generate expectations, stereotypes, and prototypical behavior, especially if the interacting parties do not know each other.

In CMC, the gender-in-context approach is consistent with proposals based on the social identity model of deindividuation effects (SIDE; Reicher, Spears, & Postmes, 1995; Spears & Lea, 1992, 1994). The SIDE model is a direct descendant of social identity (Tajfel & Turner, 1979) and self-categorization theories (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). According to the SIDE model, in situations of anonymity, the interlocutors cease to pay attention to individual differences or personal characteristics. Furthermore, when the scarcity of individuating information is combined with the presence of cues pertaining to the individuals’ affiliations (e.g., the name of a group or a particular color), the depersonalization process may be accentuated (Turner et al., 1987). By masking the idiosyncratic features of the individual group members, anonymity ensures that groups are presented in a homogeneous and standardized fashion (Lee, 2004). Therefore, people who do not have access to the specificities of each group member tend to reason on the basis of social categories.

Thus, as Postmes, Spears, and Lea (2002) indicate, “depersonalized interactions over the Internet could stimulate our natural tendency for differentiation between social categories” (p. 4).

This can have a major impact on the way in which members of groups are perceived, in particular by increasing the influence of stereotypes in anonymous online communication. This has been confirmed in several studies (Postmes & Spears, 2002; Postmes, Spears, & Lea, 1998; Spears & Lea, 1994; Spears, Lea, & Postmes, 2007). For instance, in a series of experiments, Postmes and Spears (2002) found that the activation of gender stereotypes in a context of anonymous CMC resulted in an increase in prototypical male or female behavior (e.g., tendency for women to ask more questions and be less dominant than men): “The inability to individuate group members not only increases the stereotypicality of impression formation but also of behavior” (p. 1080).

The Effects of Gender Asymmetries in CMC Research addressing gender in CMC mainly investigates whether anonymity – and fewer social cues – increases or decreases the stereotyped nature of gender perception in online interacSwiss Journal of Psychology (2016), 75 (1), 15–23

J. Guegan et al.: Gender Perception in Dyadic CMC

tion. In the present study, we go a step further by studying whether the perceptual characteristics of gender categories, resulting from asymmetrical relationships, can have a moderating effect on gender perception over the Internet.

Indeed, hierarchy can have a considerable impact on the perception of social categories (e.g., Sidanius & Pratto, 1999) that could potentially strengthen the effects predicted by the SIDE model in interactions involving dominant groups. Many studies have shown that, whatever the reasons for the social asymmetry (e.g., sex, sociooccupational category, minority group), dominant groups are perceived as heterogeneous (e.g., Linville, Salovey, & Fischer, 1986; Park, Ryan, & Judd, 1992), that is, as being composed of by specific individuals who are distinct from one another and therefore not interchangeable.

On the other hand, dominated groups are perceived as more homogeneous and therefore composed of individuals who are rather similar (Lorenzi-Cioldi, 1988, 2009; Lorenzi-Cioldi, Deaux, & Dafflon, 1998). Homogenizing a group means leaving behind the individual characteristics of its members. The members of a homogeneous group (i.e., dominated) are therefore perceived as deindividuated aggregates (i.e., under the direct influence of their group stereotype). By contrast, heterogeneous groups are viewed more as a collection of distinct, clearly individuated members (i.e., stereotype exceptions).

According to the theory of system justification (Jost & Banaji, 1994), this phenomenon is maintained by the individuals themselves. For members of dominant groups, having distinctive characteristics justifies enjoying a favored social position based on personal qualities and not social membership. In contrast, assimilation to the group – that is, minimization of distinctive characteristics – allows members of dominated groups to attribute their disadvantaged social position to their social group. In other words, they can avoid using personal characteristics to explain their disadvantaged position. In sum, this process is linked to the valuation of the self (Lorenzi-Cioldi & Dafflon, 1999) and suggests that the asymmetries of homogeneity encountered in dominant/dominated intergroup perceptions are the subject of a broad consensus among the members of these groups. If dominant groups see themselves as heterogeneous, dominated groups will perceive them as heterogeneous. On the other hand, members of dominated as well as dominant groups perceive dominated groups as homogeneous.

As a result, intergroup perceptions and stereotypes contribute to the status quo of social domination.

The SIDE model predicts that the perceived stereotypicality and homogeneity of dominated and of dominant groups will be reinforced during anonymous CMC. However, the proposals presented above help us to clarify the model: Because dominant and dominated groups are asymmetrical, the deindividuation effects of CMC should be more noticeable for the group that is usually subject to less homogenization (i.e., dominant groups).

Indeed, the effects predicted by the SIDE model (i.e., stereotypicality and homogeneity) would result in the same perceptive structure that dominated groups already have (i.e., stereotyped © 2016 Hogrefe

J. Guegan et al.: Gender Perception in Dyadic CMC

and homogeneous). In other words, a social category that is usually perceived as an aggregate of deindividuated members would be less sensitive to the deindividuating effects of anonymous online communication. In contrast, the perceptions of dominant groups – which are usually viewed as more heterogeneous and less stereotypical – would be more sensitive to the deindividuating effects of anonymous CMC. These predictions can be applied to gender categories: In an anonymous interaction, we expect a man (i.e., member of a dominant group) to be perceived in a more stereotyped way than in an interaction that allows personal identification of the interlocutors. In contrast, the perception of a woman (i.e., member of a dominated group) should be less sensitive to the anonymous or nonanonymous nature of the interaction. Therefore, the cognitive effects predicted by the SIDE model would be more or less pronounced depending on the gender groups’ asymmetrical domination/subordination relationship within the social hierarchy. The theoretical framework of the SIDE model alone does not justify this proposal. It can only be understood by the articulation of deindividuation effects described by the SIDE model and the symbolic consequences of social asymmetry on gender perceptions.

In line with these propositions, Guegan, Vergnaud, and Moliner (2011) showed that men tend to be perceived as more homogeneous in anonymous chatrooms than in social networks that allow one to identify each user. However, the perception of women, compared to men, is less influenced by the context of communication. Guegan et al. (2011) analyzed the perception of gender categories as a whole (i.e., women in general and men in general), but their study was only quasiexperimental since it did not involve actual interaction between participants. Thus, the present study proposes to test this hypothesis within a purely experimental procedure.

17

tive material that had already been established, assessed, and that remained unchanged across the different experimental conditions.

Once inside the room in which the individual sessions were conducted, the participants were invited to sit in front of a computer. They were shown a list of 15 questions they could use during the communication (see Table 1). These simple questions were general enough to elicit content-filled responses. The participants were to choose 5 of the 15 questions to ask their interlocutor.

Table 1. Questions used during the communication Questions available to participants

Confederates’ questions

1. What are the most important values in life to you?

2. What are your main qualities?

2

3. Why did you choose to join this course?

4. What are your hobbies and what do you gain from them?

1

5. How and where do you imagine yourself in 5 years’ time?

6. Do you have a lot of friends? Would you say that you are a sociable person?

3

7. Would you like to do a job that involves a lot of traveling?

8. What are you most passionate about?

9. What would you do if you won the lottery?

4

10. What place would you like to visit the most and why?

11. If you were to take three things with you to a desert island, what would they be and why?

12. Are you interested in politics?

13. Do you go out a lot in the evening?

14. Do you enjoy studying?

Method Participants Eighty-eight participants (46 women and 42 men; mean age: 19.18 years, SD = 1.97), all psychology undergraduates, took part in this research. None of the participants received remuneration, but they were given course credit for their participation.

Procedure Our goal was to study a situation of text-based standardized interpersonal communication between a man and a woman, while maintaining ecological validity despite the laboratory setting. Consequently, the methodology used in this research is based on the intervention of a confederate who used declara© 2016 Hogrefe

5

15. Would you like to do a job with responsibility that involves managing teams?

Note. The second column shows the questions asked by confederates of both genders (always in the same order from 1 to 5).

The interlocutor, who was actually a confederate, was introduced as another participant, located in another laboratory, overseen by a second experimenter. The participants were told that their interlocutor (confederate) was participating in the experiment under identical conditions, with the same list of questions. This meant that the participants would be asked to answer five questions chosen by their interlocutor.

The participants always interacted with a confederate of the opposite sex (and all participants of the same sex interacted with the same confederate). The participant and confederate took turns, employing a question/answer (Q&A) mode that allowed the participant to choose the questions he/she wanted to ask as well as respond freely to the questions he/she was asked.

Swiss Journal of Psychology (2016), 75 (1), 15–23

18

J. Guegan et al.: Gender Perception in Dyadic CMC

cions concerning the intervention of a confederate, so we did not include his data in the analysis.

Independent Variables

Figure 1. Schematic representation of the interaction between participant and confederate.

Once the instructions had been clearly understood, the participant and the confederate logged on to an Internet chatroom.

After an exchange of greetings, the Q&A session started. It lasted for 30–35 min. The participant always asked the very first question. The confederate answered, then asked his/her own first question, which the participant answered before going on to ask a second question (see Figure 1). The CMC continued in this way, in real time, until the participant and the confederate had each asked and answered five questions. As indicated above, participants freely selected the five questions to ask, from a list of 15 questions, and could give responses of their own choosing, provided that they were not too short or too long (2–5 lines). The male and female confederates, however, always asked the same five questions in the same order (see Table 1). As well as knowing which questions they were to ask, the confederates also had a list of 15 predetermined responses to the 15 possible questions. The confederates’ Q&A corpus had been assessed beforehand with respect to gender neutrality (see below) in order to disentangle the manipulated factors from the various elements expressed during the CMC. This neutral material was thus exactly the same for both confederates, the only variation being the grammatical sex used.

The success of the experiment depended on the participants’ being unaware of this and their belief that the confederates were in exactly the same situation they were.

Every message sent during the interaction appeared in a dialog window and remained on the screen until the end of the experiment. As the participants’ responses were sometimes quite personal, the entire conversation was deleted at the end of the experiment (the participants were informed about this before they started the experiment).

At the end of the CMC, the participants were asked to fill out a questionnaire measuring the stereotypicality of their perceptions of the confederate. The entire experiment took 45–50 min. After debriefing, one participant expressed suspiSwiss Journal of Psychology (2016), 75 (1), 15–23

The procedure was the same for each participant-confederate pair, with a male confederate for female participants and a female confederate for male participants (between-participants confederate sex: male vs. female) in two distinct contexts (between-participants context: anonymity vs. visibility). The Internet website created for the experiment had two chatrooms, one for each modality of the context factor.

In the first chatroom (anonymity condition, with group cues only; n = 44), both participant and confederate could be identified on the basis of their category affiliation, for when they logged on, they indicated their gender. Thus, Boy or Girl served as pseudonyms, appearing next to every message sent during the CMC. No further or more precise indication of the interlocutors’ identity was given. The latter could not communicate orally and no visual contact was possible. In this condition, the interlocutors were thus in a situation of both visual and nominal anonymity (see Lea, Spears, Watt, & Rogers, 2000).

In the second chatroom (visibility condition, with both group and individuating cues; n = 44), the first and last names of the participant and the confederate were shown. When the participants logged on, they indicated their first name, last name, and sex. Their first and last name then appeared next to every message sent during the CMC. The confederate always used the same first and last name: Nicolas Templon for the male confederate and Aurélie Templon for the female confederate.

Moreover, in the second chatroom, it was possible for the interlocutors to establish visual contact via webcams. The video windows were situated on the left side of the screen, while a window containing the textual conversation, similar to the one in the anonymity condition, was situated on the right side. The visibility condition therefore allowed for visual feedback throughout the interaction. There was, however, no opportunity for vocal communication in order to ensure that the sole means of exchanging information in each modality of the context factor was textual.

Confederate Pretest The confederates’ appearance was assessed by a group of 23 judges (psychology undergraduates) on the basis of photographs of each confederate. These photos had been taken with the computer webcam used by the participants during the experiment, and showed the confederate in situations similar to those of the CMC phase (visibility condition). The judges assessed each confederate on a 7-point scale, ranging from 1 (this person is totally characteristic of the Men category) to 7 (this person is totally characteristic of the Women category). Endpoint © 2016 Hogrefe

J. Guegan et al.: Gender Perception in Dyadic CMC

ratings meant that the confederate was a prototypical representative of his/her gender category. As mean ratings were close to the endpoints for both the male (M = 1.47, SD = 0.73) and female (M = 6.43, SD = 0.84) confederates, the confederates can be regarded as nonambiguous representatives of their respective categories.

19

endpoints of this scale corresponded to a connotative response indicating the gender of the respondent (1 = woman’s answer; 5 = man’s answer). The central, neutral zone of the scale (3) corresponded to responses for which no gender identification was possible, and which could have been given just as easily by a man or by a woman. As in the case of the questions, we selected only gender-neutral responses.

Construction of the Declarative Material Dependent Variable A total of 121 psychology undergraduates (i.e., a population similar to that of the participants) contributed to the construction of the declarative material. None of them took part in the experiment.

Questions Asked by the Two Confederates The 15 questions asked by the confederates were assessed by 15 judges. For each question, the judges were to indicate whether they thought it was a question that was more likely to be asked by a man or a woman. They did so on a 5-point scale ranging from 1 (feminine question) to 5 (masculine question), the central zone corresponding to a lack of gender connotation (3 = A man or a woman could just as easily ask this question).

By this means, we obtained 15 assessments for each question.

To select the confederates’ five questions, we looked at whether these assessments differed significantly from the norm (i.e., 3 on the scale). None of the judges’ assessments for the questions we selected differed significantly from 3, and these five questions could therefore be regarded as neutral.

Confederates’ Responses The confederates’ responses were constructed in two phases.

The first phase consisted in gathering a set of responses to each question. To this end, 41 psychology undergraduates were asked at the start of a lesson to complete a questionnaire consisting of our 15 questions. For reasons of practicality, each student responded to only half of the questions (7 or 8 questions, depending on the version). The students were asked to respond spontaneously and just as they would if the questions were asked by one of their fellow students.

In the second phase, the corpus of responses collected during the first phase was assessed by 65 judges. For each response, the judges were to indicate whether they thought it had been given by a man or a woman. The judges were informed that the original responses had all been reframed in the masculine gender. Each response was rated on a 5-point scale. The

The questionnaire administered after the CMC was designed to measure the stereotypicality of the participants’ perceptions of the confederates. It was the same questionnaire as the one used in the study by Guegan et al. (2011).

This questionnaire was based on an updated version of the Bem Sex-Role Inventory (BSRI; Bem, 1974), known as the Interpersonal BSRI (Brems & Johnson, 1990). Female participants were to rate the male confederate on nine stereotypical masculine items taken from the Interpersonal BSRI (dominant, aggressive, tendency to take sides, dynamic, confident, strong personality, quick in making decisions, leader, authoritarian). Male participants were to rate the female confederate on nine stereotypical feminine items also taken from the Interpersonal BSRI (sweet, warm, gentle, compassionate, good listener, kind, soothing, affectionate, understanding. For this study, these items were translated into French (translation and adaptation supervised by an English native speaker; Guegan et al. 2011).

Both types of items are integrated in the typology of expressive versus instrumental roles identified by Parsons and Bales (1955) and refer to the stereotypical dimensions of male and female groups (Bem, 1985). Each item was preceded by the sentence, “In your opinion, was your interlocutor . . .,” and the participants were to rate that item on a 7-point scale, ranging from 1 (not at all) to 7 (absolutely). Cronbach’s α was satisfactory for the attribution of both the feminine (α = .89) and the masculine (α = .65) characteristics1. With this material, it was possible to calculate a mean stereotypicality score for each participant, based on the ratings of the nine stereotypically feminine or masculine items.

Results The stereotypicality scores were subjected to an analysis of variance (ANOVA) with a 2 × 2 (Context [anonymity, visibility] × Confederate Sex [male, female]) design, and all factors between-participants factors. Figure 2 shows the stereotypicality scores in the different experimental conditions.

1 The coefficient α for the masculinity scale is quite moderate but enough to consider the validity of the scale. In spite of the fact that this version of the BSRI does not present all the subdimensions of masculinity (e.g., Choi, Fuqua, & Newman, 2007), it was adapted to the objectives of this research.

© 2016 Hogrefe

Swiss Journal of Psychology (2016), 75 (1), 15–23

20

J. Guegan et al.: Gender Perception in Dyadic CMC

Figure 2. Stereotypicality scores and standard errors as a function of the context and the sex of the confederate.

The ANOVA revealed a main effect of confederate sex, F(1, 84) = 15.16, p < .001, ηp2 = .15, with higher stereotypicality scores for the female confederate (M = 5.12, SD = 1.28) than for the male confederate (M = 4.40, SD = 1.57). The analysis shows no main effect of context, F(1, 84) = 3.79, p < .06, ηp2 = .04, with equivalent stereotypicality scores in the anonymity (M = 4.95, SD = 1.46) and visibility conditions (M = 4.63, SD = 1.45). Moreover, in accordance with our hypothesis, we observed a significant interaction between context and confederate sex, F(1, 84) = 6.96, p < .01, ηp2 = .08. An analysis of the simple effects revealed that, in the visibility condition, perceptions of the female confederate (M = 5.19, SD = 1.11) were more stereotyped than those of the male confederate (M = 3.98, SD = 1.53), F(1, 84) = 21.33, p < .001, ηp2 = .20. In the anonymity condition, however, stereotypicality scores for the female confederate (M = 5.06, SD = 1.44) and the male confederate (M = 4.83, SD = 1.48) were similar, F(1, 84) = 0.78, p = .37, ηp2 = .00. Therefore, while context did not influence the stereotypicality scores for the female confederate, F(1, 84) = 0.22, p = .63, ηp2 = .00, perceptions of the male confederate were more stereotyped in the anonymity condition than in the visibility condition, F(1, 84) = 11.02, p < .01, ηp2 = .12. These results confirmed our hypothesis.

Discussion This research examined gender perceptions in dyadic CMC between a man and a woman. In particular, we investigated the cognitive effects of anonymous CMC, taking into account the asymmetries characterizing gender perceptions. In accordance with the SIDE model, the study shows that the anonymity condition led to an increase in the stereotypicality of perceptions of the male confederate. Thus, in this condition, there was a significant reorganization of perceptions toward the male steSwiss Journal of Psychology (2016), 75 (1), 15–23

reotype. By contrast, neither anonymity nor visibility had an impact on stereotypicality scores for the female confederate.

These results indicate that the consequences of anonymity may depend on the asymmetry that characterizes gender categories.

Indeed, the anonymity of the CMC has a greater impact on the perceptions of a member of a dominant group – which is usually perceived as more heterogeneous and less stereotyped than dominated groups – than on the perceptions of a member of a dominated group. Thus, the cognitive effects of deindividuation may also depend on the specificities of the salient categories in the communication situation.

Moreover, the participants’ perceptions appeared to be directly conditioned by the stereotypes and social roles associated with gender identities. Despite the confederates’ standardized reactions (the same for each confederate), associated with declarative material that was gender-neutral, the participants gave a category-based reading of the CMC. Even though they were only communicating in dyads, the interlocutors can be said to have been placed in a group situation, insofar as the situation implied consistent self-categories at a group level of inclusion (e.g., Turner, 1999). Thus, as evidenced by the results of the study, the participants’ category-based perceptions were directly influenced by the context (in accordance with self-categorization theory) and, more specifically, by the anonymity of the interlocutors (in accordance with the SIDE model), while being modulated by the asymmetry of the categories.

In order to better understand the consequences of social asymmetries in CMC and to suggest future research prospects, one could articulate social categorization and the SIDE model with the reduction of uncertainty in anonymous communication. Indeed, online anonymous first meetings between men and women involve uncertainty since each person knows only the other person’s sex. According to the theory of uncertainty reduction (Berger & Calabrese, 1975), in the development of any relationship, the reduction of uncertainty is directly related to the exchange and gathering of information to predict the interlocutors’ attitudes and behaviors. This also allows us to © 2016 Hogrefe

J. Guegan et al.: Gender Perception in Dyadic CMC

get an impression of others and thus make sense of the social situation (Srull & Wyer, 1989). Tidwell and Walther (2002) therefore assumed that the reduction of social cues in CMC leads users to seek information in order to reduce uncertainty (e.g., discuss more personal themes, disclose personal information, ask more questions).

In line with these propositions, the accentuation of stereotyping and perceived homogeneity could – at least during first meetings – provide a reading grid conducive to reducing uncertainty. Since social categorization simplifies the environment (Deschamps & Devos, 1999; Rosch & Lloyd, 1978) and provides guidelines for behavior (Tajfel, 1972), by assimilating the interlocutor to the stereotype of his group (i.e., depersonalization), this process could reduce the ambiguity of anonymous CMC. Thus, if a man meets a woman anonymously on the Internet, he knows nothing about her, but he does have stereotyped knowledge about women which is all the more easier to use as the group of women is perceived as homogeneous.

Indeed, the dominated group, perceived as coherent and simplified (i.e., stereotyped and homogeneous), can be likened to a whole that generates less uncertainty (Deschamps & Beauvois, 1994). In contrast, since the dominant group is less stereotyped/homogeneous, it could generate more uncertainty. As a result, leading to an increase in stereotypicality and homogeneity, the depersonalization effects described by the SIDE model could reduce the ambiguity and uncertainty and would be even more useful if the anonymous interlocutor is a member of a dominant group.

Although these proposals were not directly examined in our study, they may explain some aspects of our results: The fact that the stereotypicality scores of the female confederate were insensitive to context could be interpreted as a ceiling effect.

In the context of anonymity, the female confederate was not seen in a more stereotyped way since the degree of stereotyping of the dominated group was sufficient to deal with the uncertainty of anonymous communication. In addition to cognitive mechanisms, the underlying process could be the use of social categorization to reduce uncertainty and to communicate with others. However, this proposal should be tested in future research to clarify new interpretations.

Finally, it should be noted that this experiment studied only situations of communication between people of the opposite sex, though it might be instructive to replicate this study with people of the same sex. This would allow us to separate the effects of the confederate’s sex from the potential influence of the participants’ sex to examine whether men and women modulate their perceptions in the same way with the same confederate. Moreover, it would allow us to compare the effect of the presence and absence of the opposite sex on the salience of gender categories (e.g., Hogg & Turner, 1987). Similarly, it would be interesting to manipulate the topic of conversation (male theme vs. female theme; as did, for example, Postmes & Spears, 2002) while using the same standardized communication protocol. The same goes for the tone of the declarative © 2016 Hogrefe

21

content. In the present experiment, the declarative material was deliberately neutral, but it would be enlightening to study the perception of gender on the basis of content assessed as male or female and expressed by the two confederates.

Moreover, individuals who communicate in chatrooms are generally identifiable by a pseudonym (sometimes in different colors, blue and pink, according to sex) and visual communication devices (e.g., Skype) often identify the names of the users (name and webcam). For reasons of ecological validity, we implemented this configuration of social cues in the present experiment. It might be interesting to incorporate intermediate configurations (e.g., nominal identification in the condition of visual anonymity) to observe nuances in the processes studied.

In conclusion, the present research contributes to a better understanding of the processes involved in CMC between people of opposite sex. This area of study is of particular interest to psychologists since many first meetings between men and women today occur on the Internet (via social networks, forums, dating sites, and so on). The results show us that, while these particular social situations can be analyzed in light of the cognitive effects of deindividuation, they may also be affected by the social asymmetry of the categories involved. Moreover, motivational effects in terms of uncertainty reduction could also be explored and future analyses of standardized interaction should open up interesting prospects for experimental research on online gender perceptions.

Acknowledgments The authors wish to thank Grégoire Zimmermann and the two reviewers for their comments and advice. The authors also thank Marie-Françoise Crété and Julien Nelson for their valuable assistance.

References Bem, S. L. (1974). The measurement of psychological androgyny.

Journal of Consulting and Clinical Psychology, 42, 155–162. doi 10.1037/h0036215 Bem, S. L. (1985). Androgyny and gender schema theory: A conceptual and empirical integration. In T. B. Sonderegger (Ed.), Psychology and gender (pp. 179–226). Lincoln NE: University of Nebraska Press.

Berger, C. R., & Calabrese, R. J. (1975). Some explorations in initial interaction and beyond: Toward a developmental theory of interpersonal communication. Human Communication Research, 1, 99–112. doi 10.1111/j.1468-2958.1975.tb00258.x Brems, C., & Johnson, M. E. (1990). Reexamination of the Bem SexRole Inventory: The Interpersonal BSRI. Journal of Personality Assessment, 55, 484–498. doi 10.1080/00223891.1990.9674086 Choi, N., Fuqua, D. R., & Newman, J. L. (2007). Hierarchical confirmatory factor analysis of the Bem Sex Role Inventory. Educational and Psychological Measurement, 67, 818–832. doi 10.1177/0013164406299106

Swiss Journal of Psychology (2016), 75 (1), 15–23

22

Deaux, K., & Major, B. (1987). Putting gender into context: An interactive model of gender-related behavior. Psychological Review, 94, 369–389. doi 10.1037/0033-295X.94.3.369 Deschamps, J.-C., & Beauvois, J.-L. (1994). Attributions intergroupes [Intergroup attributions]. In R.-Y. Bourhis & J.-P. Leyens (Eds.), Stéréotypes, discrimination et relations intergroupes (pp. 97–126). Sprimont, Belgium: Mardaga.

Deschamps, J.-C., & Devos, T. (1999). Les relations entre identité individuelle et collective ou comment la similitude et la différence peuvent covarier [The relationship between individual and collective identity or how the similarity and difference can covary]. In J.-C. Deschamps, J.-F. Morales, D. Paez, & S. Worchel (Eds.), L’identité sociale: La construction de l’individu dans les relations entre groupes (pp. 149–167). Grenoble, France: Presses Universitaires de Grenoble.

Dubrovsky, V. J., Kiesler, S., & Sethna, B. N. (1991). The equalization phenomenon: Status effect in computer-mediated and face-toface decision-making groups. Human-Computer Interaction, 6, 119–146. doi 10.1207/s15327051hci0602_2 Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum.

Fischer, A. (2011). Gendered social interactions in face-to-face and computer-mediated communication. In A. Kappas & N. C. Krämer (Eds.), Face-to-face communication over the Internet: Emotions in a web of culture, language, and technology (pp. 53–78).

Cambridge, UK: Cambridge University Press. doi 10.1017/ CBO9780511977589.005 Guegan, J., Vergnaud, E., & Moliner, P. (2011). Communications anonymes via Internet et modulation asymétrique des perceptions de genre: une approche quasiexpérimentale [Anonymous communications on the Internet and asymmetric modulation of gender perceptions: A quasiexperimental approach]. Revue Internationale de Psychologie Sociale, 24, 101–122.

Herring, S. C. (1993). Gender and democracy in computer-mediated communication. Electronic Journal of Communication, 3.

Hogg, M. A., & Turner, J. C. (1987). Intergroup behavior, self-stereotyping and the salience of social categories. British Journal of Social Psychology, 26, 325–340. doi 10.1111/j.20448309.1987.tb00795.x Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in systemjustification and the production of false consciousness. British Journal of Social Psychology, 33, 1–27. doi 10.1111/j.20448309.1994.tb01008.x Kiesler, S., Siegel, J., & McGuire, T. W. (1984). Social psychological aspects of computer-mediated communication. American Psychologist, 39, 1123–1134. doi 10.1037/0003-066X.39.10.1123 Lea, M., Spears, R., Watt, S. E., & Rogers, P. (2000). The InSIDE story: Social psychological processes affecting online groups. In T. Postmes, R. Spears, M. Lea, & S. D. Reicher (Eds.), SIDE-issues center stage: Recent developments in studies of de-individuation in groups (pp. 47–62). Amsterdam, Netherlands: KNAW. Lee, E.-J. (2004). Effects of visual representation on social influence in computer-mediated communication. Human Communication Research, 30, 234–259. doi 10.1111/j.1468-2958.2004. tb00732.x Lee, E.-J. (2007). Effects of gendered language on gender stereotyping in computer-mediated communication: The moderating role of depersonalization and gender-role orientation. Human Communication Research, 33, 515–535. doi 10.1111/j.14682958.2007.00310.x Linville, P. W., Salovey, P., & Fischer, G. W. (1986). Stereotyping and perceived distributions of social characteristics: An application to in-group-out-group perception. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice, discrimination, and racism (pp. 165–208).

San Diego, CA: Academic Press.

Lorenzi-Cioldi, F. (1988). Individus dominants et groupes dominés: Images masculines et féminines [Dominant individuals and dom-

Swiss Journal of Psychology (2016), 75 (1), 15–23

J. Guegan et al.: Gender Perception in Dyadic CMC

inated groups: Male and female images]. Grenoble, France: Presses Universitaires de Grenoble.

Lorenzi-Cioldi, F. (1994). Les androgynes [The androgynous]. Paris, France: Presses Universitaires de France.

Lorenzi-Cioldi, F. (2009). Dominants et dominés: Les identités des collections et des agrégats [Dominant and dominated: The identities of the collections and aggregates]. Grenoble, France: Presses Universitaires de Grenoble.

Lorenzi-Cioldi, F., & Dafflon, A.-C. (1999). Rapports entre groupes et identité sociale [Relationships between groups and social identity]. In J.-L. Beauvois, N. Dubois, & W. Doise (Eds.), La construction sociale de la personne (pp. 131–145). Grenoble, France: Presses Universitaires de Grenoble.

Lorenzi-Cioldi, F., Deaux, K., & Dafflon, A.-C. (1998). Group homogeneity as a function of relative social status. Swiss Journal of Psychology, 57, 255–273.

Matheson, K. (1991). Social cues in computer-mediated negotiations: Gender makes a difference. Computers in Human Behavior, 7, 137–145. doi 10.1016/0747-5632(91)90003-J Park, B., Ryan, C. S., & Judd, C. M. (1992). The role of meaningful subgroups in explaining differences in perceived variability for in-groups and out-groups. Journal of Personality and Social Psychology, 63, 553–567. doi 10.1037/0022-3514.63.4.553 Parsons, T., & Bales, R. F. (1955). Family, socialization and interaction process. New York: Free Press.

Postmes, T., & Spears, R. (2002). Behavior online: Does anonymous computer communication reduce gender inequality ? Personality and Social Psychology Bulletin, 28, 1073–1083. doi 10.1177/01461672022811006 Postmes, T., Spears, R., & Lea, M. (1998). Breaching or building social boundaries? SIDE-effects of computer-mediated communication. Communication Research, 25, 689–715. doi 10.1177/009365098025006006 Postmes, T., Spears, R., & Lea, M. (2002). Intergroup differentiation in computer-mediated communication: Effects of depersonalization. Group Dynamics: Theory, Research, and Practice, 6, 3–16. doi 10.1037/1089-2699.6.1.3 Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation phenomenon. European Review of Social Psychology, 6, 161–198. doi 10.1080/14792779443000049 Rosch, E., & Lloyd, B. B. (1978). Cognition and categorization. Hillsdale, NJ: Erlbaum.

Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press.

Spears, R., & Lea, M. (1992). Social influence and the influence of the “social” in computer-mediated communication. In M. Lea (Ed.), Contexts of computer-mediated communication (pp. 30–65). London, UK: Harvester Wheatsheaf.

Spears, R., & Lea, M. (1994). Panacea or panopticon? The hidden power in computer-mediated communication. Communication Research, 21, 427–459. doi 10.1177/009365094021004001 Spears, R., Lea, M., & Postmes, T. (2007). Computer-mediated communication and social identity. In A. N. Joinson, K. Y. A. McKenna, T. Postmes, & U.-D. Reips (Eds.), The Oxford handbook of internet psychology (Oxford Handbooks Online). Retrieved from http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/97 80199561803.001.000 1/oxfordhb-9780199561803-e-017.

Sproull, L., & Kiesler, S. (1986). Reducing social context cues: Electronic mail in organizational communication. Management Science, 32, 1492–1512. doi 10.1287/mnsc.32.11.1492 Srull, T. K., & Wyer, R. S. (1989). Person memory and judgment. Psychological Review, 96, 58–83. doi 10.1037/0033-295X.96.1.58 Tajfel, H. (1972). La catégorisation sociale [Social categorization].

In S. Moscovici (Ed.), Introduction à la psychologie sociale (Vol. 1, pp. 272–302). Paris, France: Larousse.

Tajfel, H. (1982). Social psychology of intergroup relations. Annual

© 2016 Hogrefe

J. Guegan et al.: Gender Perception in Dyadic CMC

Review of Psychology, 33, 1–39. doi 10.1146/annurev.ps.33.

020182.000245 Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relation (pp. 33–47). Monterey, CA: Brooks-Cole.

Tidwell, L. C., & Walther, J. B. (2002). Computer-mediated communication effects on disclosure, impressions, and interpersonal evaluations: Getting to know one another a bit at a time. Human Communication Research, 28, 317–348. doi 10.1111/j.14682958.2002.tb00811.x Turner, J. C. (1985). Social categorization and the self-concept: A social cognitive theory of group behavior. In E. J. Lawler (Ed.), Advances in group processes: Theory and research (Vol. 2, pp. 77–122). Greenwich, CT: JAI. Turner, J. C. (1999). Some current issues in research on social identity and self-categorization theories. In N. Ellemers, R. Spears, & B. Doosje (Eds.), Social identity: Context, commitment, content (pp. 6–34). Oxford, England: Blackwell.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell,

© 2016 Hogrefe

23

M. S. (1987). Rediscovering the social group: A self-categorization theory. Oxford, UK: Blackwell.

Williams, M. J., & Mendelsohn, G. A. (2008). Gender clues and cues: Online interactions as windows into lay theories about men and women. Basic and Applied Social Psychology, 30, 278–294. doi 10.1080/01973530802375136 Wolf, A. (2000). Emotional expression online: Gender differences in emoticon use. CyberPsychology & Behavior, 3, 827–833. doi 10.1089/10949310050191809

Jérôme Guegan Laboratoire Adaptations Travail Individu Institut de Psychologie Henri Piéron Université Paris Descartes 71 avenue Edouard Vaillant 92100 Boulogne France jerome.guegan@parisdescartes.fr

Swiss Journal of Psychology (2016), 75 (1), 15–23

Z. Tabibi Sw issJournal et al.: of Psychology Child Traffic (2016), Rule © 2016 75 Co(1), mpliance Hogrefe 25–33

Original Communication

Self-Reported Compliance with Traffic Rules in a Sample of Iranian Preschoolers Knowledge of Rules, Perception of Danger, Moral Judgment, and Self-Regulation Zahra Tabibi1, Fatemeh Grayeli2, and Mohammad Saeid Abdekhodaei1 1

Department of Psychology, Ferdowsi University of Mashhad, Iran Department of Clinical Psychology, Ferdowsi University of Mashhad, Iran

2

Abstract. Traffic injuries represent an important danger to children’s health. Safe traffic behavior requires both perceptual and cognitive abilities as well as compliance with traffic rules. The present study examines the relationship between knowledge of traffic rules, perception of traffic danger, moral judgment, self-regulation, and compliance with traffic rules among preschool children. It also examined gender differences across the study variables. A sample of 100 children aged 3–6 years participated in the study. Knowledge of traffic rules, perception of danger, and compliance with rules were assessed by interviewing each child using photos of traffic situations. Two components of self-regulation – impulse control and following an adult’s directions – were assessed using the Tower Task, the Tower Clean Up Task, and the Toy Sorting Task from the Preschool Self-Regulation Assessment. Moral judgment was assessed using Piaget’s task of stealing/clumsiness. Significant positive relationships among knowledge of rules, perception of danger, and self-reported compliance with rules were found. The ability to quickly follow an adult’s directions predicted greater self-reported compliance with rules over and above perception of danger and knowledge of rules. There were no significant gender differences in any of the variables. Using our results, educational programs for preschoolers could seek to improve their traffic knowledge of rules, perception of traffic danger, and self-regulation.

Keywords: pedestrian safety, child injury, traffic rules, impulsive control, following adults’ directions

Introduction Compared to adults, in many countries of the world children are involved in more accidents as pedestrians (World Health Organization 2004, 2009), including in Iran (Ayati 2002). In fact, childhood road traffic injuries have the highest incidence of all types of injuries in Iran to children (RezapurShahkolai, Naghavi, Vaez, Shokouhi, & Laflamme, 2009).

Significant factors influencing children’s road traffic injuries are demographic characteristics (e.g., age, sex, and socioeconomic status), environmental context, and developmental factors (Barton & Schwebel, 2007a, 2007b; Tabibi, Pfeffer, & Sharif, 2012). For example, boys and children from lower socioeconomic status (SES) backgrounds are more likely than girls and children from higher SES backgrounds to encounter traffic environments. Compared to children from higher SES, those from lower SES face more dangerous traffic environments, roads with higher traffic volume, and poorer traffic engineering.

Young children may find it difficult to cross streets safely because it requires simultaneous engagement in several cognitive and perceptual tasks (Ampofo-Boateng & Thomson, 1991; © 2016 Hogrefe DOI 10.1024/1422-4917/a000168

Dunbar, Hill, & Lewis, 2001; Dunbar, Lewis, & Hill, 1999; Hill, Lewis, & Dunbar, 2000; Pfeffer, 2005; Tabibi & Pfeffer, 2003; Tabibi et al., 2012). Several studies identified domainspecific cognitive skills (such as perceiving dangerous roadcrossing sites, route construction) relevant to pedestrians (Ampofo-Boateng & Thomson, 1991; Hill et al., 2000). Others identified more general-domain cognitive skills (such as attentional skills, executive functions) that may underlie pedestrian skills (Dunbar et al., 2001; Stavrinos et al., 2011; Tabibi et al., 2012; Whitebread & Neilson, 1999).

The failure to observe traffic safety rules has also been identified as a reason for traffic injuries (Rößger, 2011; Zavareh, 2009). Cognitive or social cognitive skills underlying compliance with traffic rules or violations have been studied extensively among drivers (Özkan, Lajunen, Chliaoutakis, Parker, & Summala, 2006; Warner, Özkan, Lajunen, & Tzamalouka, 2011). However, there are few studies on pedestrian compliance with rules among adults, and studies on pedestrian compliance with rules among children are also scarce (Harrell, 1996; Rosenbloom, 2011; Tom & Granié, 2011). The present study examines the specific and general cognitive factors important in compliance with traffic safety rules among young children.

Swiss Journal of Psychology (2016), 75 (1), 25–33

26

Development of Compliance Compliance, as a component of social competence, is the ability and willingness to modulate behavior in accordance with a specific demand or rule to be followed (Edwards & Liu, 2002).

Rule compliance develops from toddler to preschool to schoolage years (Edwards & Liu, 2002; Kochanska, Coy, & Murray, 2001). Children’s compliance behavior has been studied in a wide range of situations such as school performance, medication-taking, television-watching, and hazard avoidance (AsadiPooya, 2005; Gralinski & Kopp 1993; Morrongiello & Dawber 1998; Springer et al., 2010; Wachs, Gurkas, & Kontos, 2004).

Parents usually start teaching safety rules, including traffic safety rules, to older toddlers (Gralinski & Kopp 1993; Morrongiello, Zdzieborski, & Normand, 2010). According to Gralinski and Kopp (1993), safety rules are observed much more than any other rules by young children. MacGregor, Smiley, and Dunk (1999) also found that more than half of the children observed followed the traffic rules they had been taught when they were only unobtrusively being observed by their parents.

Compliance and Gender Studies on compliance behavior in injury-risk situations indicated a gender difference: Boys were less compliant with mothers’ requests to avoid hazards than girls; boys were also less likely than girls to avoid injury-risk hazards (Morrongiello & Dawber, 1998) or to observe pedestrian rules (Granié, 2007).

Furthermore, parents were found to perceive that their daughters complied with safety rules more than their sons (Morrongiello et al., 2010). This may explain why, across cultures, boys have higher rates of accidents and injuries than girls.

Compliance and Self-Regulation Several reasons have been proposed for the development of compliance in general and compliance with safety rules in particular. Compliance with caregivers’ requests is achieved through the development of self-regulation (Edwards & Liu, 2002; Florez, 2011; Kochanska et al., 2001). Self-regulation concerns whether children display appropriate behavior in the absence of adult supervision (Edwards & Liu, 2002). It requires the effortful control of attentional and emotional processes (Sanson, Hemphill, & Smart, 2006). Regulation develops in five stages with the first 2 years of life shaping the first stage, and the preschool years, childhood, adolescence, and adulthood covering the second, third, fourth, and final stages, respectively (Macklem, 2011). The development of self-regulation depends on the development of impulse control and knowledge of social standards (Edwards & Liu, 2002; Macklem, 2011).

Swiss Journal of Psychology (2016), 75 (1), 25–33

Z. Tabibi et al.: Child Traffic Rule Compliance

Specific to safety, poor impulse control is linked to unintentional injuries in and poor pedestrian skills of children (Barton & Schwebel, 2007a, 2007b; Schwebel, 2004). Morrongiello, Kane, McArthur, and Bell (2012) found an association between emotion dysregulation and more playground risk taking among elementary school children. Also, poor cognitive self-regulation, or executive function, is linked to greater risk-taking behavior among adolescents and adults (Gwyther & Holland, 2012; Magar, Philips, & Hosie, 2008).

Compliance and Morality Compliance may also result from a sense of personal morality and a perception of right and wrong (Rowe & Kellam, 2011; Yagil, 1998). Accordingly, people may disobey certain laws if this is not perceived as immoral (Yagil, 1998). Morality begins developing in the second year of life across cultures (Dahl, Campos, & Witherington, 2011; Helwig & Turiel, 2004; Killen & Smetana, 1999; Nucci & Turiel, 1978; Nucci & Weber, 1995). The development of morality serves to internalize social values as personal values (Turiel, 1983). With respect to safety, internalization of traffic rules is associated with compliance with traffic rules among young children (Granié, 2007) and adults (Yagil, 2000).

Compliance and Danger Perception Yagil (2000) stated that an understanding of danger is a prerequisite for compliance with safety rules. Also, the appraisal of danger involves considering other people’s safety and feeling responsible for safety (SWOV, 2010). Among adults, Yagil (2000) found that pedestrians were more likely to be noncompliant at signals when they did not perceive any danger/risk of a collision. Research with children indicates that insufficient skill in perceiving potentially dangerous situations plays an important role in their unsafe traffic behavior. Granié (2007) reported a significant relationship between pedestrian compliance with rules and perception of road danger as well as pedestrian knowledge of rules among preschool children.

Aims of the Study Traffic compliance with rules seems to involve both domaingeneral and domain-specific cognitive skills. Crucial general cognitive skills are self-regulation (Grolnick & Farkas, 2002) and moral judgment (Granié, 2007, 2009; Yagil, 1998, 2000) and important specific cognitive skills are perception and knowledge of road danger (Granié, 2007). However, the combined role of these variables in compliance with traffic rules © 2016 Hogrefe

Z. Tabibi et al.: Child Traffic Rule Compliance

among young children has not yet been explored. The present study aims to fill this gap.

Our first hypothesis was that compliance with traffic rules would be positively correlated with perception of road danger, knowledge of traffic rules, self-regulation, and moral judgment among preschool children. Our second hypothesis was that selfregulation and moral judgment would predict compliance with traffic safety rules over and above perception of road danger and knowledge of traffic rules among preschool children. We also explored gender differences in the variables.

This study focused on preschool children for four reasons: (1) traffic accidents pose a serious health problem at this age; (2) children begin playing in the street at this age; (3) the second stage of self-regulation development occurs at this age (Macklem, 2011); (4) children are expected to know some traffic rules and obey them at this age. Children are expected to know and obey some traffic rules such as sitting properly in the car or not playing on the street at this age.

Method Participants A sample of 100 preschool children (53 boys and 47 girls) with a mean age of 4.8 years (SD = 0.7 years; range: 3.1–6.0 years) participated in the study. The children were recruited from two nursery schools located in a part of the city of Mashhad, Iran, with high socioeconomic status. The head of the nursery schools gave written consent for the children to participate. In addition, oral consent was obtained from each individual child.

Instruments

27

ing the road at an intersection without traffic lights” (1 photo), “crossing the road at a blind curve” (1 photo), “not using a seat belt in a car” (1 photo), “reaching one’s hand out of a car window” (1 photo), and “crossing the road without using a pedestrian footbridge” (1 photo). The safe situations were “using a seat belt,” “playing in the park,” and “crossing the road using a pedestrian footbridge.” The safe-situation photographs were used as control situations in order to prevent a response set.

Therefore, the responses for the safe situations were not scored. The photographs were presented in random order and, for each photograph, the following questions were asked in the same order.

Perception of Traffic Danger The children were asked to categorize each item in terms of danger in two ways. First they were asked: “Is it safe or unsafe to do that?” Responses were scored 0 and 1; Cronbach’s α, based on the total score for each child, was .69. Then they were asked: “How safe/unsafe do you think it is to do that?” For this question, children were presented with a card depicting a 15cm line with green/red colored anchors at each end point, which were defined as the extreme limits of the variable being measured (e.g., not at all unsafe to extremely unsafe). Cronbach’s α, based on the total scores for each child, was .71.

Compliance with Traffic Rules The child was asked: “Do you do that?” Responses were scored 0 and 1; Cronbach’s α, based on the total score for each child, was .87.

Knowledge of Traffic Rules

Safe/dangerous traffic situations task This task was used to measure the children’s perception of traffic danger, knowledge of traffic rules, and compliance with traffic rules. Based on Granié’s (2007) study, we provided several photographs of pedestrian and vehicle passenger situations and conducted an interview. The children were presented with 13 20 × 30 cm color photographs, ten of which depicted dangerous and three of which depicted safe situations. Dangerous/safe traffic situations were taken from several educational books designed for children by the Traffic and Transport Organization of the Mashhad City Council. A total of 35 photographs were piloted with 7 children and 15 adults. Of these photographs, 13 were selected. Two criteria governed the choice of photos: (1) all children could clearly and accurately describe the situations; (2) adults identified them as dangerous/safe situations. The dangerous situations were “walking on the street” (2 photos), “playing in the street” (3 photos), “cross© 2016 Hogrefe

The child was asked whether there is a rule about what to do in each situation: “Are you allowed to do that?” Responses were scored 0 and 1; Cronbach’s α, based on the total score for each child, was .88.

The children were asked to give explanations for their responses, in order to check whether they had understood the questions and to make sure that the experimenter had understood the responses given by the children and could score them accurately. The total scores for each child were computed.

Scores ranged between 0 and 10 for each question.

Self-Regulation Tasks The Tower Task, the Tower Clean Up Task, and the Toy Sorting Task were derived from the Preschool Self-Regulation Assessment (PSRA) (Smith-Donald, Raver, Hayes, & Richardson, Swiss Journal of Psychology (2016), 75 (1), 25–33

28

Z. Tabibi et al.: Child Traffic Rule Compliance

had told him not to. While he was opening the cupboard to take some cookies, one dish fell and broke.” The children were presented with colored drawings of the stories to make the task more understandable and more concrete. Then, the children were asked: “Which of these boys did the worse thing? Why?” If the first story was chosen, we scored 1, indicating the use of outcome information in judging other’s deeds. If the second story was chosen, we scored 2, indicating the use of intention information in judging other’s deeds. Scores were between 1 and 2.

2007) and administered individually. The Tower Task requires children to take turns with the experimenter placing blocks to build a tower. It is a measure of impulse control requiring the filtering of competing stimuli. For the Tower Task, a score of 2 was given if the child successfully took turns, a score of 0 if he/she did not take turns at all, and a score of 1 if he/she occasionally took turns. For this task, the scores were between 0 and 2. The Tower Clean Up Task and the Toy Sorting Task require children to complete a task and measure how children follow an adult’s directions. That is, children were asked to sort and put away toys without playing with them. For both tasks, the only variables recorded were TTS, time taken to start taking the first toy after the instruction was given, and TTC, time taken to complete the task.

Providing evidence for the concurrent validity of the PSRA, Smith-Donald et al. (2007) found relationships between PSRA constructs and children’s social competence, behavioral problems, and early academic skills. As for reliability, Cronbach’s α for the Tower Task, Tower Clean Up Task, and Toy Sorting Task were .91, .98, and .96, respectively. It may be useful to note here that the pilot study of the tasks with children showed a cultural relevance of the tasks and seemed suitable for the children. They found the tasks interesting and grasped the instruction easily.

General Procedure The research was approved by the research committee of the Faculty of Education and Psychology, Ferdowsi University of Mashhad. The order of task administration was: safe/dangerous traffic situations task, stealing/clumsiness task, and the selected PSRA tasks.

Results We first examined the normality of each variable’s distribution.

Only TTS of the Toy Sorting Task had a skewness and kurtosis above 2.5. A log10 transformation resulted in a normal distribution. Table 1 presents the means and standard deviations of the variables, for boys and girls separately. Note that, for the traffic task, only responses to dangerous situations were analyzed. For the safe situations of “using a seat belt,” “crossing the road using a pedestrian footbridge,” and “playing in the park,” 82%, 57%, and 73% of the children, respectively, recognized them as being safe. Further analysis of the responses indicated that 43% of the children who had categorized “crossing the road using a pedestrian footbridge” as dangerous reasoned

Stealing/Clumsiness Task A pair of short stories intended to assess moral realism derived from Piaget’s (1972) work were administered individually. The stories dealt with moral stealing/clumsiness and went as follows: “Ali was playing in his room when his mother asked him to come to dinner. While Ali was walking by the table, he accidentally slipped and bumped the dishes. Fifteen dishes fell and broke.” and “One day, when Mohammad’s mother was not home, he decided to eat some cookies even though his mother Table 1. Descriptive data for all variables by sex Boys M Perception of danger

Girls SD

M

All SD

M

SD

7.3

2.1

7.1

2.3

7.2

2.2

101.7

28.8

107.2

29.0

104.1

28.9

8.0

2.6

7.9

2.8

7.9

2.7

Rule knowledge

6.8

3.6

7.8

2.7

7.3

3.3

Tower Task

1.6

.7

1.7

0.6

1.7

0.6

Rating of danger Compliance with rules

Tower Clean Up Task TTS

2.1

1.4

2.7

1.6

2.4

1.5

TTC

47.1

26.9

51.3

25.3

49.0

26.1

TTS

6.6

18.1

4.2

11.4

5.5

15.2

TTC

86.2

21.1

77.0

15.6

79.3

18.7

1.2

0.4

1.3

0.4

1.2

0.4

Toy Sorting Task

Stealing/clumsiness

Swiss Journal of Psychology (2016), 75 (1), 25–33

© 2016 Hogrefe

Z. Tabibi et al.: Child Traffic Rule Compliance

29

Table 2. Pearson/Spearman correlation coefficients between variables Measure

2

3

4

6

7

1. Perception of danger

0.89***

0.47***

0.36***

0.05

–0.09

–0.10

0.04

0.01

–0.09

2. Rating of danger

1

0.46***

0.40***

0.03

–0.07

–0.05

–0.01

–0.03

–0.07

1

0.75***

0.07

0.11

0.002

–0.31**

–0.24

–0.11

1

0.04

0.14

0.06

–0.21*

–0.03

–0.13

–0.02

0.09

–0.05

0.04

0.14

0.15

–0.07

0.04

–0.04

0.04

0.27*

0.11

1

0.24

–0.16

3. Compliance with rules 4. Knowledge of rules 5. Tower Task

5

1

8

9

10

Tower Clean Up Task 6. TTS

1

7. TTC

1

Toy Sorting Task 8. TTS 9. TTC

1

0.02

10. Stealing/clumsiness

1

Note. ***p < .001, ** p < .01, *p < .05.

that the child could fall off the bridge and 43% said that the child was not accompanied by an adult. For “playing in the park,” 18% of the children who had categorized it as dangerous reasoned that the child was not accompanied by an adult. The remaining 78% gave unrelated reasons (e.g., there are snakes in the grass, the ball will roll away).

Independent t-tests did not reveal any significant differences between boys and girls in any of the variables. Note that for all analyses the lowest bound of the p-values considered was .05.

Therefore, all correlations were computed using the scores for all children.

The relationships between the variables were calculated using Pearson’s correlation coefficient. For the stealing/clumsiness task, Spearman’s rho correlation coefficient was used. Table 2 presents the results.

Table 2 shows that compliance with rules, perception of danger, rating of danger, and knowledge of rules were significantly related to each other. Compliance with rules and knowledge of rules had negative relationships with time taken to start the Toy Sorting Task: The higher the scores in compliance with rules and knowledge of rules, the less time taken to pick up the first toy. Also, time taken to complete the Tower Clean Up Task was significantly related to the time taken to complete the Toy Sorting Task. The stealing/clumsiness task scores were not significantly related to any variables.

Next, a two-step hierarchical regression analysis was computed with compliance with traffic rules as the criterion. This was computed to test the second hypothesis, namely, that among preschool children self-regulation and moral judgment can predict compliance with traffic rules over and above perception of danger and knowledge of rules. Because of the resulting correlation coefficients, the following procedure was taken: Since the stealing/clumsiness task was not significantly correlated with compliance with traffic rules, the variable was not entered in the regression. Also, perception of danger and rating of danger were highly correlated. In order to prevent © 2016 Hogrefe

Table 3. Significant predictor variables for rule compliance for all children B

SE

β

t

Step 1 Perception of danger

.29

.09

0.24

3.37***

Rule knowledge

.54

.06

0.66

9.41***

Perception of danger

.33

.08

0.27

4.01***

Rule knowledge

.49

.06

0.59

8.49***

Toy Sorting (TTS)

–.04

.01

–0.22

–3.31***

Step 2

Note. *** ≤ .001.

multicollinearity, we entered only perception of danger. In addition, we entered the only significant variable of the Toy Sorting task, namely, time taken to start the Toy Sorting Task, in the regression. Therefore, perception of danger and knowledge of rules were entered in the first step and TTS of the Toy Sorting Task was entered in the second step. For the variables perception of danger and knowledge of rules, there was no evidence of multicollinearity in the analysis. Tolerance was .87 and VIF 1.1. This analysis was conducted for all children. The results are reported in Table 3.

For all children, the results indicated that both steps were significant. The first step variables of perception of danger and knowledge of rules significantly predicted compliance with rules, F(2, 91) = 70.2, p < .001, R2 = .61, ΔR2 = .60. The second step variables of perception of danger, knowledge of rules, and TTS of the Toy Sorting task were significant, F(3, 90) = 55.5, p < .001, R2 = .65, ΔR2 = .64. The Toy Sorting Task was able to add as much as 4% (Fchange(1.90) = 10.9, p < .001) to the prediction.

Next, we analyzed the explanations children provided for their danger responses. Children’s responses to the question “Why do you think it is dangerous?” were divided into three categories: behavioral, consequences, and danger factor. BeSwiss Journal of Psychology (2016), 75 (1), 25–33

30

Z. Tabibi et al.: Child Traffic Rule Compliance

Table 4. Percentage of responses given to the question “Why do you think it is dangerous?” for all children, and for boys and girls separately Category

All children (N = 617)

Boys (n = 321)

Girls (n = 296)

Being on the street

29.82

31.15

28.37

No seat belt/hands out

18.80

19.62

17.90

Running in street

10.53

10.59

10.47

Not being with adults

10.58

6.58

15.20

Behavioral

Not using pedestrian footbridge Total

1.13

1.86

0.33

70.86

69.80

72.27

21.23

21.18

21.28

1.78

0.62

3.04

Consequence Will be hit Getting kidnapped Total

23.01

21.8

24.32

Dangerous factor Speed of cars

1.94

3.42

0.33

Number/distance/direction of cars

1.29

1.55

1.01

Being distracted/not looking

2.26

2.49

2.02

Not being seen

0.32

0.62

0

Total

5.81

8.08

3.36

havioral refers to responses that mention what the child is doing or not doing or what the child is expected to do or not do.

Consequences refers to responses that mention the possible consequence of what the child is doing. Danger factor refers to responses that indicate what makes the situation dangerous or safe. For the analysis, the percentage of each response was calculated against the total number of danger responses. The calculation was conducted for all children and separately for boys and girls. The results are reported in Table 4.

As can be seen in Table 4, the most frequent responses belonged to the behavioral category. Children mostly mentioned that “being in the street” made the situation dangerous. The responses least frequently belonged to the danger factor category. In this category, “being distracted and not looking” was the most frequent response. Nearly a quarter of the responses belonged to the consequences category. The responses given by boys and girls did not reveal many differences. However, girls were more likely than boys to reason that “not being with adults” and “getting kidnapped” make the situation dangerous (15.2% vs. 6.58%, and 3.04% vs.

0.62%, respectively). Boys were more likely than girls to reason that “speed of cars” is a danger factor (3.42% vs.

0.33%).

Discussion This study found that young children who possessed a greater understanding of dangerous traffic situations were more likely to be aware of the rules for those situations and to comply with Swiss Journal of Psychology (2016), 75 (1), 25–33

the rules. Also, young children who knew more about traffic rules were more likely to comply with them. Additionally, perception of dangerous traffic situations was significantly related to danger ratings. Young children who perceived a traffic situation as being dangerous estimated the extent of danger to a higher degree. These results support the findings of Granié’s (2007) as well as Morrongiello and Matheis’ (2007) studies, which were conducted in other countries.

Furthermore, this study showed that the reasons children gave about why the presented situations were dangerous were most likely to fit the behavioral category. That is, young children mainly reasoned that what the child in the photograph was doing was dangerous: The child’s behavior as such was dangerous. For example, the child in the photograph was engaging in dangerous behavior because he/she was in the street and the street is a dangerous place; or because his/her hand was reaching out of the car window. This type of response may indicate what the child is expected to do or not expected to do. Almost 71% of young children’s responses were correctly categorized as dangerous traffic behavior.

However, the children were less able to realize what factors make the behavior dangerous. Close to 6% of the children’s responses specified that playing in the street is dangerous because, for example, one could get distracted. Surprisingly, a small percentage of responses about dangerous situations stated the possible consequence of what the child was doing.

Nearly 23% of the responses maintained that, for example, since the child in the photograph is playing in the street, he/she could be hit by a car.

There were no significant differences between boys and girls in perceiving dangerous traffic situations, estimates of © 2016 Hogrefe

Z. Tabibi et al.: Child Traffic Rule Compliance

traffic danger, knowledge of rules, and self-reported compliance with the traffic rules. In the present study, boys and girls almost always gave the same reasons for dangerous situations. However, there were small differences in a few of the reasons given. In comparison to boys, girls were more likely to reason that “not being with adults” and “getting kidnapped” make the behavior dangerous. Compared to girls, boys were more likely to reason that “speed of cars” is a dangerous factor (3.42% vs. 0.33%). The lack of gender differences in perceiving traffic dangers and in estimating the amount of danger is consistent with Granié’s (2007), Hill et al.’s (2000), and Hillier and Morrongiello’s (1998) findings.

However, Granié (2007) found significant differences between young girls and boys in self-reported compliance with traffic rules and knowledge of traffic rules, which contrasts with the results of the present study. One reason could be that the sample in the present study was recruited from areas of the city with relatively high socioeconomic status (SES).

In Iran, the parents of children from high SES areas are more likely to supervise their children than those from low SES areas (Hadai, Soori, Mehrabi, & Alamdar, 2007). Parental supervision may influence children’s traffic safety behavior and knowledge (Morrongiello & Barton, 2009). The other reason could be that the tasks were derived from educational books available in nursery schools. Therefore, both sexes were taught traffic rules in the same way. However, these are tentative explanations that require further research.

The children’s performance on traffic tasks was not significantly related to their performance on self-regulation tasks – with one exception: In the Toy Sorting Task, children who immediately responded to the examiner’s request to start sorting the toys were more likely to comply with traffic rules.

They also possessed more knowledge about traffic rules. The Toy Sorting Task assesses how well young children follow adults’ directions, which is a measure of self-regulation.

Contrary to our expectation, there were no significant relationships between young children’s performance on the traffic task and on Piaget’s moral task. Young children who were able to reason based on the intention in Piaget’s moral task did not necessarily reported higher compliance with traffic rules. A total of 76% of the children said that Ali, who accidentally broke 15 dishes, did the worse thing. This contradicted the results of studies that found an association between morality and compliance with rules among adults and young children (Granié, 2007; Rowe & Kellam, 2011; Yagil, 1998). However, Granié assessed the morality of children according to social domain theory (Turiel, 1983, 1998). Also, three-quarters of the children scored at the lower level for the dichotomous Piagetian task, leaving scant sensitivity to show a relationship.

The regression results showed that perception of danger, knowledge of traffic rules, and time taken to start the Toy Sorting Task were associated with self-reported compliance with traffic rules. Young children’s knowledge about traffic rules was the strongest predictor for self-reported compli© 2016 Hogrefe

31

ance with traffic rules. It was interesting to find that young children’s ability to immediately follow adults’ instructions had an independent effect on their self-reported compliance with traffic rules. This is consistent with previous findings that children with higher self-regulation ability are less likely to take playground risks, are more competent in social and academic skills, and in observing rules (Florez, 2011; Kochanska et al., 2001; Macklem, 2011; Morrongiello et al., 2012; Sanson et al., 2006).

The results of the present study highlight two points for the design of prevention programs. One is the importance of young children’s ability to understand traffic safety rules and their ability to perceive safe/dangerous traffic situations. The young children in this study only perceived 617 of a possible 1,000 dangerous traffic situations as being dangerous. The second point is the importance of self-regulation in order for young children to comply with traffic safety rules.

However, the study has limitations. First, of the five measures of self-regulation, only one was significantly associated with compliance with rules. Thus, the result should be interpreted with great caution due to multiple analyses. One possible explanation may be the order in which the self-regulation tasks were administered. Because of the instructions, the tasks were administered in the same order for all children.

Perhaps the relationship only turned up on the second task performed because more discipline was required to get going on yet another task. Another issue is that some measures exhibit greater variance than others. For example, TTS had lower variance in the Tower Clean Up Task (< 3 s) than in the Toy Sorting Task (6.6 s for boys and 4.2 s for girls). This may partly explain the absence of a significant effect of TTS for the Tower Clean Up Task. The Tower Clean Up Task may be less cognitively demanding than the Toy Sorting Task. Second, the present study used a self-report measure to assess children’s compliance with different traffic rules. There might have been demand effects: When children know what is allowed, they might signal obedience. The key question then is whether this carries over to behavior. In the research literature, self-report is a common methodology for investigating adult behaviors, and there is some evidence for the relationship between reported and actual behavior (Grolnick & Farkas, 2002; Victoir, Eertmans, Van den Bergh, & Van den Broucke, 2005). Nevertheless, observing the actual traffic behavior of children or asking parents about their perceptions of their children’s compliance with traffic safety rules might increase the reliability of this measure (MacGregor et al., 1999; Zeedyk, Wallace, Carcary, Jones, & Larter, 2001).

Third, to measure self-regulation, we used only two of its components, namely, following adults’ directions and impulse control. It is well established that executive functions are nonunitary, and that tasks often measure various subcomponents (e.g., see Miyake et al., 2001). Therefore, other components of self-regulation such as attention, resisting temptation, and the ability to inhibit responses should be assessed.

Swiss Journal of Psychology (2016), 75 (1), 25–33

32

References Ampofo-Boateng, K., & Thomson, J. A. (1991). Children’s perception of safety and danger on the road. British Journal of Psychology, 82, 487–505. doi 10.1111/j.2044-8295.1991.tb02415.x Asadi-Pooya, A. A. (2005) Drug compliance of children and adolescents with epilepsy. Seizure, 14, 393–395. doi 10.1016/j.seizure.

2005.05.003 Ayati, E. (2002). [A survey of road accidents in the city of Mashhad: Forming computer-based information and analyzing data based on TRL-MAAP]. Mashhad, Iran: Shahrdari Mashhad.

Barton, B. K., & Schwebel, D. C. (2007a). The influences of demographics and individual differences on children’s selection of risky pedestrian routes. Journal of Pediatric Psychology, 32, 343–353. doi 10.1093/jpepsy/jsl009 Barton, B. K., & Schwebel, D. C. (2007b). The roles of age, gender, inhibitory control, and parental supervision in children’s pedestrian/safety. Journal of Pediatric Psychology, 32, 517–526. doi 10.1093/jpepsy/jsm014 Dahl, A., Campos, J. J., & Witherington, D. C. (2011). Emotional action and communication in early moral development. Emotion Review, 3, 147–157. doi 10.1177/1754073910387948 Dunbar, G. L., Hill, R., & Lewis, V. (2001). Children’s attentional skills and road behavior. Journal of Experimental Psychology: Applied, 7, 227–234. doi 10.1037//1076-898X.7.3.227 Dunbar, G. L., Lewis, V., & Hill, R. (1999). Control processes and road crossing skills. The Psychologist, 12, 398–399.

Edwards, C. P., & Liu, W.-L. (2002). Parenting toddlers. In M. H. Bornstein (Ed.), Handbook of parenting: Children and parenting (2nd ed., Vol. 1, pp. 45–71). Mahwah, NJ: Erlbaum.

Florez, I. R. (2011). Developing young children’s self-regulation through everyday experiences. Young Children, 66(4), 46–51.

Gralinski, J. H., & Kopp, C. B. (1993). Everyday rules for behavior: Mothers’ requests to young children. Developmental Psychology, 29, 573–584. doi 10.1037/0012-1649.29.3.573 Granié, M. A. (2007). Gender differences in preschool children’s declared and behavioral compliance with pedestrian rules. Transportation Research Part F: Traffic Psychology and Behavior, 10, 371–382. doi 10.1016/j.trf.2007.02.002 Granié, M. A. (2009). Effects of gender, sex-stereotype conformity, age and internalization on risk-taking among adolescent pedestrians.

Safety Science, 47, 1277–1283. doi 10.1016/j.ssci.2009.03.010 Grolnick, W. S., & Farkas, M. (2002). Parenting and the development of children’s self-regulation. In M. H. Bornstein (Ed.), Handbook of parenting: Practical issues in parenting (2nd ed., Vol. 5, pp. 89–111). Mahwah, NJ: Erlbaum.

Gwyther, H., & Holland, C. (2012). The effect of age, gender and attitudes on self-regulation in driving. Accident Analysis & Prevention, 45, 19–28. doi 10.1016/j.aap.2011.11.022 Hadadi, M., Soori, H., Mehrabi, Y., & Alamdar, S. (2007). [The role of parents on road traffic injuries of children aged 6 to 9 years].

Journal of Medical Organization of the Islamic Republic of Iran, 25, 170–177.

Harrell, W. A. (1996). The safety of older pedestrians at signal-controlled crossings. International Journal of Aging & Human Development, 42, 65–79. doi 10.2190/9UGV-26V6-WX64-4YFE Helwig,C. C., & Turiel, E. (2004). Children’s social and moral reasoning.

In P. K. Smith & C. H. Hart (Eds.), Blackwell handbook of childhood social development (pp. 475–490). Oxford, UK: Blackwell.

Hill, R., Lewis, V., & Dunbar, G. L. (2000). Young children’s concepts of danger. British Journal of Developmental Psychology, 18, 103–119. doi 10.1348/026151000165607 Hillier, L. M., & Morrongiello, B. A. (1998). Age and gender differences in school-age children’s appraisals of injury risk. Journal of Pediatric Psychology, 23, 229–238. doi 10.1093/jpepsy/23.4.229 Killen, M., & Smetana, J. G. (1999). Social interactions in preschool

Swiss Journal of Psychology (2016), 75 (1), 25–33

Z. Tabibi et al.: Child Traffic Rule Compliance

classrooms and the development of young children’s conceptions of the personal. Child Development, 70, 486–501. doi 10.1111/1467-8624.00035 Kochanska, G., Coy, K. C., & Murray, K. T. (2001). The development of self-regulation in the first four years of life. Child Development, 72, 1091–1111. doi 10.1111/1467-8624.00336 MacGregor, C., Smiley, A., & Dunk, W. (1999). Identifying gaps in child pedestrian safety: Comparing what children do with what parents teach. Transportation Research Record, 1674(1), 32–40. doi 10.3141/1674-05 Macklem, G. L. (2011). Evidence-based school mental health services: Affect education, emotion regulation training, and cognitive behavioral therapy. New York: Springer-Verlag.

Magar, E. C. M., Philips, L. H., & Hosie, J. A. (2008). Self-regulation and risk-taking. Personality and Individual Differences, 45, 153–159. doi 10.1016/j.paid.2008.03.014 Miyake, A., Friedman, N. P., Rettinger, D. A., Shah, P., Arbor, A., & Hegarty, M. (2001). How are visuospatial working memory, executive functioning, and spatial abilities related? A latent-variable analysis. Journal of Experimental Psychology: General, 130, 621–540.

Morrongiello, B. A., & Barton, B. K. (2009). Child pedestrian safety: Parental supervision, modeling behaviors, and beliefs about child pedestrian competence. Accident Analysis & Prevention, 41, 1040–1046. doi 10.1016/j.aap.2009.06.017 Morrongiello, B. A., & Dawber, T. (1998). Toddlers’ and mothers’ behaviors in an injury-risk situation: Implications for sex differences in childhood injuries. Journal of Applied Developmental Psychology, 19, 625–639. doi 10.1016/S0193-3973(99)80059-8 Morrongiello, B. A., & Matheis, S. (2007). Understanding children’s injury-risk behaviors: The independent contributions of cognitions and emotions. Journal of Pediatric Psychology, 32, 926–937. doi 10.1093/jpepsy/jsm027 Morrongiello, B. A., Zdzieborski, D., & Normand, J. (2010). Understanding gender differences in children’s risk taking and injury: A comparison of mothers’ and fathers’ reactions to sons and daughters misbehaving in ways that lead to injury. Journal of Applied Developmental Psychology, 31, 322–329. doi 10.1016/j. appdev.2010.05.004 Morrongiello, B. A., Kane, A., McArthur, B. A., & Bell, M. (2012).

Physical risk taking in elementary-school children: Measurement and emotion regulation issues. Personality and Individual Differences, 52, 492–496. doi 10.1016/j.paid.2011.11.003 Nucci, L. P., & Turiel, E. (1978). Social interactions and the development of social concepts in preschool children. Child Development, 49, 400–407. doi 10.2307/1128704 Nucci, L. P., & Weber, E. K. (1995). Social interactions in the home and the development of young children’s conceptions of the personal. Child Development, 66, 1438–1452. doi 10.2307/1131656 Özkan, T., Lajunen, T., Chliaoutakis, J. E., Parker, D., & Summala, H. (2006). Cross-cultural differences in driving behaviors: A comparison of six countries. Transportation Research Part F: Traffic Psychology and Behavior, 9, 227–242. doi 10.1016/j.trf.2006.01.002 Pfeffer, K. (2005). Rural and urban children’s understanding of safety and danger on the road. In G. Underwood (Ed.), Traffic and transport psychology: Theory and application (pp. 27–36). Amsterdam, The Netherlands: Elsevier.

Piaget, J. (1972). The moral judgment of the child. London, UK: Routledge and Kegan Paul.

Rezapur-Shahkolai, F., Naghavi, M., Vaez, M., Shokouhi, M., & Laflamme, L. (2009). Injury incident, healthcare consumption and avenues for prevention: A household survey on injury in rural Twiserkan, Iran. Public Health, 123, 384–389. doi 10.1016/ j.puhe.

2009.03.010 Rößger, L. (2011, November). Acceptance toward the traffic rules and compliance. Paper presented at the Second International Workshop on Acceptance of Technological and Organizational Changes in Transport, Paris.

© 2016 Hogrefe

Z. Tabibi et al.: Child Traffic Rule Compliance

Rosenbloom, T. (2011). Traffic light compliance by civilians, soldiers and military officers. Accident Analysis & Prevention, 43, 2010–2014. doi 10.1016/j.aap.2011.05.019 Rowe, J., & Kellam, C. (2011). Ethics and moral development: Core ingredients of a compliance culture. Home Health Care Management Practice, 23(1), 55–59. doi 10.1177/1084822310382020 Sanson, A., Hemphill, S. A., & Smart, D. (2006). Temperament and social development. In P. K. Smith & C. H. Hart (Eds.), Blackwell handbook of childhood social development (pp. 97–116). Oxford, UK: Blackwell.

Schwebel, D. C. (2004). Temperamental risk factors for children’s unintentional injury: The role of impulsivity and inhibitory control. Personality and Individual Differences, 37, 567–578. doi 10.1016/j.paid.2003.09.027 Smith-Donald, R., Raver, C. C., Hayes, T., & Richardson, B. (2007).

Preliminary construct and concurrent validity of the Preschool Self-regulation Assessment (PSRA) for field-based research.

Early Childhood Research Quarterly, 22, 173–187. doi 10.1016/ j.ecresq.2007.01.002 Springer, A. E., Kelder, S. H., Barroso, C. S., Drenner, K. L., Shegog, R., Ranjit, N., & Hoelscher, D. M. (2010). Parental influences on television watching among children living on the Texas-Mexico border. Preventive Medicine, 51, 112–117. doi 10.1016/j.ypmed.2010.05.013 Stavrinos, D., Biasini, F. J., Fine, P. R., Hodgens, J. B., Khatri, S., Mrug, S., & Schwebel, D. C. (2011). Mediating factors associated with pedestrian injury in children with attention-deficit/hyperactivity disorder. Pediatrics, 128, 296–302. doi 10.1542/peds.2010-3829 SWOV Fact Sheet (2010). Hazard perception in traffic. Leidschendam, Netherlands: Institute for Road Safety Research. Retrieved from http://www.swov.nl/rapport/Factsheets/UK/FS_Training_hazard_perception.pdf Tabibi, Z., & Pfeffer, K. (2003). Choosing a safe place to cross the road: The relationship between attention and identification of safe and dangerous road-crossing sites. Child Care, Health and Development, 29, 237–244. doi 10.1046/j.1365-2214.2003.00336.x Tabibi, Z., Pfeffer, K., & Sharif, J. T. (2012). The influence of demographic factors, processing speed and short-term memory on Iranian children’s pedestrian skills. Accident Analysis & Prevention, 47, 87–93. doi 10.1016/j.aap.2012.01.013.

Tom, A., & Granié, M. A. (2011). Gender differences in pedestrian compliance with rules and visual search at signalized and unsignalized crossroads. Accident Analysis & Prevention, 43, 1794–1801. doi 10.1016/j.aap.2011.04.012 Turiel, E. (1983). The development of social knowledge: Morality and convention. Cambridge, MA: Cambridge University Press.

Turiel, E. (1998). The development of morality. In N. Eisenberg (Ed.),

© 2016 Hogrefe

33

Handbook of child psychology: Social, emotional and personality development (Vol. 3, pp. 863–932). New York: Wiley.

Victoir, A., Eertmans, A., Van den Bergh, O., & Van den Broucke, S. (2005). Learning to drive safely: Social-cognitive responses are predictive of performance rated by novice drivers and their instructors. Transportation Research Part F: Traffic Psychology and Behavior, 8(1), 59–74. doi 10.1016/j.trf.2005.01.002 Wachs, T. D., Gurkas, P., & Kontos, K. (2004). Predictors of preschool children’s compliance behavior in early childhood classroom settings. Journal of Applied Developmental Psychology, 25, 439–457. doi 10.1016/j.appdev.2004.06.003 Warner, H. W., Özkan, T., Lajunen, T., & Tzamalouka, G. (2011). Crosscultural comparison of drivers’ tendency to commit different aberrant driving behaviors. Transportation Research Part F: Traffic Psychology and Behavior, 14, 390–399. doi 10.1016/j.trf.2011.04.006 Whitebread, D., & Neilson, K. (1999). Learning to cross the road: Cognition in action, The Psychologist, 12, 403–405.

World Health Organization. (2004). World report on road traffic injury prevention: Summary. Geneva, Switzerland: Author.

World Health Organization. (2009). Global status report on road safety: Time for action. Geneva, Switzerland: Author.

Yagil, D. (1998). Gender and age-related differences in attitudes toward traffic laws and traffic violations. Transportation Research Part F: Traffic Psychology and Behavior, 1, 123–135. doi 10.1016/S1369-8478(98)00010-2 Yagil, D. (2000). Beliefs, motives and situational factors related to pedestrians’ self-reported behavior at signal-controlled crossings. Transportation Research Part F: Traffic Psychology and Behavior, 3, 1–13. doi 10.1016/S1369-8478(00)00004-8 Zavareh, D. K. (2009). Toward safety promotion among road users: Epidemiology and prevention of road traffic injuries in Iran (Unpublished doctoral dissertation). Karolinska Institutet, Stockholm, Sweden.

Zeedyk, M. S., Wallace, L., Carcary, B., Jones, K., & Larter, K. (2001).

Children and road safety: Increasing knowledge does not improve behavior. British Journal of Educational Psychology, 71, 573–594. doi 10.1348/000709901158686

Zahra Tabibi Department of Psychology Ferdowsi University of Mashhad Mashhad Korasan Razavi Iran tabibi@um.ac.ir

Swiss Journal of Psychology (2016), 75 (1), 25–33

A. Mignon et al.: Judgment of Agency, Sw issJournal Competence, of Psychology Morality, (2016), © and 2016 75Sociability (1), Hogrefe 35–45

Original Communication

Effect of Political Orientation on Judgment of Agency, Competence, Morality, and Sociability The French Presidential Election of 2012 Astrid Mignon1, Patrick Mollaret2, Odile Rohmer3, and Céline Bagès4 1

SCALab CNRS UMR 9193, U.F. R. de Psychologie, Université de Lille, France

2

Laboratoire Parisien de Psychologie Sociale, Université de Paris 8, France Laboratoire de Psychologie des Cognitions, Université de Strasbourg, France

3 4

Psitec, Université de Lille, France

Abstract. This study tested the effect of voter’s political orientation on the importance they placed on agency, competence, morality, and sociability as qualities required for being President of the French Republic. It furthermore assessed the participants’ judgments of the five main candidates of the French Presidential Election of 2012 on those dimensions. The results showed that, compared to left-wing voters, right-wing voters placed greater importance on agency and less importance on morality and sociability as qualities required for being president.

Right-wing voters differentiated between candidates with respect to agency more than competence, whereas the reverse pattern was observed for left-wing voters. Voters differentiated between candidates with respect to morality as much as sociability. Finally, the participants’ ratings of candidates showed a partisan effect: Left-wing voters rated left-wing candidates as being more competent, moral, and sociable, the same being true of right-wing candidates by right-wing voters. Ratings of candidates’ agentic qualities were less dependent on the voter’s political orientation. These data provide support for how people perceived qualities necessary for being President in France.

Keywords: agency, competence, morality, sociability, left- and right-wing

Numerous studies have focused on voters’ perception of the personality of politicians (e.g., Caprara, Vecchione, & Schwartz, 2009; Schwartz, Caprara, & Vecchione, 2010). The very fact that citizens are inclined to describe candidates’ personalities may be seen as a consequence of the decline of ideologies (Bennett, 2012). As a consequence, the public at large may conceptualize issues in terms of candidates more than in terms of political orientations or parties. Indeed, research on people’s perceptions of politicians indicates that they tend to prefer politicians with a personality profile similar to their own (Caprara, Vecchione, Barbaranelli, & Fraley, 2007). This framework, which makes the personalization of politicians a nonideological process, may explain why the role of voters’ political orientation in their perception of politicians’ personalities has received so little attention. Some studies found evidence that left- and right-wing voters differ with respect to how they perceive politicians’ personalities (e.g., Roets & Van Hiel, 2009). This paper aims to show that the Big Two personality dimensions (i.e., agency/competence vs. sociability/morality) provide a useful framework for making predictions concerning how voters’ political orientation determines (1) the qualities voters believe one must possess to exercise the function of Pres© 2016 Hogrefe DOI 10.1024/1422-4917/a000169

ident of the French Republic and (2) how voters judge real candidates for the office of President of France in the 2012 election.

The Two Basic Dimensions of Social Judgment in Politics There is growing evidence in the literature that social judgment maps onto two fundamental dimensions, regardless of whether the targets are individuals, groups, or oneself (e.g., Abele, Cuddy, Judd, & Yzerbyt, 2008; Abele & Wojciszke, 2007; Beauvois & Dubois, 2009; Fiske, Cuddy, Glick, & Xu, 2002; Peeters, 1992). These two dimensions of judgment are often referred to as competence and warmth (Fiske et al., 2002) or as agency and communion (Abele & Wojciszke, 2007). Agency and competence refer to the efficient pursuit of one’s goals; this dimension is considered self-profitable (Peeters, 1992) and is linked to individualism (Dubois & Beauvois, 2005). Communion and warmth refer to the formation and maintenance of the quality Swiss Journal of Psychology (2016), 75 (1), 35–45

36

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

of social relationships and is frequently conceptualized as other-profitable, meaning that the consequences of being warm are more relevant for people surrounding the trait holder than for the trait holder himself (Peeters, 1992; Wentura, Rothermund, & Bak, 2000; Wojciszke, BaryBa, Parzuchowski, Szymków, & Abele, 2011).

Despite the validity and usefulness of this bidimensional view of social judgment, recent work suggested that the two fundamental dimensions can be broken down into various subcomponents.

In doing so, it conceptualized agency and competence more narrowly than the classical uses of these terms and investigated the relative contribution of each of these components in predicting social status (Carrier, Louvet, Chauvin, & Rohmer, 2014; Dubois, 2010). These authors defined agency as the motive to promote the self (e.g., ambitious, independent, and self-confident) and competence as efficiency in the achievement of tasks (e.g., competent, intelligent, and capable). Through factor analyses, they found that agency and competence operate as distinct dimensions in social judgment and that high-status targets are evaluated more favorably than low-status targets with respect to agency, whereas there was no difference between the two types of targets with respect to competence. In the same vein, even if communion and warmth are defined and operationalized in a consistent manner (Abele et al., 2008), empirical work has shown that communion/warmth encompasses two specific subcomponents: sociability, which reflects empathy and understanding for others’ interests (e.g., warm, likeable, friendly), and morality, which refers to the standards of human virtue and correctness that direct social behaviors (e.g., sincere, honest, trustworthiness). It has been shown that sociability and morality emerged as distinct concepts in evaluations of behaviors, and that people are more concerned about gathering information about morality than sociability to form an impression about targets (Brambilla, Rusconi, Sacchi, & Cherubini, 2011; Leach, Ellemers, & Barreto, 2007).

Social judgments were also analyzed in terms of Beauvois and Dubois’ (2009) social value theory, according to which the two fundamental dimensions serve to communicate social affordances, that is, opportunities to act on others. Here, personality traits are considered evaluative labels that communicate evaluative knowledge about others. More precisely, people not only use traits to describe the target’s stable dispositions, but also to communicate information about what one can (or must) do to the target (Mignon & Mollaret, 2002, 2012). For example, people may consider a person to be honest and communicate that he/she is “a person you can trust” (evaluative knowledge). In this theoretical framework, the two fundamental dimensions are social desirability and social utility. Social desirability refers to an individual’s capacity to gain others’ social approval, that is, to provoke positive (or negative) reactions. Social utility refers to an individual’s capacity to satisfy (or not) the functional requirements of a given organization and therefore the degree to which he/she can succeed (or not) in the organization. In this approach, social desirability is composed of morality and sociability, and social utility is composed Swiss Journal of Psychology (2016), 75 (1), 35–45

by competence and agency (Dubois & Beauvois, 2012). In line with this theoretical approach, this paper aims to show that political judgments can be considered a type of social value judgment in a highly evaluative context, namely, a presidential election.

Competence and morality appear to be the two prevalent dimensions guiding voters’ perception of politicians. The role of competence in determining voters’ choice of candidates is obvious (Kinder & Sears, 1985; Page, 1978). The candidates’ competence is quickly inferred from nonverbal information and the most competent-looking political candidates are more likely than others to produce an overall positive impression (Olivola & Todorov, 2010; Poutvaara, Jordahl, & Berggren, 2009). The role of morality in determining voting behavior has been shown by Wojciszke and Klusek (1996), who found that honesty and integrity strongly determine voters’ attitude toward politicians. Morality probably plays a more important role than warmth in voters’ perception of politicians. Nevertheless, considering voters’ political orientation may introduce another aspect of their perception of politicians’ personality.

Personality Judgment of Politicians and Political Orientations Some indications are available to predict how political orientation may influence voters’ perception of politicians with respect to competence, agency, morality, and warmth.

Competence Previous findings showed that the perception of competence drives voting intention and suggested, quite evidently, that voters of any ideological orientation are reluctant to elect incompetent candidates (e.g., Cislak & Wojcizke, 2006). More direct support for this is provided by Roets and Van Hiel (2009). Adopting the five-factor approach to personality assessment (McCrae & Costa, 2006), these authors investigated perceptions of a politician’s ideal personality among voters with different political orientations and showed that conscientiousness – the closest dimension to competence – was the most important dimension in determining the ideal politician for all voters. Competence, defined as the cognitive ability to fulfill intentions or goals (Peeters & Czapinski, 1990), is certainly a consensual and nonideological dimension on which politicians are perceived.

Agency According to Dubois (2010), agency traits (dominant, ambitious, at ease with competition) are the most socially useful © 2016 Hogrefe

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

for success in competitive contexts. Dubois found that agentic traits, compared to competence traits, were used by subordinates to describe agents ranking higher in social hierarchies (i.e., bosses). Agentic traits refer to a social attitude that promotes interindividual competition. Because (1) the presidential function is the highest social function of a political career and (2) the presidential election is a competition between candidates, we reasoned that agency could be a quality that voters regard as required for exercising the function of president. Nevertheless, some indications converge to show that agency is an ideological dimension in that some voters are more inclined than others to look for agency in politicians. Findings from the social dominance orientation approach (SDO, Sidanius & Pratto, 2001) showed that highSDO individuals are primarily motivated to climb social hierarchies and at ease with competitive contexts. Interestingly, items of the SDO scale (e.g., “to get ahead in life, it is sometimes necessary to step on other groups”) have clearly agentic content. There is some evidence that the willingness to valorize the self within highly competitive contexts is congruent with political conservatism. The link between a high-SDO orientation and political conservatism is now firmly established (e.g., Sidanius, Pratto, & Bobo, 1996). In another vein of research, authors have shown that people who scored high in the dominant subscales of the extraversion dimension were the most likely to hold a right-wing ideology (Caprara, Barbaranelli, & Zimbardo, 1999; Van Hiel & Mervielde, 2002). Because voters may prefer politicians with a personality profile similar to their own (Caprara et al., 2007), we hypothesize that right-wing voters are much more interested in politicians’ agency than left-wing voters are.

Morality As stated above, morality is one of the most important criteria by which politicians are judged. The taxonomy of values proposed by Schwartz (1992) is very useful for locating morality with respect to agency. Schwartz identified ten values that sum up the requirements of human existence; in his taxonomy of values, self-transcendence values (e.g., tolerance and protection for the welfare of all people) are opposed to self-enhancement values (success, achievement, and ambition). Caprara et al.

(2009) showed that personal values were good predictors of political preferences. As expected, self-transcendence values were found to be linked to left-wing ideology and self-enhancement values to right-wing ideology. Moreover, these authors confirmed that self-transcendence was regarded as most important by left-wing voters (Caprara et al., 1999; Caprara, Schwartz, Vecchione, & Barbaranelli, 2008; Caprara et al., 2007). Because agentic traits refer to self-enhancement values and morality to self-transcendence values, we hypothesize that left-wing voters will be more interested in the morality of presidential candidates than right-wing voters will. © 2016 Hogrefe

37

Sociability Compared to morality, sociability appears to play a negligible role in the making of political decisions. Nevertheless, the above considerations, which contrasted morality and agency, still apply to sociability, which is a more transcendent value than agency. Roets and Van Hiel’s (2009) study directly supported this view: Left-wing voters identified the ideal politician to be higher in agreeableness than right-wing voters did. Thus, we believe that left-wing voters will be more interested in the sociability of presidential candidates than right-wing voters.

Objectives and Hypothesis The first objective of the present study was to examine how judgments of competence, agency, morality, and sociability are qualities regarded as required for exercising the function of president and how sensitive these judgments are to voters’ political orientations. By focusing on the two basic dimensions of social judgment, we assumed that social utility would be seen as a more important required quality than social desirability.

More precisely, we expected competence to be a fundamental criterion for both left- and right-wing voters, agency to be more important for right-wing voters, and morality and sociability to be more important for left-wing voters.The second objective was directly related to the last presidential election in France.

We studied the effect of political orientation on voters’ differentiation between the five main candidates of the French presidential election (the five candidates above 10% of voting intentions according to opinion polls). From a psychosocial point of view, differentiating between individuals with respect to a dimension is an indicator of the intensity of social judgments (see Mignon & Mollaret, 2002, 2012). The more a series of individuals are rated as being different from each other, the more effective a social judgment is. Conversely, a very tenuous differentiation indicates that participants are reluctant to judge individuals. Thus, the strength of voters’ differentiation of candidates provides an indication of the relevance of that dimension for judgment of politicians. We explored the extent to which the five main candidates were differentiated with respect to each subdimension as a function of voters’ political orientation.

The third objective was to analyze the extent to which voters’ rating were partisan. We thus decided to take into account how candidates’ ideological orientations interact with voters’.

If the participants are motivated to make partisan ratings, they should judge candidates with an ideological orientation similar to their own more positively and deny that other candidates have positive traits. We explored whether this potential partisanship bias exists for each dimension.

Swiss Journal of Psychology (2016), 75 (1), 35–45

38

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

Method Participants The participants were recruited online through LimeSurvey (2012); 674 participants responded to all of the questions. The participants were to indicate their political orientation by moving a cursor on a 200-point continuum, going from left (–100) to right (+100). A total of 449 participants indicated that they were left-wing (–100 to –1), 201 said they were right-wing (+1 to +100), and 24 lay in-between (0). To create a categorical independent variable political orientation with two balanced groups to be used in ANOVA analyses, 201 of the 449 left-wing participants, selected randomly, were retained. Thus, 402 participants, of which 172 were women (Mage = 41.71 years, SDage = 17.15, age range = 18–80 years) and 230 were men (Mage = 43.14 years, SDage = 17.15, age range = 18–84 years), were included in the analysis.

Procedure and Selection of the Candidates Since 1965, the President of the French Republic has been elected by direct universal suffrage. The President is elected by absolute majority vote in a two-round procedure. In 2000, the length of the term was reduced from 7 to 5 years, and since 2008 a president may not serve more than two consecutive terms. The presidential election of 2012 was held on April 22, 2012, with a run-off held on May 6, 2012. In order to qualify for the first round, a candidate had to collect the signatures of at least five hundred elected representatives (mayors, general councilors, regional councilors, deputies, senators, members of the European Parliament elected in France). On March 19, 2012, the Constitutional Council announced that ten candidates had qualified to run for office. The official campaign began on March 20, 2012.

Data were collected between April 4–22, 2012, via website.

The study was presented as a scientific study (and not as a sample survey) about voters’ perception of the five main candidates. Five candidates were selected from the ten because they represented the five main tendencies of the political French context and each assembled above 10% of voter intention in a Gallup poll. From the extreme left-wing to the extreme rightwing, these candidates were: Jean-Luc Mélenchon (Left Front, radical left-wing), Francois Hollande (Socialist Party, left-wing), François Bayrou (Democratic Movement, center right-wing), Nicolas Sarkozy who was President from 2007 to 2012 (Union for a Popular Movement, right-wing), Marine Le Pen (National Front, extreme right-wing).

Participants were asked to report their political orientation; the importance they placed on agency, competence, warmth, and morality as qualities required to exercise the function of president; and their ratings of the five candidates on those subdimensions. They also provided demographic information, including their age and sex1.

Measures Political Orientation The participants reported their political orientation by responding to the question: “How would you characterize your political orientation?” by moving a cursor on a line. It was said that the extreme left part of the line corresponded to the most left-wing orientation and the extreme right part of the line corresponded to the most right-wing orientation. The location of the cursor was directly encoded by a number ranging from –100 (the most left-wing location) to +100 (the most right-wing location). This measure was used to facilitate the communication of participants’ political orientation because in France people are not inclined to reveal it, in particular when they take an extreme right-wing position. On average, participants were slightly leftwing (M = –3.93, SD = 55.02). To create the categorical independent variable political orientation, participants whose political orientation ranged from –100 to –1 were left (M = –51.54, SD = 27.53) and those whose political orientation ranged from +1 to +100 were right (M = 43.68, SD = 27.50).

Judgment of Qualities Required to Exercise the Function of President Each dimension was operationalized by two positive traits: leader and self-assured for agency, competent and intelligent for competence, sincere and honest for morality, friendly and sociable for sociability. These traits had been selected from previous research on social judgment (e.g., Abele & Wojicizke, 2007; Brambilla et al., 2011; Dubois, 2010) and seemed relevant to the political domain. Using a scale ranging from 1 (not required at all) to 7 (completely necessary), participants rated the extent to which they thought each trait was required to exercise the function of president.

Perception of the Five Candidates and Differentiation Among Candidates Participants rated each candidate on eight bipolar scales composed of the positive traits used for the evaluation of the qual-

1 Note that participants were asked to answer to other questions (e.g., political orientation estimation of each candidate) in relation to a part of this research that is not considered in this paper.

Swiss Journal of Psychology (2016), 75 (1), 35–45

© 2016 Hogrefe

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

ities and their respective opposite (for agency: follower/leader and hesitating/self-assured; for competence: incompetent/competent and unintelligent/intelligent; for morality: hypocrite/sincere and honest/dishonest; for sociability: unpleasant/friendly and cold/sociable). Participants rated the extent to which each item applied to each candidate on a 7-point scale ranging from –3 (very Trait A) to 0 (in between the two traits) to +3 (very Trait B). For each candidate, we calculated the mean score based on the two items of each dimension. Those scores were used to measure the partisan ratings. To assess differentiation among the five candidates, we computed the standard deviation between the scores that a participant assigned to each candidate on a dimension.

(The standard deviation is calculated as following for each _ (x−x)2 participant: SD = Σ were xi is the score assigned by a n−1 _ participant to a candidate i on a subdimension, and x is the mean of the xi and n is the number of candidates.)

A high score indicated that the five candidates were judged very differently on a dimension, whereas a low score indicated that the candidates were judged similarly (see Mignon & Mollaret, 2002, 2012).

⎯⎯⎯⎯ √

Results Note that we considered all effects associated with a p > .05 nonsignificant because the associated partial eta squared ηp2 was systematically < .01, which corresponds to a very weak effect size. All decompositions of significant interactions included Bonferroni correction. Participant sex was introduced first in the ANOVAs; no significant interaction was observed with this variable, so we didn’t take this variable into account in the ANOVA presented in this results section.

Judgment of Qualities Required to Exercise the Function of President Principal factors extraction with varimax rotation2 was performed on the eight items used to assess the qualities required to exercise the function of president. Principal components extraction was used prior to principal factors extraction to verify how items correctly loaded on two different components, as provided in the theoretical model. The two components extracted were consistent with the literature and explained a total of 52.63% of the variance: sincere (λ = .86), honest (λ = .83), sociable (λ = .54), and friendly (λ = .42) loaded on the first component (i.e., social desirability); leader (λ = .76), self-assured (λ = .76), intelligent (λ = .63), and

39

competent (λ = .58) loaded on the second factor (i.e., social utility). Orthogonal rotation was retained in this first step because of conceptual model and ease of description. Moderate correlations (r = –.24) were revealed when oblique rotation was requested. However, because these correlations were under .30, orthogonal rotation was deemed adequate in this preliminary step (Tabachnick & Fidell, 2007). In a second step, we subjected these eight items to a principal component extraction (PCA) with oblique rotation, in order to empirically investigate whether the subcomponents of the two principal components would appear. This analysis revealed four factors: intelligent (λ = .94) and competent (λ = .76) loaded on the first factor (i.e., competence); sincere (λ = –.89) and honest (λ = –.89) on the second factor (i.e., morality); friendly (λ = .84) and sociable (λ = .84) on the third factor (i.e., sociability); leader (λ = .93) and self-assured (λ = .69) on the fourth (i.e., agency). The expected correlations appeared between the factors competence and agency (r = .36) and between the factors morality and sociability (r = .28). Following these preliminary steps, we were able to create average scores and submit these data to a 2 × 2 (Political Orientation [right-wing, left-wing] × Dimension [Social Desirability, Social Utility]) factorial ANOVA with dimension as a within-participants factor. The dependent variable was the mean score within each component, that is, the averaged scores of sociability and morality for social desirability (SD) and the averaged scores of agency and competence for social utility (SU).

The main effect of political orientation was not significant, F(1, 400) = 2.25, p = .14, ηp2 = .006. As expected, the main effect of component was significant, indicating that social utility (M = 4.61, SD = 0.45) was perceived as being a quality that is more important for exercising the function of president than social desirability (M = 4.01, SD = 0.63; F(1, 400) = 283.14, p < .001, ηp2 = .41). The Political Orientation × Dimension interaction was also significant, F(1, 400) = 36.31, p < .001, ηp2 = .08.

The pairwise comparisons confirmed the main effect of component since SU was perceived, both for left- and right-wing voters, as being a quality that is more important for exercising the function of president than SD (for left-wing voters, MSU = 4.55, SDSU = 0.43, MSD = 4.21, SDSD = 0.54; F(1, 400) = 58,34, p < .001, ηp2 = .13; for right-wing voters, MSU = 4.67, SDSU = 0.47, MSD = 3.96, SDSD = 0.69; F(1, 400) = 261.11, p < .001, ηp2 = .39). As expected, the effect of political orientation on SD and SU was different. More precisely, left-wing voters placed more importance on SD than right-wing voters did, F(1, 400) = 16.70, p < .001, ηp2 = .04, whereas left-wing voters placed less importance on SU than right-wing voters did, F(1, 400) = 7.04, p = .008, ηp2 = .02. The means and standard deviations are presented in Figure 1.

To further explore these results, we conducted two sepa-

2 The eigenvalues are ≥ 1. © 2016 Hogrefe

Swiss Journal of Psychology (2016), 75 (1), 35–45

40

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

Figure 1. Mean judgment score within social desirability and social utility as a function of political orientation. Standard errors are represented by the error bars.

desirability: Fleft-wing(1,400) = 140.02, p < .001, ηp2 = .26; Fright-wing(1, 400) = 126.37, p < .001, ηp2 = .24). The means and standard deviations are presented in Figure 2.

In the same way, we conducted a 2 × 2 (Political Orientation [right-wing, left-wing] × Social Utility [agency, competence]) factorial ANOVA with social utility as a within-participants variable on the judgment of qualities required to exercise the function of president. As expected, participants rated competence (M = 4.76, SD = 0.48) as being a quality that is more required than agency (M = 4.45, SD = 0.59; F(1, 400) = 119.21, p < .001, ηp2 = .23). The expected Political Orientation × Social Utility interaction was significant, F(1, 400) = 25.6, p < .001, ηp2 = .06. Decomposition of the interaction confirmed the main effect of social utility since left- and right-wing voters rated competence (MCompetence = 4.77, SDCompetence = 0.41; MCompetence = 4.75, SDCompetence = 0.54, respectively) as a quality that is more important than agency (MAgency = 4.32, SDAgency = 0.6; MAgency = 4.59, SDAgency = 0.56, respectively; effect of social utility: Fleft2 wing(1, 400) = 127.64, p < .001, ηp = .24; Fright-wing (1, 400) = 2 17.16, p < .001, ηp = .04). Compared to left-wing voters, rightwing voters considered agency to be a quality that is more required (F(1, 400) = 20.38, p < .001, ηp2 = .05) than left-wing voters did, but right- and left-wing voters did not differ with respect to the importance placed on competence (F(1, 400) = .22, p = .64, ηp2 = .001). The means and standard deviations are presented in Figure 3.

Figure 2. Mean judgment score for morality and sociability as a function of political orientation. Standard errors are represented by the error bars.

rate ANOVAs analyzing the effect of political orientation within factor. The dependent variable was the mean score of the two traits illustrating a subdimension (i.e., factor). First, we conducted a 2 × 2 (Political Orientation [right-wing, leftwing] × Social Desirability [morality, sociability]) factorial ANOVA with social desirability as a within-participants factor on the judgment of qualities required to exercise the function of president. Morality (M = 4.45, SD = 0.75) was rated as being a more important quality than sociability, M = 3.72, SD = 0.79), F(1, 400) = 226.21, p < .001, ηp2 = .39. The Political Orientation × Social Desirability interaction was not significant F(1, 400) = .18, p = .64, ηp2 = .001 (see Figure 2).

The pairwise comparisons revealed that the two main effects were significant, that is, left-wing voters (MMorality = 4.59, SDMorality = 0.59; MSociability = 3.84, SDSociability = 0.75) placed more importance on morality and sociability than right-wing voters did (MMorality = 4.32, SDMorality = 0.86, MSociability = 3.61, SDSociability = 0.82; effect of political orientation: FMor2 ality(1, 400) = 13.53, p < .001, ηp = .033); FSociability (1, 400) = 8.85, p < .001, ηp2 = .022) and that both voters placed more importance on morality than sociability (effect of social Swiss Journal of Psychology (2016), 75 (1), 35–45

Figure 3. Mean judgment score for agency and competence as a function of political orientation. Standard errors are represented by the error bars.

In sum, our hypotheses were confirmed. The results showed that political orientation determined the importance placed on qualities required for exercising the function of President of the French Republic. Neither kind of voter differed with respect to the fact that (1) social utility is a quality that is more important than social desirability, (2) morality is more important than warmth, and (3) competence is more important than agency.

However, compared to left-wing voters, right-wing voters placed (1) more importance on agency and (2) less importance on morality and warmth. © 2016 Hogrefe

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

41

Differentiation Among Candidates We conducted a 2 × 2 (Political Orientation [right-wing, leftwing] × Components [social desirability, social utility]) factorial ANOVA with dimension as a within-participants factor on differentiation among candidates. The dependent variable was average score between warmth and morality for SD and average score between agency and competence for US. The political orientation effect was not significant, F(1, 400) = 1,18, p = .28, ηp2 = .003. The significant main effect of component showed that candidates were more differentiated on SD (M = 1.75, SD = 0.62) than on SU (M = 1.36, SD = 0.56), F(1, 400) = 196.25, p < .001, ηp2 = .33. The Political Orientation × Component interaction was significant, F(1, 400) = 128.90, p < .001, ηp2 = .24. Pairwise comparisons revealed that, for left-wing voters, differentiation is higher on SD (M = 1.88, SD = 0.55) than on SU (M = 1.17, SD = 0.51; F(1, 400) = 321.62, p < .001, ηp2 = .45), but this effect was not significant for right-wing voters (MSD = 1.62, SDSD = 0.66; MSU = 1.54, SDSU = 0.56; F(1, 400) = 3.53, p = .06, ηp2 = .009). Moreover, the effect of political orientation was significant for both dimensions. More precisely, as shown in Figure 4, left-wing voters differentiated candidates with respect to SD more than right-wing voters did, F(1, 400) = 18.21, p < .001, ηp2 = .04, whereas left-wing voters differentiated candidates with respect to SU less than rightwing voters did, F(1, 400) = 47.76, p < .001, ηp2 = .11.

In order to analyze the effect of political orientation within each component, we conducted two separate ANOVAs. The 2 × 2 (Political Orientation [right-wing, left-wing] × Social Desirability [morality, sociability]) ANOVA with social desirability as a within-participants factor revealed that neither the effect of social desirability (MMorality = 1.77, SDMorality = .70; MSociability = 1.72, SDSociability = .70; F(1, 400) = 2.95, p = .087, ηp2 = .007) nor the interaction were significant, F(1, 400) = .11, p = .71). The pairwise comparisons confirmed the two main effects previously described. The simple effects of social desirability were not significant in the left-wing condition (MMorality = 1.91, SDMorality = 0.61): MSociability = 1.84, SDSociability = 0.64; Fleft-wing(1, 400) = 2.10, p = .15) or in the right-wing condition (MMorality = 1.64, SDMorality

Figure 4. Differentiation of candidates with respect to social desirability and social utility as a function of political orientation. Standard errors are represented by the error bars.

© 2016 Hogrefe

Figure 5. Differentiation of candidates with respect to morality and sociability as a function of political orientation. Standard errors are represented by the error bars.

= 0.75 MSociability = 1.59, SDSociability = 0.73; Fright-wing (1, 400) = .95, p = .33, ηp2 = .002). The simple effects of the political orientation were significant in the morality and sociability conditions. More precisely, left-wing voters differentiated candidates with respect to morality and sociability more than rightwing voters did (FMorality(1, 400) = 15.68, p < .001, ηp2 = .04; FSociability(1, 400) = 13.01, p < .001, ηp2 = .03, respectively). The means and standard deviations are presented in Figure 5.

The 2 × 2 (Political Orientation [right-wing, left-wing] × Social Utility [agency, competence]) factorial ANOVA with social utility as a within-participants factor revealed that the effect of social utility was not significant, F(1, 400) = 1.61, p = .21, ηp2 = .004. The Political Orientation × Social Utility interaction was significant, F(1, 400) = 20.72, p < .001, ηp2 = .05. Right-wing voters differentiated candidates with respect to agency (M = 1.61, SD = .69) more than competence (M = 1.48, SD = .65) whereas the opposite pattern was observed for left-wing voters (MAgency = 1.06, SDAgency = .61; MCompetence = 1.29, SDCompetence = .69; F(1, 400) = 5.39, p = .021, ηp2 = .01 and F(1, 400) = 16.74, p < .001, ηp2 = .04, respectively). Moreover, right-wing voters differentiated candidates with respect to competence more than left-wing voters did; the same was true for agency (F(1, 400) = 8.09, p = .005, ηp2 = .02; F(1, 400) = 70.57, p <

Figure 6. Differentiation scores for agency and competence as a function of political orientation. Standard errors are represented by the error bars.

Swiss Journal of Psychology (2016), 75 (1), 35–45

42

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

.001, ηp2 = .15, respectively). The means and standard deviations are presented in Figure 6.

In sum, both kinds of voters differentiated the candidates with respect to social desirability more than social utility. However, left-wing voters differentiated candidates with respect to social desirability more than right-wing voters did while rightwing voters differentiated with respect to social utility more than left-wing voters did. For both kinds of voters, there were more differences between candidates with respect to sociability than morality. Right-wing voters differentiated on the basis of social utility more than left-wing voters did. However, rightwing voters differentiated on the basis of agency more than competence, whereas left-wing voters differentiated on the basis of competence more than agency.

Figure 7. Mean competence for the five candidates as a function of political orientation. Standard errors are represented by the error bars.

Partisan Ratings To analyze partisan ratings, we conducted four separate ANOVAs differentiated by the outcome measure: mean on agency, mean on competence, mean on morality, and mean on sociability. Each ANOVA followed a 2 × 5 (Political Orientation [right-wing, left-wing] × Candidates [Mélenchon, Hollande, Bayrou, Sarkozy, Le Pen]) design with candidates as a withinparticipants factor. The interaction was adapted to examine whether voters had given partisan ratings3. The Political Orientation × Candidates interaction was significant on the four subdimensions, FAgency(4, 1600) = 45.32, p < .001, ηp2 = .10; FCom2 petence(4, 1600) = 125.87, p < .001, ηp = .24; FMorality(4, 1600) = 2 166.16, p < .001, ηp = .29; and FSociability(4, 1600) = 153.61, p < .001, ηp2 = .28 . It appeared that for competence, morality and sociability, the judgments were partisan: As illustrated in Figure 7, Figure 8, and Figure 9, left-wing candidates (Mélenchon and Hollande) were rated as being more competent, moral, and sociable by left-wing voters than by right-wing voters: for Mélenchon, FCompetence(1, 400) = 173.34, p < .001, ηp2 = .30; FMorality(1, 400) = 161.87, p < .001, ηp2 = .29; FSociability(1, 400) = 143.17, p < .001, ηp2 = .27; for Hollande, FCompetence(1, 400) = 186.75, p < .001, ηp2 = .32; FMorality(1, 400) = 187.19, p < .001, ηp2 = .32; FSociability(1, 400) = 97.64, p < .001, ηp2 = .20 . The reverse was observed for the right-wing candidates (Sarkozy and Le Pen): They were rated as being more competent, moral, and sociable by right-wing voters than by left-wing voters: for Sarkozy, FCompetence(1, 400) = 71.85, p < .001, ηp2 = .15; FMoral2 ity(1, 400) = 150.31, p < .001, ηp = .27; FSociability(1, 400) = 2 181.90, p < .001, ηp = .31; for Le Pen, FCompetence(1, 400) =

Figure 8. Mean morality for the five candidates as a function of political orientation. Standard errors are represented by the error bars.

Figure 9. Mean sociability for the five candidates as a function of political orientation. Standard errors are represented by the error bars.

3 The two main effects are significant, but they are not central to our purpose. In fact, the effect of political orientation only indicates whether rightand left-wing voters judged all of the candidates more or less positively on a subdimension and the effect of candidates only indicates whether candidates were judged differently from each other on a subdimension, and this information is better captured by the dependent variable differentiation among candidates. However, we bring to the attention of the reader the results of the two main effects. The results showed significant effects of political orientation on agency and competence, FAgency(1, 400) = 79.07, p < .001, η2 = .17 and FCompetence(1, 400) = 13.29, p < .001, η2 = .03, but not on warmth or morality, FSociability(1, 400) = 79.07, p < .001, η2 = .17 and FMorality(1, 400) = 13.29, p < .001, η2 = .03. The effect of candidates was significant in each case, FAgency(4, 1600) = 282.48, p < .001, η2 = .41; FCompetence(4, 1600) = 31.25, p < .001, η2 = .07; FMorality(4, 1600) = 78.77, p < .001, η2 = .165; and FSociability(4, 1600) = 128.30, p < .001, η2 = .24.

Swiss Journal of Psychology (2016), 75 (1), 35–45

© 2016 Hogrefe

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

Figure 10. Mean agency for the five candidates as a function of political orientation. Standard errors are represented by the error bars.

30.13, p < .001, ηp2 = .07; FMorality(1, 400) = 90.15, p < .001, ηp2 = .18; FSociability(1, 400) = 107.77, p < .001, ηp2 = .21.

Left- and right-wing voters did not differ in their rating of the central candidate (Bayrou) with respect to competence, morality, or sociability (all Fs < 1, FCompetence(1, 400) = .37, p = .54, ηp2 = .0001; FMorality(1, 400) = .079, p = .78, ηp2 = .001; FSociabil2 ity(1, 400) = .021, p < .88, ηp = .0001).

For ratings of agency, another pattern of results obtained (Figure 10): The partisan rating was only seen for the two leftwing candidates (Hollande, FAgency(1, 400) = 183.83, p < .001, ηp2 = .32 and Mélenchon, FAgency(1, 400) = 46.09, p < .001, ηp2 = .10), the right-wing candidates (Sarkozy and Le Pen) were seen as being strongly agentic by both kinds of voters: FAgency(1, 400) = .21, p = .65, ηp2 = .001, for Le Pen, FAgency(1, 400) = 2,79, p = .10, ηp2 = .007. Left-wing voters rated the central candidate (Bayrou) as being more agentic than rightwing voters did, FAgency(1, 400) = 15.75, p < .001, ηp2 = .04.

Discussion In line with the social value approach, the aim of the present research was to examine how judgments of social utility (i.e., competence and agency) and social desirability (i.e., morality and sociability) are qualities required to exercise the function of president and how these requirements might be sensitive to voters’ political orientations. This study was conducted in the specific context of the French presidential election of 2012. Our theoretical objective was to show that voters’ ideology (operationalized by political orientation) would determine their judgments of candidates’ personality. First, we showed that left- and right-wing voters hold different views about the personality traits needed to be president. Compared to right-wing voters, left-wing voters placed more importance on socially desirable traits and less importance on socially useful traits. Moreover, taking the subdimensions into account, even if both categories of voters placed more importance on competence than agency, right-wing voters considered agency to be a more important quality in a president than left-wing voters did. Second, our results showed that political orientation influenced the participants’ judgments of real candi© 2016 Hogrefe

43

dates’ personality. Right-wing voters differentiated candidates with respect to social desirability less and with respect to social utility more than left-wing voters did. Moreover, unlike right-wing voters, left-wing voters used social desirability more than social utility to differentiate candidates. Importantly, left-wing voters differentiated candidates with respect to competence more than agency and the reverse pattern was found for right-wing voters.

Third, voters attributed more morality, sociability, and competence to candidates with an ideological orientation similar to their own. By contrast, the results revealed no partisan ratings with respect to agency.

Most of our results confirm our hypotheses about the link between descriptions of politicians and political orientations of voters. However, the voters’ perception of candidates with respect to agency was surprising. Both left- and right-wing voters characterized Mélenchon (radical left wing), Sarkozy (right wing), and Le Pen (radical right wing) as agentic candidates. Only François Hollande gave rise to clearly partisan ratings: Only left-wing voters found him to be moderately agentic. Overall, voters were not reluctant to attribute agency to candidates with different political views than their own. However, the results concerning competence followed a very different pattern. Right-wing voters denied that Mélenchon was competent whereas left-wing voters found that Sarkozy and Le Pen were not competent. One plausible explanation for this noncorrespondence between agency and competence ratings (the two subcomponents of social utility) is that all voters may consider competence to be the most necessary dimension to become president. If competence is the primary criterion for voters’ political choices, voters would be expected to ascribe competence to their favorite candidate. Our methodology does not allow us to make any clear conclusions about which dimensions are primary criteria for voting decisions. Further research is necessary to explore how voters attribute competence and agency to politicians. Moreover, our results cannot be generalized to the whole political domain. The prevalent role of social utility (and particularly agency) in the voters’ choice could be limited to the highest political functions (the presidential election is the most important election in France). One could hypothesize that morality and warmth (social desirability) are more valued for more local responsibilities (e.g., mayor of a small town). Behaving as a leader (agency) is advantageous for reaching the top of social hierarchies, which is not the goal of all politicians.

In sum, our results confirm the relevance of distinguishing between different subdimensions of traits within social utility and social desirability. However, because our research was conducted with only two traits per subdimension, further work is necessary before drawing firm conclusions. More precisely, our experiment was conducted with only two traits in order to prevent the excessive fatigue of the participants which were to make a series of judgments. The distinction between different subdimensions within the Big Two should be confirmed by means of simpler studies in which participants are instructed to rate many traits of a few well-known heads of state. Although the political orientation is a recurrent variable involved in the Swiss Journal of Psychology (2016), 75 (1), 35–45

44

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

prediction of political choices, it is probably not sufficient to understand voters’ evaluations of politicians. In future work, we will seek to verify whether voters’ support of more explicit ideological beliefs (i.e., justification of the actual political system) could be a strong determinant of their perception of politicians. In particular, the belief that a nonegalitarian political system is fair and should be maintained (Kay & Jost, 2003) could orient voters to look for the most agentic political leaders.

In this vein, it is also important to study voters’ perception of politics with respect to their social status. High-status voters’ motivation to maintain their social position relative to others may also play a role in the perception of politicians.

Our results are compatible with Caprara, Barbaranelli, and Zimbardo (2002) argument that voters’ perception of politicians’ personality is only based on two factors. Although the Big-Two factors (communion/morality and competence/agency) can be seen as two meta-traits corresponding to genuine psychological tendencies (Vecchione, Alessandri, Barbaranelli, & Caprara, 2011; Wiggins & Trapnell, 1996), our results are more compatible with an evaluative view of the dimensions expressing a variety of beliefs or values. Voters may use personality traits to express their ideological beliefs. In that sense, the fact that people are interested in politicians’ personality is more than a simple personalization of political issues. We propose that voters may use traits as ideological criteria on the basis of which they judge candidates in order to make their voting decisions. Thus, the evaluative view of personality description held by social psychologists (e.g., Dubois & Beauvois, 2011) has the potential to conceptualize the link between personality judgments and ideological beliefs. Given that the Big-Two approach has proved its efficiency for studying a variety of social and societal issues like the rationalization of gender stereotypes (Fiske et al., 2002) and the justification of inequality of resources (Oldmeadow & Fiske, 2007, 2010), we think it is better adapted to study political issues than the traditional fivefactor model.

Finally, our findings confirm the legitimacy of dissociating different facets of the Big-Two. Only a few studies view agency and competence as distinct facets of the second fundamental dimension (for two exceptions, see Carrier et al., 2014; Dubois, 2010).

This study further demonstrates that competence and agency correspond to two very different facets of social utility. Competence (i.e., the ability to master a task) is valued by people of any ideological orientation, which is most certainly not the case for agency (i.e., the willingness to perform better than others). Beyond political psychology, our results provide a strong argument for the dissociation of agency and competence.

Acknowledgments This research was supported by a grant from the Maison Européenne des Sciences de l’Homme et de la Société (USR 3185, Lille, North of France). We thank Tanguy Leroy for his comment on a statistical issue.

Swiss Journal of Psychology (2016), 75 (1), 35–45

References Abele, A. E., Cuddy, A. J. C., Judd, C. M., & Yzerbyt, V. Y. (2008). Fundamental dimensions of social judgment. European Journal of Social Psychology, 38, 1063–1065. doi 10.1002/ejsp.574 Abele, A. E., & Wojciszke, B. (2007). Agency and communion from the perspective of self versus others. Journal of Personality and Social Psychology, 93, 751–763. doi 10.1037/0022-3514.93.5.751 Beauvois, J.-L., & Dubois, N. (2009). Lay psychology and the social value of persons. Social and Personality Psychology Compass, 6, 1082–1095. doi 10.1111/j.1751-9004.2009.00225.x Bennett, L. (2012). The personalization of politics: Political identity, social media, and changing patterns of participation. The Annals of the American Academy of Political and Social Science, 644(1), 20–39. doi 10.1177/0002716212451428 Brambilla, M., Rusconi, P., Sacchi, S., & Cherubini, P. (2011). Looking for honesty: The primary role of morality (vs. sociability and competence) in information gathering. European Journal of Social Psychology, 41, 135–143. doi 10.1002/ejsp.744 Caprara, G. V., Barbaranelli, C., & Zimbardo, P. G. (1999). Personality profiles and political parties. Political Psychology, 20, 175– 197. doi 10.1111/0162-895X.00141 Caprara, G. V., Barbaranelli, C., & Zimbardo, P. G. (2002). When parsimony subdues distinctiveness: Simplified public perception of politicians’ personality. Political Psychology, 23, 77–96. doi 10.1111/0162-895X.00271 Caprara, G. V., Schwartz, S. H., Vecchione, M., & Barbaranelli, C. (2008). The personalization of politics: Lessons from the Italian case. European Psychologist, 3, 157–172. doi 10.1027/10169040.13.3.157 Caprara, G. V., Vecchione, M., Barbaranelli, C., & Fraley, R. C. (2007).

When likeness goes with liking: The case of political preference.

Political Psychology, 28, 609–632. doi 10.1111/j.1467-9921.

2007.00592.x Caprara, G. V., Vecchione, M., & Schwartz, S. H. (2009). The mediational role of values in linking personality traits to political orientation. Asian Journal of Social Psychology, 12, 82–94. doi 10.1111/j.1467-839X.2009.01274.x Carrier, A., Louvet, E., Chauvin, B., & Rohmer, O. (2014). The primacy of agency over competence in status perception. Social Psychology, 45, 347–356. doi 10.1027/1864-9335/a000176 Cislak, A., & Wojciszke, B. (2006). The role of self-interest and competence in attitudes toward politicians. Polish Psychological Bulletin, 37, 203–212. doi 10.1002/ejsp.554 Dubois, N. (2010). Theory of the social value of persons applied to organizations: Typologies of good leaders and recruitment. European Review of Applied Psychology, 60, 255–266. doi 10.1016/ j.erap.2010.01.002 Dubois, N., & Beauvois, J.-L. (2005). Normativeness and individualism. European Journal of Social Psychology, 35, 123–146. doi 10.1002/ejsp.236 Dubois, N., & Beauvois, J.-L. (2011). Are some rabbits more competent and warm than others? The lay epistemologist is interested in object value, not in descriptive parameters. Swiss Journal of Psychology, 70, 63–73. doi 10.1024/1421-0185/a000040 Dubois, N., & Beauvois, J.-L. (2012). The social value of persons: Theory and applications. In G. Rossi (Ed.), Psychology – selected papers (pp. 307–330). Rijeka, Croatia: InTech Europe.

Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82, 878–902. doi 10.1037/ 0022-3514.82.6.878 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.

© 2016 Hogrefe

A. Mignon et al.: Judgment of Agency, Competence, Morality, and Sociability

Journal of Personality and Social Psychology, 85, 823–837. doi 10.1037/0022-3514.85.5.823 Kinder, D. R., & Sears, D. O. (1985). Public opinion and political action. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed., Vol. 2, pp. 659–741). New York: Random House.

Leach, C. W., Ellemers, N., & Barreto, M. (2007). Group virtue: The importance of morality (vs. competence and sociability) in the positive evaluation of in-groups. Journal of Personality and Social Psychology, 93, 234–249. doi 10.1037/0022-3514.93.2.234 LimeSurvey Project Team/Carsten Schmitz. (2012). LimeSurvey: An Open Source survey tool [Internet Software]. Hamburg, Germany: LimeSurvey Project. Retrieved from http://www.limesurvey.org.

McCrae, R. R., & Costa, P. T. (2006). Perspectives de la théorie des cinq facteurs (TCF): traits et culture [A five-factor theory perspective on traits and culture]. Psychologie Française, 51, 227– 244. doi 10.1016/j.psfr.2005.09.001 Mignon, A., & Mollaret, P. (2002). Applying the affordance conception of traits: A person perception study. Personality and Social Psychology Bulletin, 10, 1327–1334. doi 10.1177/014616702236825 Mignon, A., & Mollaret, P. (2012). Perceiving the target’s state or state provoked by the target? An analysis of the descriptive and evaluative knowledge in person perception. British Journal of Social Psychology, 51, 583–605. doi 10.1111/j.2044-8309.2011.02025.x Oldmeadow, J. A, & Fiske, S. T. (2007). System justifying ideologies moderate status = competence stereotypes: Roles for belief in a just world and social dominance orientation. European Journal of Social Psychology, 37, 1135–1148. doi 10.1002/ejsp.428 Oldmeadow, J. A., & Fiske, S. T. (2010). Social status and the pursuit of positive social identity: Systematic domains of intergroup differentiation and discrimination for high- and low-status groups.

Group Processes and Intergroup Relations, 13, 425–444. doi 10.1177/1368430209355650 Olivola, C. Y., & Todorov, A. (2010). Fooled by first impressions? Reexamining the diagnostic value of appearance-based inferences. Journal of Experimental Social Psychology, 46, 315–324. doi 10.1016/j.jesp.2009.12.002 Page, B. I. (1978). Choices and echoes in presidential elections: Rational man and electoral democracy. Chicago, IL: University of Chicago Press.

Peeters, G. (1992). Evaluative meanings of adjectives in vitro and in context: Some theoretical implications and practical consequences of positive-negative asymmetry and behavioral-adaptive concepts of evaluation. Psychologica Belgica, 32, 211–231.

Peeters, G., & Czapinski, J. (1990). Positive-negative asymmetry in evaluations: The distinction between affective and informational negativity effects. European Review of Social Psychology, 1, 33–60. doi 10.1080/14792779108401856 Poutvaara, P., Jordahl, H., & Berggren, N. (2009). Faces of politicians: Babyfacedness predicts inferred competence but not electoral success. Journal of Experimental Social Psychology, 45, 1132–1135.doi 0.1016/j.jesp.2009.06.007 Roets, A., & Van Hiel, A. (2009). The ideal politician: Impact of voters’

© 2016 Hogrefe

45

ideology. Personality and Individual Differences, 46, 60–65. doi 10.1016/j.paid.2008.09.006 Schwartz, S. H. (1992). Universals in the content and structure of values: Theory and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 1–65). New York: Academic Press.

Schwartz, S. H., Caprara, G. V., & Vecchione, M. (2010). Basic personal values, core political values, and voting: A longitudinal analysis. Political Psychology, 31, 421–452. doi 10.1111/j.14679221.2010.00764.x Sidanius, J., & Pratto, F. (2001). Social dominance: An intergroup theory of social hierarchy and oppression. New York: Cambridge University Press.

Sidanius, J., Pratto, F., & Bobo, L. (1996). Racism, conservatism, affirmative action and intellectual sophistication: A matter of principled conservatism or group dominance? Journal of Personality and Social Psychology, 70, 476–490. doi 10.1037/00223514.70.3.476 Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.). Boston, MA: Pearson International Edition.

Van Hiel, A., & Mervielde, I. (2002). Social identification among political party voters and members: An empirical test of optimal distinctiveness theory. Journal of Social Psychology, 142, 202–209. doi 10.1080/00224540209603895 Vecchione, M., Alessandri, G., Barbaranelli, C., & Caprara, G. V. (2011). Higher-order factors of the Big Five and basic values: Empirical and theoretical relations. British Journal of Psychology, 102, 478–498. doi 10.1111/j.2044-8295.2010.02006.x Wentura, D., Rothermund, K., & Bak, P. (2000). Automatic vigilance: The attention-grabbing power of approach- and avoidance-related social information. Journal of Personality and Social Psychology, 78, 1024–1037. doi 10.1037//0022-3514.78.6.1024 Wiggins, J. S., & Trapnell, P. D. (1996). A dyadic-interactional perspective on the five-factor model. In J. S. Wiggins (Ed.), The fivefactor model of personality: Theoretical perspectives (pp. 88–162). New York: Guilford.

Wojciszke, B., BaryBa, W., Parzuchowski, M., Szymkow, A., & Abele, A. E. (2011). Self-esteem is dominated by agentic over communal information. European Journal of Social Psychology, 41, 617–627. doi 10.1002/ejsp.791 Wojciszke, B., & Klusek, B. (1996). Moral and competence-related traits in political perception. Political Psychology Bulletin, 27, 319–324.

Astrid Mignon SCALab CNRS UMR 9193 U. F. R. de Psychologie Université de Lille B. P. 60149 59653 Villeneuve d’Ascq Cedex France astrid.mignon@univ-lille3.fr

Swiss Journal of Psychology (2016), 75 (1), 35–45

A. C. Leite et al.: Ambiguous Sw issJournal Normatives of Psychology and Deviants’ (2016), © 2016 75Tolerance (1), Hogrefe 47–52

Short Research Note

Do Ambiguous Normative Ingroup Members Increase Tolerance for Deviants?

Ana C. Leite1, Isabel R. Pinto2,3, and José M. Marques2,3 1

School of Psychology, University of Kent, UK

2

Faculty of Psychology and Educational Sciences, University of Porto, Portugal

3

Institute of Social Sciences, University of Lisbon, Portugal

Abstract. Subjective group dynamics theory (Marques, Páez, & Abrams, 1998) proposes that deviant ingroup members who threaten the positive value of the group members’ social identity are evaluated negatively. In an experiment, we investigated whether group members evaluate deviant ingroup members less negatively when the normative member’s commitment to the ingroup is ambiguous. Participants evaluated one normative and one deviant ingroup or outgroup member. Two conditions were contrasted, in which the normative target showed high versus low commitment to the group. As predicted, the participants evaluated deviant ingroup targets more negatively and normative ingroup targets more positively than their respective outgroup counterparts – but only when the normative member’s commitment to the ingroup was unambiguous. When presented with a normative member with ambiguous commitment, the deviant ingroup member was evaluated less negatively.

We discuss these results in light of subjective group dynamics theory.

Keywords: deviance in groups, tolerance for deviance, black sheep effect, subjective group dynamics

Subjective group dynamics theory (SGDT; Marques, Páez, & Abrams, 1998; Pinto, Marques, Levine, & Abrams, 2010) proposes that group members evaluate deviant ingroup members negatively because they jeopardize the group’s positive image and thus threaten the group members’ social identity. According to Marques and Páez (1994), group members judge deviant ingroup members negatively in an attempt to maintain a positive social identity, and they judge normative ingroup members positively because they enhance social identity. In support of this idea, evidence on the black sheep effect (e.g., Marques & Páez, 1994; Marques, Yzerbyt, & Leyens, 1988) shows that people judge desirable ingroup members more favorably and undesirable ingroup members more unfavorably than their respective outgroup counterparts, especially when their social identity is threatened or insecure (Marques, Páez, & Abrams, 1998; see also Marques, 2004; Marques & Páez, 2008).

Research on the black sheep effect has concentrated on the conditions under which deviant ingroup members represent a relevant threat, and normative ingroup members yield relevant support, to positive social identity. These conditions typically include the activation of a prescriptive focus on group members’ behavior (Marques, Páez, et al., 1998), the existence of low intragroup cohesiveness around a violated norm (Marques, Abrams, & Serôdio, 2001), the lack of social validation of that norm outside the group (Marques et al., 2001), or the intragroup status of normative and deviant members (Pinto et al., 2010). In addition, the focus of such research was directed toward the negative impact of deviant members on group members’ social identity. To our knowledge, no attention has been © 2016 Hogrefe DOI 10.1024/1422-4917/a000170

paid to the role of normative members in this process. The present study attempts to fill this gap. We propose that the black sheep effect results not only because people recognize that deviant ingroup members threaten their social identity, but also because they recognize that normative ingroup members offer strong support for the norm that justifies a reaction against deviants.

Normative ingroup members are important because they inform about the appropriate behavior expected from group members (Marques, Abrams, Páez, & Martinez-Taboada, 1998). As a result, group members should only be more motivated to exert normative pressure on deviant ingroup members, as opposed to avoiding them, when there is high ingroup support for that norm (see also Frings, Abrams, Randsley de Moura, & Marques, 2010). Indeed, the adoption of a prescriptive focus toward deviant ingroup members would depend on the norm’s perceived resilience to deviant threats. Such resilience should depend on whether salient members adopt normative conduct as well as the amount of perceived commitment of these members to the norm they embody.

As a rule, previous research characterized normative members as those who adopt socially desirable conduct, or endorse generic prescriptive norms (Marques et al., 2001), and simultaneously hold a positive orientation toward the group. This may have led participants to consider target members’ desirable or undesirable behavior as the equivalent of their commitment or lack of commitment, respectively, to the group. However, we can conceive of situations in which group members adopt normative behavior while not being genuinely committed Swiss Journal of Psychology (2016), 75 (1), 47–52

48

to the group. Conversely, in other situations, group members may be genuinely committed to the group while adopting behavior deemed deviant by other group members (see Packer, 2008). Thus, in many social situations, the reaction to deviance may depend on the level of commitment that normative members exhibit. Therefore, it would be interesting to separate the normative component of the behavior from the level of group commitment of the member adopting such behavior, that is, to examine how the normative member’s level of commitment to the group affects group members’ judgments of normative and deviant members, as well as their stronger or weaker adherence to the norm.

In this investigation, we do not expect ambiguous normative members (i.e., members who adopt the expected normative behavior while showing little commitment to the group) to be perceived as providing normative support strong enough to activate a prescriptive focus. As a result, we do not expect to observe a black sheep effect under such conditions.

Indeed, because of their lack of commitment to the group, ambiguous normative members should be perceived as being nonreferential for the ingroup’s normative position. In this case, deviant ingroup members should be evaluated with relative leniency, and the normative members should not be evaluated more positively than in situations in which normative members’ behavior and commitment to the group are in line with each other.

A. C. Leite et al.: Ambiguous Normatives and Deviants’ Tolerance

sition can be validated, we expected the differential agreement between the normative and deviant positions to be associated with evaluative differentiation between the normative and the deviant targets, especially in the ingroup/unambiguous condition. We expected evaluations of an ambiguous normative ingroup member to be negatively influenced by these members’ lack of commitment to the group and thus less positive than evaluations of normative ingroup members whose normative behavior emerges along with a strong commitment to the group.

Method Participants and Design A total of 26 female and 22 male students (N = 48) who were enrolled in one of two schools at the University of Porto (School of Arts or School of Architecture) participated in this experiment and were randomly assigned to conditions. We used a 2 × 2 × 2 design (Group [ingroup, outgroup] × Normative Targets’ Commitment [unambiguous, ambiguous] × Targets’ Position [normative, deviant]); the first two were between-participants factors and the latter was a within-participant factor.

Overview and Hypotheses The participants were university students who were told that students from their university were involved in an evaluation of the Bologna Process. They were presented with one normative and one deviant target, both of which were either students from their own school of the university (ingroup) or another school of the university (outgroup). Depending on the experimental condition, the normative target either showed high (unambiguous condition) or low (ambiguous condition) commitment to their university.

In the unambiguous condition, we expected the participants to judge the deviant ingroup target more negatively than the deviant outgroup target, and to judge the normative ingroup target more positively than the normative outgroup target (black sheep hypothesis). However, in the ambiguous condition as compared to the unambiguous condition, we expected the participants to judge the normative ingroup targets less positively and the deviant ingroup targets less negatively. Following the same reasoning, we expected the participants to agree more with the normative targets and disagree more with the deviant targets in the ingroup/unambiguous condition than in the other conditions (agreement hypothesis). Finally, as the presence of an unambiguous normative target (in the ingroup/unambiguous condition) increases the likelihood that the normative poSwiss Journal of Psychology (2016), 75 (1), 47–52

Procedure One experimenter informed participants that she was working for a department of their university whose mission was to track progress in the implementation of the Bologna Process. She proceeded to inform the participants that, as a part of this evaluation process, students would be invited to take part in a series of forthcoming group discussions on important aspects of the Bologna Process, and that student teams were being created for that purpose. Participants, the experimenter went on, were taking part in a validation process, in which they were asked to help establish whether the opinions previously voiced by students who had been selected to participate in the teams were representative of the opinions of the students of their respective schools. The participants were then handed two folders, each concerning one target student (normative target and deviant target) who, ostensibly, had participated in a recent discussion.

Each folder contained information about the target’s school and the target’s opinion about the involvement of students in the evaluation of the Bologna Process. In the folder concerning the normative target, there was also information about the target’s position on the participation of ingroup students (commitment manipulation). © 2016 Hogrefe

A. C. Leite et al.: Ambiguous Normatives and Deviants’ Tolerance

49

Group

– “As a student, how representative of your school do you

The targets were presented as either studying in the same school as the participant (ingroup condition) or in the other school (outgroup condition).

– “How much do you identify with your school?”

consider yourself to be?”

We averaged the responses to the four items to compute a social identification score (Cronbach’s α = .70).

Targets’ Position The normative target endorsed a socially desirable position (“University students should be involved in the evaluation of the Bologna Process”), while the deviant target endorsed a socially undesirable position (“University students are not mature enough to participate in the evaluation of the Bologna Process”). These positions were adapted from Pinto et al. (2010).

Agreement with Targets’ Position After the experimental manipulations, we asked the participants: “How much do you agree with Student A’s/B’s (normative/deviant) position?” (1 = totally disagree, 7 = totally agree).

Evaluation of Targets Commitment Participants learned that the normative target either supported or opposed the involvement of own-school students in that process. In the ambiguous condition, the normative target agreed with the normative position, but stated that “the students of my school should not take part in this process or become members of the student committee that will represent our university.” In the unambiguous condition, participants read no statement by the normative target1.

The participants evaluated the targets on five 7-point scales (1 = bad fellow, unreasonable, selfish, boring, disloyal; 7 = good fellow, reasonable, altruistic, stimulating, loyal). For each participant, we averaged the evaluations of each target with respect to these traits to create a normative target score (Cronbach’s α = .88) and a deviant target score (Cronbach’s α = .86).

Results Measures Participants responded to three sets of questions tapping social identification, agreement with targets’ position, and targets’ evaluation, respectively.

Social Identification In order to control for a priori differences in social identification, participants responded to the following questions (1 = not at all; 7 = completely): – “How competent are the other students in your school?”

– “How similar are you to the other students in your school?”

Social Identification On average, participants identified with the ingroup, M = 5.36, SD = 0.74. A 2 × 2 (Group [ingroup, outgroup] × Commitment [unambiguous, ambiguous]) ANOVA on the social identification score yielded a significant effect of group, F(1, 44) = 5.51, p = .023, ηp2 = . 111 (remaining effects Fs < 3.09, ps ≥ .086, ηp2s ≤ .066). We, thus, controlled for potential effects of a priori differences in social identification on all of our dependent measures using a regression analysis (Muller, Yzerbyt, & Judd, 2008). The regression analysis revealed no significant effects of social identification on the model terms (lowest B = –0.43, SE = 0.41, β = –.21, p = .303). Social identification did not significantly change any effects on the dependent measures.

1 It could be argued that participants perceive the unambiguous normative member as being more normative than the ambiguous normative member. In this case, participants’ judgments would be influenced by the normative character instead of the group representativeness of the normative member. In order to assure the meaning of commitment manipulation, we conducted a postexperimental study in which we manipulated the commitment of normative ingroup targets and measured the perceived normativeness and ingroup representativeness of these targets. We expected both normative targets to be perceived as being equally normative, and the ambiguous normative targets as being less representative of the group, suggesting that they are less central to the group. The results supported our expectations: for target’s normativeness: ambiguous condition (M = 5.41, SD = 1.46), unambiguous condition (M = 5.69, SD = 0.90), t(48) = 0.18, p = .417, d = 0.06, 95% CI [–0.41, 0.97]; for target’s representativeness: ambiguous condition (M = 4.24, SD = 1.45), unambiguous condition (M = 5.04, SD = 1.10), t(48) = 2.20, p = .033, d = 0.64, 95% CI [0.07, 1.53]. Thus, the commitment manipulation has an impact on the participants’ perception of how representative of the group the target is and not on how normative he/she is.

© 2016 Hogrefe

Swiss Journal of Psychology (2016), 75 (1), 47–52

50

A. C. Leite et al.: Ambiguous Normatives and Deviants’ Tolerance

Agreement With Targets’ Position

Evaluation of Targets

We expected participants in the ingroup/unambiguous condition to agree more with the normative targets’ position, and to disagree more with the deviant targets’ position, than participants in the remaining conditions would. To test this, we ran a 2 × 2 × 2 (Group [ingroup, outgroup] × Commitment [unambiguous, ambiguous] × Targets’ Position [normative, deviant]) ANOVA on agreement with the normative and deviant targets’ position scores (see Table 1). We found a significant effect of agreement, F(1, 44) = 80.92, p < .001, η2 = .648, showing that participants agreed more with the normative (M = 5.47, SD = 1.34) than with the deviant target (M = 3.34, SD = 1.23). In addition, in partial support of our prediction, the significant Commitment × Targets’ Position interaction, F(1, 44) = 15.38, p < .001, ηp2 = .259, shows that participants agreed less with the normative target in the ambiguous than in the unambiguous condition (M = 4.82, SD = 1.44, and M = 6.13, SD = 0.84), respectively, t(46) = 3.87, p < .001, d = 1.14, 95% CI [–2.00, –0.63]. However, contrary to what we expected, there was no difference in agreement with the deviant target in the ambiguous versus unambiguous conditions, t(46) = 1.50, p = .140, d = 0.44, 95% CI [–0.18, 1.23]. The remaining effects were nonsignificant (Falways ≤ 2.57, ps ≥ .116, ηp2s ≤ .032).

We only expected to observe the black sheep effect in the unambiguous condition. Therefore, we predicted that the deviant ingroup target would be evaluated less negatively, and the normative ingroup target evaluated less positively, in the ambiguous as compared to the unambiguous condition. To test these predictions, we conducted a 2 × 2 (Group [ingroup, outgroup] × Commitment [unambiguous, ambiguous]) ANOVA on the normative and deviant targets’ scores (see Table 2). This analysis yielded a significant effect of targets’ position, F(1, 44) = 24.80, p < .001, ηp2 = .360, and significant interactions between Commitment × Targets’ Position, F(1, 44) = 10.70, p = .002, ηp2 = .196, and Group × Commitment × Targets’ Position, F(1, 44) = 15.23, p < .001, ηp2 = .257. The effects of Group, Commitment, and the interactions between Group × Targets’ Position as well as Group × Commitment (Falways ≤ 2.36, ps ≥ .132, ηp2s ≤ .051) were not significant.

Table 1. Means (and standard deviations) of agreement with targets’ position as a function of group and commitment Group Commitment

Target

Ingroup

Outgroup

Ambiguous

Normative

4.86 (1.67)

4.79 (1.31)

Deviant

3.63 (0.66)

3.58 (1.44)

Unambiguous

Normative

6.65 (0.45)

5.69 (0.86)

Deviant

2.95 (1.23)

3.18 (1.34)

We did not obtain a significant full interaction. However, the mean pattern of results obtained across conditions was consistent with our hypothesis, especially regarding agreement with the normative target. We thus conducted independent-sample t-tests between conditions (all possible combinations), which showed that participants agreed more with the normative target in the ingroup/unambiguous condition than in the remaining conditions, t(23) = 4.49, p < .001, d = 1.87, 95% CI [1.05, 2.72]. Conversely, the obtained pattern of results did not support our prediction that participants would agree less with the deviant target in the ingroup/unambiguous condition than in all other conditions (talways ≤ 1.56, p ≥ .136, d = 0.72, 95% CI [–1.59, 0.23]).

In sum, these results provide partial support for our hypothesis. Specifically, they indicate that participants tended to agree more with a normative ingroup member with an unambiguous position than with a normative ingroup member with an ambiguous position or any normative outgroup member.

Swiss Journal of Psychology (2016), 75 (1), 47–52

Table 2. Means (and standard deviations) of evaluation of targets as a function of group and commitment Group Commitment

Target

Ingroup

Outgroup

Ambiguous

Normative

5.02 (1.44)

4.84 (1.00)

Deviant

5.11 (1.19)

4.29 (1.10)

Normative

5.81 (1.01)

4.88 (0.75)

Deviant

3.94 (0.62)

4.50 (0.78)

Unambiguous

The significant effect of targets’ position shows that participants evaluated the normative target more favorably (M = 5.11, SD = 1.09) than the deviant target (M = 4.44, SD = 1.00). Moreover, the interaction between Commitment × Targets’ Position shows that participants’ evaluations of the normative (M = 5.31, SD = 0.98) and deviant (M = 4.24, SD = 0.75) targets only differed in the unambiguous condition, t(23) = 4.94, p < .001, d = 1.23, 95% CI [0.62, 1.51]. In the ambiguous condition, the differences between their evaluations of the normative target (M = 4.92, SD = 1.18) and the deviant target (M = 4.63, SD = 1.19) were not significant, t(23) = 1.26, p = .220, d = 0.24, 95% CI [–0.18, 0.76].

In order to test our black sheep hypothesis, we decomposed the significant Group × Commitment × Targets’ Position interaction according to the commitment factor. As expected, Group × Targets’ Position was significant in the unambiguous condition, F(1, 45) = 11.14, p = .002, ηp2 = .198, but not in the ambiguous condition, F(1, 45) = 1.36, p = .249, ηp2 = .029. In addition, as predicted, in the unambiguous condition, participants evaluated the ingroup normative target more positively (M = 5.81, SD = 1.01) than the outgroup normative target (M = 4.88, SD = 0.75), t(22) = 2.57, p = .017, d = 1.10, 95% CI [0.18, 1.67]. The deviant ingroup target (M = 3.94, SD = 0.62) tended to be evaluated more negatively than the outgroup de© 2016 Hogrefe

A. C. Leite et al.: Ambiguous Normatives and Deviants’ Tolerance

viant target (M = 4.50, SD = 0.78), t(22) = 1.92, p = .068, d = 0.82, 95% CI [–1.17, 0.44].

We also decomposed the interaction according to the group factor. Commitment × Targets’ Position was significant in the ingroup condition, F(1, 45) = 22.73, p < .001, ηp2 = .336, but not in the outgroup condition, F(1, 45) = 0.19, p = .667, ηp2 = .004. Participants evaluated the normative ingroup target positively, regardless of the targets’ commitment, t(19) = 1.47, p = .159, d = 0.67, 95% CI [–1.92, 0.34] (unambiguous condition, M = 5.81, SD = 1.01; ambiguous condition, M = 5.02, SD = 1.44). Moreover, as predicted, participants evaluated the deviant ingroup target more positively in the ambiguous condition (M = 5.11, SD = 1.19) than in the unambiguous condition (M = 3.94, SD = 0.62), t(19) = 2.87, p = .010, d = 1.32, 95% CI [0.32, 2.02]. These results fully support our predictions. The participants only evaluated normative ingroup members more positively and deviant ingroup members more negatively than their respective outgroup counterparts (the black sheep pattern) when the normative member was perceived as being committed to the ingroup. Moreover, deviant ingroup members were only evaluated more negatively when they were presented with an unambiguous normative member.

Association Between Agreement with Targets’ Position and Evaluation of Targets We predicted that the association between participants’ differential agreement with the normative and deviant targets and differential evaluation of these same targets would be stronger in the ingroup/unambiguous condition than in the other conditions. The correlations between these two scores within experimental conditions are consistent with this prediction. In the ingroup/unambiguous condition, the more participants disagreed with the deviant target and agreed with the normative target, the more their evaluations of these targets differed, r = .59, p = .055, N = 11 (remaining conditions: ralways < .14, ns).

This suggests that agreement with the normative and deviant positions is associated with targets’ evaluations, especially when the normative ingroup member is perceived as being committed to the group.

Discussion The results support our prediction that participants would agree more with an opinion espoused by a normative ingroup target with an unambiguous as compared to ambiguous position. However, the results do not support our prediction that participants would disagree more with the deviant ingroup target when presented with a normative ingroup target with an © 2016 Hogrefe

51

unambiguous position. These results indicate that uncommitted normative targets are not sufficient to trigger an opinion change toward the deviant opinion, but enhance the positive influential role of an unambiguous normative member to potentiate adherence to the normative position.

Regarding evaluations of the targets, the results support the idea that participants are more lenient toward deviant ingroup members when these members emerge in parallel with ambiguous normative ingroup members: We only observed a black sheep effect when there was unambiguous normative support for the ingroup. Furthermore, the deviant ingroup target was evaluated more positively in the ambiguous than in the unambiguous condition. Finally, the positive association between evaluative differentiation and the difference in agreement with the normative and the deviant targets found in the ingroup/unambiguous condition suggests that it was in this condition that participants more strongly attempted to validate the normative position.

Taken together, the results offer preliminary, yet compelling, evidence in support of the idea that, when people are faced with deviant ingroup members, their reaction to them depends on the extent to which normative ingroup members are referential by supporting the ingroup position and thereby showing their commitment to the group. Their commitment seems to facilitate participants’ tendency to address the nefarious effects of deviance by evaluating the normative ingroup targets positively and the deviant ingroup targets negatively while simultaneously increasing their agreement with the normative position in question. Conversely, people may be willing to tolerate deviant members when the group lacks the necessary normative support provided by committed members.

Previous research only found the black sheep effect when both the normative and the deviant targets were highly representative ingroup members (Pinto et al., 2010). The present work extends Pinto et al.’s findings by showing that participants’ evaluations of deviant members were more negative when the normative position was supported by a committed group member. However, when the normative group member held an ambiguous position, participants showed more leniency toward the deviant member. Thus, the attitudinal (commitment) component is an important factor in determining whether the group has unambiguous and sufficient normative support to highlight the prescriptive attribute of the violated norm and thus to react to deviance. Our results support the idea that normative members who are less committed to the group might not be enough to trigger the black sheep effect. In turn, they may result in lenient judgments of deviant members. Although ambiguous normative members are not sufficient to trigger opinion change toward the deviant position, they result in tolerance for deviant positions.

These results may have important implications for SGDT, which inspired the present study, as well as for our understanding of reactions to deviance. SGDT proposes that, in the presence of deviant ingroup members, individuals try to re-establish Swiss Journal of Psychology (2016), 75 (1), 47–52

52

the group’s positive value by engaging in behavior that results in the black sheep effect (Marques et al., 1988). Although recent research has focused on the conditions in which deviant ingroup members are tolerated (e.g., Abrams, Randsley de Moura, & Travaglino, 2013; see also Abrams, Randsley de Moura, Marques, & Hutchison, 2008; Randsley de Moura & Abrams, 2013), this research did not investigate the impact of normative members on the evaluation of deviant group members. Our results suggest that group members’ recognition that normative members are committed to the ingroup is crucial for their commitment to protecting the group against the risks of deviance, both by evaluating deviant members negatively and increasing their adherence to the norms violated by deviant members. Although more conclusive research is clearly required to elucidate this issue, the present study offers a potentially valuable extension to SGDT and to our understanding of the processes involved in the way groups react to deviance.

Acknowledgments This research was supported by FCT (Fundação para a Ciência e a Tecnologia) Grants SFRH/BD/66324/2009 and EXCL/MHCPSO/0800/2012.

References Abrams, D., Randsley de Moura, G., Marques, J. M., & Hutchison, P. (2008). Innovation credit: When can leaders oppose their group’s norms? Journal of Personality and Social Psychology, 95, 662–678. doi 10.1037/0022-3514.95.3.662 Abrams, D., Randsley de Moura, G., & Travaglino, G. (2013). A double standard when group members behave badly: Transgression credit to ingroup leaders. Journal of Personality and Social Psychology, 105, 799–815. doi10.1037/a0033600 Frings, D.,Abrams, D., Randsley de Moura, G.,& Marques, J. M. (2010).

The effects of cost, normative support, and issue importance on motivation to persuade in-group deviants. Group Dynamics: Theory, Research, and Practice, 14, 80–91. doi 10.1037/a0016092 Marques, J. M. (2004). Déviance, normativité, et dynamique de groupes subjective [Deviance, normativeness, and subjective group dynamics]. New Review of Social Psychology, Festschrift in Honor of Willem Doise, 1–2, 29–37.

Swiss Journal of Psychology (2016), 75 (1), 47–52

A. C. Leite et al.: Ambiguous Normatives and Deviants’ Tolerance

Marques, J. M., Abrams, D., Páez, D., & Martinez-Taboada, C. (1998).

The role of categorization and ingroup norms in judgments of groups and their members. Journal of Personality and Social Psychology, 75, 976–988. doi 10.1037/0022-3514.75.4.976 Marques, J. M., Abrams, D., & Serôdio, R. G. (2001). Being better by being right: Subjective group dynamics and derogation of in-group deviants when generic norms are undermined. Journal of Personality and Social Psychology, 81, 436–447. doi 10.1037//00223514.81.3.436 Marques, J. M., & Páez, D. (1994). The “black sheep effect”: Social categorization, rejection of ingroup deviates, and perception of group variability. European Review of Social Psychology, 5, 37–68. doi 10.1080/14792779543000011 Marques, J. M., & Páez, D. (2008). Dynamique de groupes subjective: Un cadre théorique pour l’effet brebis galeuse [Subjective group dynamics: A theoretical framework for the black sheep effect]. In R. V. Joule & P. Huguet (Eds.), Bilans et perspectives en psychologie sociale (Vol. 2, pp. 71–115). Grenoble, France: Presses Universitaires de Grenoble.

Marques, J. M., Páez, D., & Abrams, D. (1998). Social identity and intragroup differentiation as subjective social control. In S. Worchel, J. F. Morales, D. Páez, & J.-C. Deschamps (Eds.), Social identity: International perspectives (pp. 124–141). London, UK: Sage.

Marques, J. M., Yzerbyt, V. Y., & Leyens, J.-P. (1988). The black sheep effect: Judgmental extremity toward ingroup members as a function of ingroup identification. European Journal of Social Psychology, 18, 1–16. doi 10.1002/ejsp.2420180102 Muller, D., Yzerbyt, V. Y., & Judd, C. M. (2008). Adjusting for a mediator in models with two crossed treatment variables. Organizational Research Methods, 11, 224–240. doi 10.1177/1094428106296639 Packer, D. J. (2008). On being both with us and against us: A normative conflict model of dissent in social groups. Personality and Social Psychology Review, 12, 50–72. doi 10.1177/1088868307309606 Pinto, I. R., Marques, J. M., Levine, J. M., & Abrams, D. (2010). Membership status and subjective group dynamics: Who triggers the black sheep effect? Journal of Personality and Social Psychology, 99(1), 107–119. doi 10.1037/a0018187 Randsley de Moura, G., & Abrams, D. (2013). Bribery, blackmail, and the double standard for leader transgressions. Group Dynamics: Theory, Research, and Practice, 17(1), 43–52. doi 10.1037/ a0031287

Ana C. Leite School of Psychology Centre for the Study of Group Processes University of Kent, Canterbury Kent CT2 7NP UK a.castro-leite@kent.ac.uk

© 2016 Hogrefe

Swiss Journal of Psychology © 2016 Hogrefe

Instructions to Authors Overview. The Swiss Journal of Psychology is published by the Swiss Psychological Society (member of the International Union of Psychological Science). It is a peer-reviewed journal that serves as a forum for the scientific discussion of all areas of psychology and their applications. Its central concern is the mediating function between different fields of psychology, between research and applications, and between psychology in culturally different areas both within Switzerland as well as in other nations. Internet: http://www. ssp-sgp.ch.

Manuscript Types. Two kinds of articles are published: Regular Articles and Short Research Notes. In addition, occasional special issues, invited articles or editorial statements are published. Economy of presentation is strongly recommended.

Manuscript Length and Format. Regular Articles should usually not be longer than 6000 words (including main text plus notes, acknowledgments, and appendices). Short Research Notes are limited to 3500 words (including main text plus notes, acknowledgments, and appendices) and not more than 25 items on the reference list. Manuscripts should be prepared according to the 6th edition of the Publication Manual of the American Psychological Association (APA) (http://apastyle.org). In particular, they should be typed with wide margins and double spaced throughout. Figures and tables should be supplied on separate sheets grouped at the end of the manuscript with their location in the text clearly indicated. Footnotes should be avoided if possible. Authors are responsible for obtaining and supplying written permission (with the manuscript) to reprint any material from copyrighted sources. References to published work must be in APA style in the text (names, year) and in alphabetical order at the end of the manuscript. The first page should contain a concise but informative title followed by the names of the authors, the names of the institution(s) with which the authors are affiliated, and a running title not exceeding 48 letters and spaces. The manuscript should be preceded by an abstract of maximum 200 words, and 4–6 keywords, both in English. At the end of the manuscript the full postal address of the author responsible for checking proofs and subsequent correspondence should be given along with telephone, telefax numbers and e-mail address.

It is recommended that authors who are not native speakers of English have their papers checked and corrected by a native speaker colleague before submission. Standard US American spelling and punctuation as given in Webster’s New Collegiate Dictionary should be followed.

Manuscript Submission. Only original papers written in English will be considered for publication. Submissions are only accepted online at www.editorialmanager.com/sjp. The online manuscript submission and review system offers easy and straightforward log-in and submission procedures.

This system supports a wide range of submission file formats: for manuscripts – Word, WordPerfect, RTF, TXT and LaTex; for figures – TIFF, GIF, JPEG, EPS, PPT, and Postscript. Please follow the online instructions for submission.

Should you have any technical queries regarding this process, please contact: gregory.mantzouranis@chuv.ch Submission of manuscripts is taken to imply that neither the manuscript nor any component of it has already been published or is currently under consideration by another Swiss Journal of Psychology © 2016 Hogrefe

journal. Authors should ensure that the manuscripts do not contain any information explicitly identifying them, with the exception of the title page and acknowledgments. Manuscripts will be initially screened for scientific quality and compliance with formal aspects before being sent out for peer review (generally two referees). After revision, the final version of the article accepted for publication should also be supplied in electronic form.

Offprints. The publisher will send the corresponding author of each accepted paper an e-offprint (PDF) of the published version of the paper when it is first released online.

This e-offprint is provided for the author’s personal use, including for sharing with coauthors (see also “Online Rights for Journal Articles” on the publisher’s website at http://www.verlag-hanshuber.com/informationen).

Copyright Agreement. The first author confirms and guarantees on behalf of himself or herself and any coauthors that he or she holds all copyright in and titles to the submitted contribution, including any figures, photographs, line drawings, plans, maps, sketches, and tables, and that the article and its contents do not infringe in any way on the rights of third parties. Upon acceptance of the article for publication, the author agrees 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 Swiss law. Furthermore, the author transfers to the publisher the following exclusive rights to the article and its contents: a) 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. b) 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 on 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. c) The rights to store the article and its contents 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 to third parties for viewing or for 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. d) 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. e) The right to transfer any or all of the rights mentioned in this agreement as well as the rights retained by the Verwertungsgesellschaft «WORT» including the corresponding royalty rights to third parties within or outside Switzerland.