216 HEERDINK.VAN KLEEF.HOMAN.AND FISCHER publicat the start of th groupdiscussion.which oud follow th which y rch (see Steinel et a and they were offered an opp ity to co eal thei rmity by co his foun ed th th Subsequ tly.in the per cral conditio zly to 100 vere ne ormality o heres、cans that you have le d the roup was coded as conf formity (6).and 1 No m vement (N)was coded as missing.s this behavior terms of conformity/de After leaming how thei pe ed with that of the other group nea re informed that a few inan of the moo and Ihe group expects me and uld ask participants to on checks The manipulation of majority emotior motion that he or she the (from 1= e at all to c)how much anser and gfrom the group s had been communicate group membe se emotions by having ts answer the sar on for nd th at r surprise,disappoi tment,,and s dness).The ma and that the s int 91) Upon the em red,the participant told that the group discussion ha participants they had be in case his or her result to nding Ecept for one weeks er participants had ta part in the exper ut I'm happy lan cone rated the pain questions about art.It was plained that afte t of he group.an ing to h ist of rating s and ar e-maladthe ipant saw that three group members had the same emo The posttest was terminated 4 weeks after the end of the initial lal ticipants and their comments did not yield any insights and will st which was another evaluation of the focal paintine.Par ot be discussed furt fer indica icipants were parti Then.we first ting (a random pa scond time.It was explained that these ratings would be made ing selccted from the 40 nonfocal paintings valuated in the give them some more information about their fellow group mem￾bers, which was used to manipulate prototypicality. In this proce￾dure, which was adapted from previous research (see Steinel et al., 2010; van Kleef et al., 2007), participants first read that their scores in the personality test had been compared with those of the other group members. Subsequently, in the peripheral condition, they read, “There is little overlap between your personality and your fellow group members. This means that you have little in common with the other participant and that you’re not typical for your group.” (emphasis in original). In the prototypical condition, the italicized words were changed to “much” (twice) and “very,” respectively. Majority emotion manipulation. After learning how their per￾sonality compared with that of the other group members, partici￾pants learned that aesthetic preferences are, in large part, deter￾mined by the emotions that people experience while viewing art and that personality is an important determinant of the emotions that people experience. To investigate whether emotions could indeed account for the link between personality and aesthetic preferences, the experimenter would ask participants to complete a group communication task in which every participant could indi￾cate the emotion that he or she experienced while viewing the painting that received the most different ratings from the group members. On the following screen, the focal painting was dis￾played, together with the ratings of the other group members. These were presented anonymously and were all around the 90th scale point and thus constituted a descriptive norm. Since the painting that received most different ratings had been selected, the participant could deduce from this information that he or she had been the one with the different rating and that the other group members could see the participant’s deviant rating on their screen. After participants had been exposed to the group norm, the emotion communication round was introduced, which contained the manipulation of majority emotion. In this task, all group members could select an emoticon (e.g., a smiley) from a sample of 15 emoticons to represent the emotion they experienced while viewing the focal painting. If they wanted, they could also write a comment. This emoticon and comment would be sent to the other group members. First, participants learned they had been randomly selected to be the last group member to indicate their emotions. Then, depending on the condition, the group member to go first selected either an angry emoticon or a happy emoticon (i.e., a smiley) and wrote, “I don’t really feel a certain emotion about this painting, but I’m happy [angry] that someone rated the painting so differently from the rest of the group, and we’re going to have to talk about that. It’ll be an interesting discussion!” The second group member selected the same emoticon as the first without sending a comment, and the third group member selected the same emoticon as the other two and commented “Me too!” Thus, the participant saw that three group members had the same emotion (either happiness or anger), which one of them attributed to the behavior of the participant, while the others concurred. Finally, the participant could pick an emoticon and write a comment to send to the other group members. Analysis of the emoticons chosen by the participants and their comments did not yield any insights and will not be discussed further. T2 conformity. After indicating and commenting on their own emotions, participants proceeded to rate the same focal painting a second time. It was explained that these ratings would be made public at the start of the group discussion, which would follow the computerized session. Thus, participants felt accountable for their behavior, and they were offered an opportunity to conceal their previous nonconformity by conforming in this round. Then, par￾ticipants rated the same focal picture again using a slider (from 0
ugly to 100
beautiful). Because of severe nonnormality of the difference scores between the first and second ratings, the change in judgment was coded as a binary variable. Movement toward the group norm was coded as conformity (N
26), and movement away from the group norm was coded as deviance (N
57). No movement (N
3) was coded as missing, as this behavior could not be interpreted in terms of conformity/deviance. Acceptance/rejection scale. After completing the conformity measure, participants were informed that a few more measures would be completed before the group discussion would begin. Perceived rejection was measured using a six-item scale composed of the four items used in Studies 1, 2, and 4 and two new items (“I feel pressure to yield to the group,” and “The group expects me to stay in line”; Cronbach’s
.81). Manipulation checks. The manipulation of majority emotion was checked by asking participants to indicate on 7-point scales (from 1
none at all to 7
very much) how much anger and happiness had been communicated by their fellow group members during the communication round. We avoided drawing too much attention to these emotions by having participants answer the same question for six more emotions that had also been represented by emoticons earlier during the emotion communication round (dis￾gust, surprise, disappointment, contempt, and sadness). The ma￾nipulation of prototypicality was checked with four items (e.g., “I have little in common with my fellow group members” (reverse￾scored), and “My fellow group members’ personality is compara￾ble to mine.”), which were answered on 7-point scales (Cronbach’s
.91). T3 conformity. Upon completing all measures, participants were instructed to notify the experimenter. When the experimenter entered, the participant was told that the group discussion had unfortunately been canceled due to practical reasons. The experi￾menter then asked whether the participant would be willing to be contacted for some further measurements, in case his or her results showed inconsistencies or unexplainable findings. Except for one, all participants agreed and left their telephone numbers. Three weeks after participants had taken part in the experiment, they were contacted via e-mail and invited for a posttest consisting of some more questions about art. It was explained that after analyzing the data, some questions had remained unanswered, which we wanted to resolve using the posttest. The task would consist of rating some more paintings and answering some mis￾cellaneous questions about art. Participants were contacted a max￾imum of three times, once via e-mail, and then twice by telephone. The posttest was terminated 4 weeks after the end of the initial lab experiment, at which point 61 participants had taken the posttest. Participation in the posttest was not predicted by (any combination of) the manipulations. The first part of the posttest only contained one measure of interest, which was another evaluation of the focal painting. Par￾ticipants were told that we wanted them to rate some more paint￾ings, some of which could be familiar, some of which could be new. Then, we first presented a familiar painting (a random paint￾ing selected from the 40 nonfocal paintings evaluated in the This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 276 HEERDINK, VAN KLEEF, HOMAN, AND FISCHER