ON THE SOCIAL INELUENCE OF EMOTIONS IN GROUPS 275 freely accessible galleries and museums (of Bad Art. en that no et.As art ca y dim adarorg)on the Inteme cin the group.Prot typical n mber nown to be influ ced by psychological variables (e.gn are stable nd the 2007 our selection of paintings as constant as p ssible Thus,we usc affected by the majority's emotions than the behavior of peripheral nembers Method The group norm could be kept constant between participan Fo hich was camed that their fo cted for a follow- up posttest.in which the sa but all participants expressed a mild dislike of their focal painting conta Thus,two to strongly like this pa oint as th ale,we ensured】 at participants could shift ard o on the to be dis narticina neir participatio ntal situation.and an additional six parti After completing the measure o dual aesthetic pre nts indicated that they of the persons in the Greer.Jehn and Konins (010)served to mak the situatio to the which ed a pro ndu p dr appear ev nd 64 n (M 2142.range 1835 years).who were randomly whichwould b to the 2 ith th ubj cts design. to the memb f th group men which te the relation between persc hed in the In the first which mplete a p The sex of the people in the pi tures was arr ding the participant)always ion about art with thr .wh were taki After having seen the photos o other group members,participants learned that we wanted to n the same group who kn ach other per ality th n.but the cover story was needed to make the nd the manipulations appear that was used for presenting all instructions and recording all 6 ual ratings. and taking a nality test.which was used to set the change any ng had be 100-betif).These paintings were collected froma number ofexpect to find that only peripheral members would conform more after receiving an angry reaction from a majority (given that no alternative groups are available), as they are especially concerned with gaining acceptance in the group. Prototypical members, on the other hand, are less likely to be affected by the opinion of their peers because they are stable group members and therefore less concerned with gaining acceptance (van Kleef et al., 2007). Thus, we predicted that prototypical members’ behavior would be less affected by the majority’s emotions than the behavior of peripheral members. Method The experiment, which was inspired by a study by Griskevicius et al. (2006), consisted of two parts. The first session took place in the laboratory. During this session, participants were placed in a group, evaluated several paintings, received emotional feedback from a majority about one of their evaluations, and finally rated this specific painting again. Then, 3– 4 weeks later, participants were contacted for a follow-up posttest, in which the same painting was rated again. Thus, two measures of conformity could be calculated, one at the second rating during the session (Time 2 [T2] conformity) and one a few weeks later (T3 conformity). Participants and design. Ninety-seven participants were re￾cruited for a study about aesthetic preferences. In exchange for their participation, they received either course credit or 3.50 euro. Four participants spontaneously expressed suspicion about the reality of the experimental situation, and an additional six partic￾ipants indicated that they knew at least one of the persons in the photos that represented their fellow group members (see “Intro￾duction to the group” in the following text), which posed a prob￾lem for the credibility of the emotion manipulation. After dropping these participants, the final sample consisted of 22 men and 64 women (Mage
21.42, range 18 –35 years), who were randomly assigned to the conditions of a 2 (prototypicality: peripheral or prototypical)  2 (majority emotion: angry or happy) between￾subjects design. Materials and procedure. Upon arrival in the laboratory, participants learned that the aim of the experiment was to inves￾tigate the relation between personality and art preferences, and the way people communicate about art. The experiment would consist of two parts. In the first part, they would complete a personality test and then gave their opinion about a number of paintings individually. In the second part, they would have a group discus￾sion about art with three other participants, who were taking part in the same experiment at the same time in a different building on the university campus, as we wanted to avoid having participants in the same group who knew each other personally. In reality there were no other participants taking part, nor would there be a group discussion, but the cover story was needed to make the situation and the manipulations appear genuine. After obtaining informed consent, participants were seated individually behind a computer that was used for presenting all instructions and recording all answers. Individual ratings. After entering their demographic informa￾tion and taking a bogus personality test, which was used to set the stage for the manipulation of prototypicality, the participant pro￾ceeded to evaluate 41 paintings using a slider (from 0
ugly to 100
beautiful). These paintings were collected from a number of freely accessible galleries and museums (e.g., Museum of Bad Art, http://museumofbadart.org) on the Internet. As art can differ on many dimensions, and preferences on some of these dimensions are known to be influenced by psychological variables (e.g., need for cognitive closure increases preferences for figurative art; Wi￾ersema, van der Schalk, & van Kleef, 2012), we attempted to keep our selection of paintings as constant as possible. Thus, we used only nonfigurative art that was photographed in full color. Pictures were cropped and/or resized to have an identical on-screen size. Apart from lending credibility to the cover story that we were interested in the relation between personality and aesthetic prefer￾ences, rating such a large number of paintings had another aim: The group norm could be kept constant between participants. For this purpose, the painting that was rated closest to the 40th scale point was selected. We refer to this painting as the focal painting. Later in the experiment, participants learned that their fellow group members had consistently rated this focal painting around the 90th scale point. Thus, each participant had a different focal painting, but all participants expressed a mild dislike of their focal painting in the initial round and later learned that the group consensus was to strongly like this painting. By using the 40th scale point as the critical value, we ensured that participants could shift toward or away from the group norm on a second rating of the focal painting. Thus, on the second rating (to be discussed later), participants could both conform or express their deviance from the group if they wanted to do so.7 Introduction to the group. After completing the measure of individual aesthetic preferences, the participant was introduced to the other group members. This procedure, adapted from Homan, Greer, Jehn, and Koning (2010), served to make the situation appear even more genuine and to induce a sense of “groupiness.” First, participants were instructed to make a photo of themselves using the webcam, which would be sent to the other group mem￾bers via the network. They could make as many photos as they liked, until they were happy with the result. After sending their photo to the members of their group, participants saw photos of the other group members appearing one by one on the screen. To minimize chances that the actual participants knew the people in the photos, which could arouse suspicion, we used photos of first￾and second year students from other disciplines who were photo￾graphed in the same cubicles as those in which the study was conducted. The sex of the people in the pictures was arranged so that the four-person groups (including the participant) always contained two males and two females. Prototypicality manipulation. After having seen the photos of the other group members, participants learned that we wanted to 7 This is a slightly simplified description of the actual procedure, in which both the 40th and 60th scale point were used as critical values. After all paintings had been rated, difference scores to these values were calcu￾lated, and the painting with the smallest difference score to either of these values was used as the focal painting. In case of ties (e.g., a painting rated at 39 and one rated at 61), one was selected randomly from the ties. For participants who rated the focal painting at (or around) 60, the group norm was around the 10th scale point. Including the group norm (higher vs. lower than initial rating) as a factor in the analysis did not change any of the results, and this factor is therefore not discussed further. For simplicity, all effects are discussed as if the original painting had been rated at the 40th scale point (thus reversing the ratings for participants that initially rated their focal painting at or around 60). 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. ON THE SOCIAL INFLUENCE OF EMOTIONS IN GROUPS 275