64 DONG,DAL,AND WYER of qu size and were told that we had a different task for each group was as were instructed to folle an expo to leam four ting that the s in s the Durbin-Wat oefficients ranged from 1 55 them to count aloud as they perfommed each movement.In asy M)d( is exercis nditio rthe the ne nature of the study design roo cepts models)The conc ear the the rswere given. HLM remained unchanged from conc ions base 'safphhcakBehevioalnchoayNesmanipulaied and were presented with nM=317.SD=1.10F1.142)=19.11D 001 Reporis about the asked to ch of three brands in ch of f vcs(4.00s.330.F1,142)=7.99p<01.F1. e undoubtedly individual differe in the or p who (In the case of sot 719 other indiv d a sofa from Wildor om Sky Fumiture an r actual bet he on whether s chose the n the indic es that the ma conformit nd participar data yielded s,P( 42) d the ming the exer ha hem feel that th chronous cond (M th in asy to which m and(b)the extent to wh ng the e 97)than in the latter (M 10.60.3 cported the hich they felt s that actor vely correlated with thei eg ch the id th 10)w all)to 5 (ven ch). ally ac d the ively correlated with their pe ceptions that the actors beha .14.D 0).The t t that these correlations synchronously.along the same scale used for other items Table I Results Actors'and Observers'Conformity asa Function of Behavioral Saatistical stratee The design of this and sub Synchrony:Experiment iments t me ough a within or obse Ro Syn Asynchronou g part o B Actor 193 measures in all five experiments.This coefficient,calculated forof equal size and were told that we had a different task for each group. On this pretense, one group was assigned to be actors and the other to be observers. In synchronous exercise conditions, actors were instructed to follow an experimenter to learn four simple exercises involving head, arm, waist, and leg movements. They were told to do the exercises in synchrony with other actors (i.e., to perform the same actions in the same way at the same pace as others). To facilitate their attainment of this objective, we asked them to count aloud as they performed each movement. In asyn￾chronous exercise conditions, the procedure was the same except that actors were not instructed to do the exercises synchronously. While this was going on, the observers waited in another room and could not hear the instructions the actors were given. After the actors had practiced for 2 minutes, the observers reentered the experimental room and watched the actors perform for 10 minutes. All participants then moved to an ostensibly unrelated study (con￾ducted by a different researcher) in which they completed a prod￾uct preference questionnaire similar to that employed by Berger and Heath (2007). Participants were instructed that the researchers were interested in consumers’ preferences and were presented with information allegedly excerpted from Consumer Reports about the market share of different brands. Under this guise, participants were asked to choose one of three brands in each of five different product categories (e.g., car navigation system, sunglasses, and sofa). The alternatives in each category varied in terms of the proportion of persons who owned each. (In the case of sofa, for example, participants were told that, according to an online con￾sumer survey, 71% of other individuals owned a sofa from Wildon Home, 19% owned one from Skyline Furniture, and 10% owned one from Catnapper.) Participants’ choices were coded as 1, 2, or 3 depending on whether participants chose the option with the lowest, middle, or highest market share, respectively. Choices were summed over items to provide a score from 5 to 15, with higher numbers indicating a greater tendency to copy others’ preferences. After the brand choice task, actors indicated the extent to which performing the exercises had made them feel that their freedom of behavior was restricted, whereas observers reported both (a) the extent to which performing the exercises restricted the actors’ freedom and (b) the extent to which observing the exercise made them think about restrictions on their own freedom. In addition, observers reported the extent to which they felt socially excluded and the extent to which they felt they were not an important part of the experiment. All judgments were reported along scales from 1 (not at all) to 5 (very much). Finally, actors reported the extent to which they felt that they did the exercise synchronously, and observers indicated whether they felt the actors did the exercise synchronously, along the same scale used for other items. Results Statistical strategy. The design of this and subsequent exper￾iments meant that the current dependent measures, although as￾sessed individually, was nested within actors’ or observers’ groups. To examine the degree of variability in responses at the individual level that is attributed to being part of the group, we computed the intraclass correlation coefficient, ICC (1) (Bartko, 1976; James, 1982; Kozlowski & Klein, 2000), for the dependent measures in all five experiments. This coefficient, calculated for both the main dependent variables (conformity) and the great majority of other variables in our experiments, was typically small in size and nonsignificant (ranging from .13 to .11, ps  .10), suggesting that the effects of nesting were minimal and the data could be analyzed at the individual level. Furthermore, the Durbin–Watson coefficients ranged from 1.55 to 2.09 for all the analyses across the five experiments, suggesting that the independence assumption for the generalized linear model (GLM) was satisfied (see Durbin & Watson, 1950, 1971; Garson, 2012). Nevertheless, given the nested nature of the study design, we re-ran the regressions testing all hypotheses using hierarchical linear modeling (HLM; random intercepts models). The conclu￾sions based on HLM remained unchanged from conclusions based on the GLM approach. Manipulation check. Behavioral synchrony was manipulated successfully. Analyses of perceptions of synchrony as a function of behavior and participant role indicated that participants perceived a higher degree of synchrony in the synchronous exercise condi￾tion (M 3.91, SD 0.94) than in the asynchronous exercise condition (M 3.17, SD 1.10), F(1, 142) 19.11, p  .001, 2 .12, and that this difference did not depend on whether participants were actors (3.84 vs. 3.06), F(1, 142) 11.41, p .001, or observers (4.00 vs. 3.30, F(1, 142) 7.99, p  .01), F(1, 142) .06, p  .81. That said, there are undoubtedly individual differences in the degree to which actors were successful in engaging in synchronous behavior. Data pertaining to these differences are unavailable. However, our conceptualization pertains to individuals’ percep￾tions of their behavior and not their actual behavior. The fact that observers, like actors, perceived differences in the behavior in the two conditions indicates that the manipulation was successful. Conformity. Participants’ conformity to others’ product choices is shown in Table 1 as a function of behavioral synchrony and participant role. Analyses of these data yielded a significant interaction of these variables, F(1, 142) 10.87, p .001, 2 .07. That is, actors copied others’ preferences to a greater extent in synchronous conditions (M 10.88, SD 2.34) than in asyn￾chronous conditions (M 9.77, SD 1.90), F(1, 142) 4.88, p  .05, whereas observers conformed less in the former condi￾tions (M 9.47, SD 1.97) than in the latter (M 10.60, SD 1.77), F(1, 142) 5.11, p  .05. Supplementary analyses indi￾cated that actors’ conformity was positively correlated with their perceptions of the degree to which they did the exercises synchro￾nously (r .17, p  .10), whereas observers’ conformity was negatively correlated with their perceptions that the actors behaved synchronously (r .14, p  .10). The fact that these correlations Table 1 Actors’ and Observers’ Conformity as a Function of Behavioral Synchrony: Experiment 1 Role Behavior Synchronous Asynchronous Actor 10.88 (2.34)a 9.77 (1.90)b,c Observer 9.47 (1.97)c 10.60 (1.77)a,b Note. Cells with unlike subscripts differ at p  .05. Standard deviations are indicated in parentheses. 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. 64 DONG, DAI, AND WYER