The following is an excerpt of the intervention methodology. For more information, please see the full text of the article on the publisher's website or through your institution's library.

Experiment 1 

Method 

Participants. Eighty students [...]  participated in this experiment [...] individually or in small groups.

Design 

The experiment used a 2 (context group: academic vs. political) x 2 (confirmation: confirm vs. disconfirm) x 2 (trait dimension rated: academic vs. political) design with the first two factors between- and the last within participants. Participants formed a stereotype of a focal group of college students who were academically motivated and politically liberal. [...] Participants first learned the stereotype of one of the two other college student groups. Each context group was described on one dimension that was shared with the focal group but of the opposite pole [...]  and on another dimension that was irrelevant to the focal group stereotype. In the academic context condition, members of the context group were described as academically unmotivated and artistic. In the political context condition, members of the context group were described as politically conservative and leaders on campus. In a subsequent phase of the experiment, participants were exposed to additional focal group members who either confirmed or disconfirmed the stereotype that focal group members were academically motivated and liberal.

Procedure 

The experiment involved two phases: a stereotype formation phase and a stereotype change phase. Upon arrival at the laboratory, a female experimenter instructed participants that they would learn about two fraternities [...].

In the stereotype formation phase of the experiment, participants were first asked to form an impression of one of the two context fraternities by reading statements that described one behavior for each of the 24 fraternity members. Participants in the academic context condition read 12 statements that described behaviors that were academically unmotivated [...] and 12 statements that described behaviors that reflected interest in the arts [...]. Participants in the political context condition read 12 statements that described politically conservative behaviors [...] and 12 statements that described leadership behaviors [...]. The 24 statements were presented on a computer screen in random order and participants were able to read through the statements at their own pace.

After learning about one of the two context fraternities, participants were introduced to the second (focal) fraternity. Participants formed an impression of the focal fraternity by reading 24 statements that described behaviors performed by fraternity members. Twelve of these statements described academically motivated behaviors [...] and 12 described politically liberal behaviors [...]. Again, the 24 statements about the focal group were presented in a random order. After learning about both fraternities, participants were allowed as much time as needed to generate descriptions of both groups [...]. The descriptions were subsequently coded to assess the impact of context on focal group stereotypes, prior to presentation of the confirming or disconfirming instances.

In the stereotype change phase of the study, participants were told that they would learn about additional members of the focal fraternity. Participants in the confirm condition then read about 12 additional fraternity members whose behaviors were consistent with the stereotype they had formed in the first phase of the experiment. [...] Participants in the disconfirm condition [...] read [...] about several group members who disconfirmed the group stereotype. [...] The remaining four statements described stereotype-consistent [...] behaviors.

After learning about the additional members of the focal group, participants completed measures of stereotype central tendency and variability of the focal fraternity as well as measures of perceived typicality of the 12 group members learned in the second phase of the experiment.

Dependent Variables 

Initial open-ended descriptions. Prior to receiving the additional fraternity members during the stereotype change phase of the study, participants wrote a few sentences that gave their impressions of the focal fraternity. These impressions were rated by judges to assess the perceived extremity of participants’ group stereotype based on the initial stereotype formation information. After the stereotype change phase of the study, participants completed a set of additional measures to assess their stereotypes of the focal group at that point.

After the stereotype change phase of the study, participants completed a set of additional measures to assess their stereotypes of the focal group at that point.

Percentage estimate task. Participants completed a percentage estimate task in which they were asked to estimate the prevalence of eight attributes among members of the focal fraternity. Two stereotypic and two counterstereotypic items for each dimension of the focal group stereotype were assessed [...].

Mean/range task. The perceived mean and range of the focal group on each attribute used in the percentage estimate task were also assessed. Participants estimated the standing of the average member of the focal fraternity on each attribute by placing an ‘X’ on a 178-mm line that represented the possible range [...] of that attribute. Participants also put marks on the dimension to indicate where the group members who possessed the least and the greatest amounts of the attribute were located. The perceived range of the group was computed by taking the difference in these two ratings.

Within-group similarity ratings. Participants rated how similar members of the focal group were to each other on the academic and political dimensions. They marked their ratings on seven-point scales that ranged from extremely dissimilar to extremely similar.

Typicality ratings. The set of behaviors that were included in the stereotype change phase of the experiment was again presented to participants. Participants rated how typical each group member was of the overall group using a seven-point scale [...].

Experiment 2

Method 

Participants. One hundred and sixty students [...] participated in this experiment [...] individually or in small groups.

Design 

The experiment employed a 2 (context: academic vs. political) x 3 (confirmation: confirming vs. disconfirming vs. no additional information)  x 2 (trait dimension rated: academic vs. political) mixed design, with the first two factors varying between-participants and the last varying within them. [...] Participants learned about the focal group in either an academic or political context. In the stereotype change phase of the experiment, participants were either given no additional information about the focal group [...] or were given stereotype-confirming or stereotype-disconfirming information. Participants then rated the focal group on dimensions of academic achievement and political orientation.

Procedure 

The procedure of this experiment was similar to that used in Experiment 1, with three exceptions. Most significantly, this experiment included a no additional information condition in which participants did not complete the stereotype change phase of the experiment and instead completed the dependent measures (except for the typicality measure), immediately after the stereotype formation phase of the experiment. 

Two other methodological changes were made. First, the groups were identified as ‘student groups on a college campus’ rather than as fraternities to reduce the application of pre-existing stereotypes to the groups. Second, the mean/range task described in the second set of dependent measures used in Experiment 1 was replaced with a histogram task. In this task, participants were asked to generate a perceived frequency distribution of group members along each of the eight attributes used in the first experiment. Participants were presented with five blank bars of equal height that were labeled to reflect a continuum that ranged from very few group members to very many group members. Participants were asked to fill in each bar to the height that would indicate the proportion of group members they believed belonged in that segment of the continuum.