Experiment 1 

Method 

Subjects and design. Twenty-four male and female undergraduates at the University of Wales College of Cardiff [...].

The study had a 2 (task instruction: stereotype-suppression or control) X 2 (construction: Passage 1 or Passage 2) mixed design with repeated measures on the second factor. 

Procedure and stimulus materials. Each subject [..] was [...] randomly assigned to one of the experimental conditions. 

The experimenter then explained that the study comprised an investigation of people's ability to construct life event details from visual information. All subjects were shown a color photograph of a male skinhead. It was explained that [..] subjects would be given 5 min to compose a brief passage describing a typical day in the life of the target. Before performing this task, half of the subjects were informed that previous psychological research has established that our impressions and evaluations of others are consistently biased by stereotypic preconceptions. In the present task, however, it was stressed that subjects should actively try to avoid thinking about the target in such a manner. [..] The control subjects, in contrast, were given no such instruction before constructing the passage. [...]

After they completed the passage, subjects were shown a color photograph of a different male skinhead. They were then all requested to spend another 5 min constructing a story describing a day in the life of this person. [...] In this case, none of the subjects received any admonishment about the avoidance of stereotyping. [...]

Dependent measure. The dependent variable in this experiment was the level of stereotypicality of the constructed passages.

To compute this measure, two independent raters read each passage and, using a 9-point rating scale [...] gave an estimate of the stereotypicality of its contents. These raters were blind to experimental condition and the purpose of the study. Given the high level of agreement in their estimates [...] scores were collapsed and a single measure of the stereotypicality of each passage was calculated.

Experiment 2 

Method 

Subjects and design. Twenty-four male and female undergraduates at the University of Wales College of Cardiff [...].

The study had a single-factor (task instruction: stereotype suppression or control) between-subjects design. 

Procedure and stimulus materials. [...] Each subject [...] was greeted by a female experimenter and randomly assigned to one of the treatment conditions.

The procedure, at this stage, was identical to that of Experiment 1. [...]

Following their completion of the passage, the experimenter explained that she would take the subject next door to meet the person [...] who was depicted in the photograph. On entering the adjacent room, however, the subject was confronted not by the skinhead, but by a row of eight empty seats. A denim jacket and bag were placed on the first seat and the experimenter explained that these items belonged to the person. She then remarked that he must have popped out of the room for a few minutes [...]. Until he reappeared, each subject was instructed to occupy one of the vacant seats [...] and wait for his return. It was anticipated that subjects' seating preference would be affected by the earlier passage completion task. That is, to the extent that stereotype inhibition promotes the resurgence of stereotypical thinking, we expected stereotype suppressors, rather than subjects in the control condition, to maintain a larger social distance from the target.

[...] Once each subject had selected a seat, the seat number was noted and the experimenter brought the study to an end. [...]

​​Dependent measures. Two dependent measures were used in this study. 

First, we considered the stereotypicality of the contents of the constructed passages. [...] Scores were again collapsed to produce a single measure reflecting the stereotypicality of each passage. Second, to assess the behavioral expression of rebounded stereotypic beliefs, we noted each subject's choice of seating position. [...] As Seat 1 was located next to the target's belongings, successive seat numbers reflect an increase in the social distance between the subject and the target.

Experiment 3 [...]

Method 

Subjects and design. Twenty-four male and female undergraduates at the University of Wales College of Cardiff [...].

The study had a 2 (task instruction: stereotype suppression or control) X 2 (word type: critical or distracter) mixed design with repeated measures on the second factor.

Procedure and stimulus materials. [...] Each subject [...] was [...] randomly assigned to one of the treatment conditions.

At this stage, the experiment was identical to Experiments 1 and 2. [...]

Following the completion of the passage, each subject was seated facing the screen of an Apple Macintosh Microcomputer [...] and requested to perform another [...] task, specifically a lexical decision task. On the presentation of a letter string in the center of the computer screen, subjects had to judge, by means of a key press, whether it was a "word" or "nonword." The list of to-be-identified letter strings comprised 14 words that were stereotypic of skinheads, 14 matched distracters, and 28 nonwords. [...] The distracters [...] were irrelevant with respect to the category of skinhead. Presentation of the letter strings was randomized, for each subject, by the computer. Lexical decision times [...] were recorded by the computer and measured using the Macintosh's Extended Time Manager. It was anticipated that residual action potential, when present, would facilitate subjects' responses to the skinhead-related words. [...] If the stereotype remained accessible, lexical decisions would be faster for the critical than for the distracter words. [...] To provide supplementary baseline data, an additional 12 subjects performed the lexical decision task without previously completing a life event passage. As no stereotype was activated in this condition, facilitatory priming effects were not expected in the lexical decision task. [...]

Dependent measures. As before, we considered the stereotypicality of the contents of the constructed passages.

Two independent raters read each passage and gave an estimate of its stereotypicality [...]. Given the level of interrater reliability in these estimates [...] the data were collapsed to produce a single measure reflecting the stereotypicality of each passage. Lexical decision latencies were taken as an indicator of stereotype accessibility.