| Data collection method: |
Laboratory experiments and questionnaire data with human participants, using individual participants as respondents. Study 1a developed the Self-Other Intergroup Anxiety Scale (SOIAS). Participants' scores on the SOIAS were key independent variables in Studies 2 to 5. The research team developed a diverse pool of 85 pilot questions, which 495 participants answered using a 7-point scale from -3 ('Strongly disagree') to +3 ('Strongly agree'). Study 1b confirmed the factor structure and tested for discriminant and convergent validity by administering the SOIAS questionnaire to 198 participants in mass testing sessions. Study 2 contained assessments of physiological reactivity including electrodermal activity (skin conductance), measures of cardiac performance and facial electromyogram (EMG). Fifty-one participants were recruited for the Study. They were randomly assigned to one of two experimental conditions: participants were told that they would shortly meet a person to work together on a task, and that this person had a history of schizophrenia (experimental condition) or average musical reactivity (control condition). Study 3 included behavioural data from an interaction with a confederate who participants were led to believe had a history of schizophrenia. Sixty-one female undergraduate students took part in a design that was implemented in four stages. First, a pretest included measures of the motivation to control prejudice (Plant & Devine, 1998) and the two measures of intergroup anxiety (the SOIAS and a conventional unidimensional measure). Second, during a preinteraction phase, participants were told that they would meet and work on a task with another person. Participants were told that this person had a history of schizophrenia (experimental condition) or an unspecified personality type (control condition). Participants were briefly introduced to the confederate at this point to increase the face validity of the anticipated interaction. Physiological responses, self-reported anxiety (including the SOIAS), and expectancies were recorded at this point. Third, there was a structured interaction with the confederate (who was blind to experimental condition) and this interaction was unobtrusively recorded for later behavioural coding by independent judges. Finally, in the post-interaction phase, participants and confederates completed attitude measures about each other. Study 4 was designed to test the effectiveness of a well-known technique to reduce the effects of implicit or automatic prejudice. The compunction intervention (Devine, Monteith, Zuwerink, Elliot, 1991; Monteith, 1993) involves making participants aware of situations when they have behaved in ways that are not consistent with their personal standards against prejudice. Sixty-five participants completed a novel Single Category Implicit Attitude Test (Karpinski & Steinman, 2006) towards people with mental health difficulties. Half were given false feedback that they were high in prejudice towards people with mental health difficulties (high prejudice feedback condition). Half had feedback that they were low in prejudice towards people with mental health difficulties (low prejudice feedback condition). Participants were then told that they would meet someone who had a history of schizophrenia (in reality a confederate who was blind to condition). The paradigm was very similar to that of Study 3, with a pretest, a preinteraction phase, a structured interaction, and a post interaction phase. We used a very similar battery of dependent measures (physiological, behavioural, and self-reported), with an additional follow-up telephone interview to explore the effects of the intervention in the longer term. Study 5 tested a promotion intervention (which encourages participants to approach an anticipated future interaction as an opportunity to have an enjoyable dialogue). The design of Study 5 was virtually identical to Studies 3 and 4 apart from the manipulation. Sixty-four participants were told that they would shortly meet a person with a history of schizophrenia. Half the participants were told that this interaction was an opportunity to relax and have an enjoyable dialogue (experimental condition), and the other half were given no further instructions (control condition). We used the same range of pretest and dependent measures as in Study 4. In addition, participants completed a Stroop task as a measure of resource depletion. |
