Others' Views Task. Each item was scored from 0 (definitely a female thing), through to 6 (definitely a male thing). In order that higher scores on both the male and female subscales would represent greater implicit stereotypic knowledge, the female subscales items were scored in the reverse. This scoring procedure was also employed for the YrVws, however, in this, higher scores were interpreted as representing higher levels of sexism.
Word Association Task. In response latency paradigms, such as this one, responses occurring before a certain point are considered to reflect factors other than aspects of cognitive processing that they intend to measure (e.g., anticipation, mechanical errors). According to patterns of response latency there is a natural minimum value that can be taken by an outlying score. When this is not accounted for the distribution of scores tends to be negatively skewed. As such the program employed to measure response latencies in this study (dmaster), contains a device that eliminates response latency recordings that fall below 200ms.
In this task, participants were exposed to 108 prime-target pairs overall. The design of the task was a 3 (category) x 3 (trait) design (Table 4), which resulted in nine conditions, each containing 12 pairs. In order to analyse the response latencies (measured in ms.) of the category-trait pairs according to condition, latencies for correct responses were averaged across the 12 pairs in each set.
The proportion of errors for each of the nine conditions was calculated by averaging the number of incorrect responses across the 12 pairs. These averages were then converted to represent the proportion of errors obtained in each of the nine conditions.
Sentence Completion Task. For this task, a sentence was scored to be explanatory in nature if the explanation provided by the participant directly related to the behaviour described in the sentence beginning (von Hippel, 1997). In a scoring procedure demonstrated by von Hippel (personal communication, June 18, 1998) a measure of the tendency to explain stereotype incongruent behaviours, was obtained by subtracting the number of explanations provided for stereotype congruent behaviours from the number provided for stereotype incongruent behaviours. Hence, as there were six stereotype-incongruent and six stereotype-congruent sentences in each task, the range of possible scores was -6 to +6, where a positive score indicated that the participants possessed a form of implicit knowledge.
Ambivalent Sexism Inventory. Each item on this scale was scored from 0 (strongly disagree) through to 5 (strongly agree). There were, however, six items that were scored in the reverse. The higher the score on the ASI the more prejudiced or sexist participants' attitudes were interpreted to be, and the lower the score, the more egalitarian.
Women in Society Questionnaire. The WSQ was scored similarly to the ASI, although in this case two items required reverse scoring. As with the ASI, higher scores were understood to indicate more sexist attitudes, and lower scores more egalitarian.
Competence-related behaviour. The rejection of influence rate was calculated as the proportion of trails in which the participant stayed with his response compared to the number of trials in which a response was made. Responses to the post-experimental ability-related questions were measured through the use of an anchored line scale, in which responses could range from 0 to 60. In the particular item of interest, the higher the score, the higher the perceived competence of self relative to other.