Development and verification of a gender-role stereotyping index for physical activities

The purpose of this study was to develop a valid and reliable instrument to measure the degree to which parents, teachers, and children (4-1/2 to 7) label physical activities according to gender. The research design involved two phases: (a) development/validation of the instrument, and (b) reliability/stability assessment.

Participants for the development/validity assessment were 12 experts in the area of motor and psychosocial development of children who were divided into two groups of six judges (three females, three males). Both groups of judges rated 45 research-based, gender-specific and gender-neutral physical activities according to the degree to which each activity involved each of 14 (nine male, five female) research-based gender-specific physical activity characteristics. Based on the ratings of the Group I judges, 30 physical activities, representative of 10 male, 10 female, and 10 gender-neutral physical activities, were identified and combined to make up the preliminary questionnaire.

Minor modifications were made in the preliminary questionnaire due to unfamiliarity with items and inconsistent response patterns. This modification reduced the final instrument to 24 items, representative of eight male, eight female, and eight gender-neutral physical activities.

In order to assess the construct validity of the instrument, two steps were taken. First two Pearson product-moment correlations were calculated comparing the ratings of the two groups of judges for the total list of activities for: (a) male activity characteristics (r = .98, p < .01) and female activity characteristics (r r = .82, p < .01). Second, six one-way ANOVA's were calculated to determine if there were a significant difference between the two groups of judges on their ratings of the underlying constructs of gender-related activity characteristics for the three categories of physical activities. The results substantiated construct validity as the two sets of judges were not significantly different (p > .05) on any construct.

To assess the reliability/stability of the instrument, ninety participants from each of the three population groups responded to the questionnaire on two occasions, three days apart. Reliability coefficients were calculated using the Pearson product-moment method and descriptive statistics were obtained for each population group. Results indicated that the instrument has high reliability/stability (r = .94 and r = .95, p = .0001) for the parents and teachers respectively, and moderate but acceptable reliability/stability (r = .77, p = .0001) for children between the ages of 4-1/2 and 7.