Establishing Content Validity of the Heterosexism in Sport-Lesbian Scale

Background/Purpose

While qualitative researchers have established that heterosexism is a serious concern within women's athletics, little is known about the psychological ramifications. Quantitative researchers have unsuccessfully relied on non-sport specific questionnaires to link heterosexism with psychological constructs. The purpose of this study was to establish content validity of the Heterosexism in Sport-Lesbian Scale (HAS-L) using quantitative methodology.

Method

The researcher generated 40 sport-specific items dispersed among the four hypothesized factors of the HAS-L: affective, behavioral, cognitive, and avoidance of the lesbian label. Experts (N = 22) indicated the fit of each item to the factors. Repeated Measures ANOVAs, Aiken's V coefficients, and Estimate Sizes were conducted to determine the factor structure.

Analysis/Results

Three items were removed due to non-significant RM ANOVAS. Items (n = 16) which loaded onto the intended domain were maintained. Of the 18 items which statistically loaded onto two factors, 10 were reworded and 8 were removed. Items (n = 3) which loaded onto a non-keyed domain were moved to the non-keyed domain for the second iteration of the HAS-L. Twenty-nine items were maintained with Aiken's V coefficients ranging from .68-.92 while ES ranged from 0.23-2.87.

Conclusions

Using rigorous statistical analysis, 40 items from the HAS-L were maintained, reworded, or removed to create the second iteration of the HAS-L. Using inferential statistics to reach a consensus among a panel of experts removes some of the subjectivity inherent in the traditional methods of questionnaire development. Future versions of the HAS-L should also undergo factor analyses establish construct validity.