Psychometrics of the Heterosexist Attitudes in Sport—Lesbian Scale

Background/Purpose: Homophobia in women’s athletics has been investigated using qualitative methodology (e.g., Blinde & Taub, 1992; Kauer & Krane, 2006; Knight & Guiliano, 2003; Krane 1996, 1997; Krane & Barber, 2003; Lenskyj, 1991, 1997).  Mullin (2013) developed the Heterosexist Attitudes in Sport – Lesbian (HAS-L) scale, which demonstrated adequate construct validity and internal consistency. The purpose of the questionnaire is to quantify heterosexist attitudes within the women’s collegiate athletic environment. The scale consists of four subscales: Cognitive/Affective (C/A), Language Behaviors (LB), Inclusion Behaviors (IB), and Avoidance of the Lesbian Label (ALL). Presently, the purpose was to determine if the HAS-L demonstrated evidence of concurrent and convergent validity and test-retest reliability in the same population.

Method: Female collegiate athletes (N = 225) completed the following:  HAS-L,  Attitudes toward Lesbians and Gay Males Short Form (ATLG; Herek & McLemore, 2011),  Religious Attitude Inventory (RAI; Ausbel & Schpoont, 1957),  Attitudes towards Women Scale (ATWS; Spence & Helmrich, 1973),  Marlowe-Crowne Social Desirability Scale (MC-SDS; Crowne & Marlowe, 1960), and  demographic questionnaire. Participants completed the HAS-L again 4 weeks later.  Correlational analyses were conducted among the HAS-L and the ATLG, the RAI, the ATWS, the MC-SDS, and items from the demographic questionnaire. Intraclass correlation coefficients (ICCs) were measured between responses to the first and second session of completing the HAS-L. 

Analysis/Results: All HAS-L subscales were significantly correlated to the 2 subscales of the ATLG (r = .17 - .67, p < .05). The C/A subscale was significantly correlated with the RAI (r = -.33, p < .01). All HAS-L subscales were significantly correlated with the ATWS (r = -.20 –  -.30, p  < .01). The LB subscale was significantly correlated with the MC-SDS (r = -.20, p = .002). At least one of the subscales of the HAS-L was also significantly related to age, number of known LGBT individuals, political affiliation, and/or sexual orientation. The subscales of the HAS-L had moderate to strong test-retest reliability (ICC = .64-.91). 

Conclusions: The HAS-L demonstrated strong evidence of concurrent validity.  Subscales were significantly correlated with the appropriate questionnaires, suggesting adequate evidence of convergent validity with other measures of homophobia/heterosexism correlates. The HAS-L also demonstrated adequate temporal stability. With additional evidence of reliability and validity, the HAS-L is a useful tool for researchers and practitioners to quantitatively examine the potential psychological, performance, or developmental ramifications of heterosexism in women’s collegiate athletics.