As a result of the numerous significant correlations between the AHBS and the CAS-A subscales, stepwise multiple regression was used to determine unique variance employing the four subscales of the AHBS as the criteria and the six subscales of the CAS-A as predictors.
Table 3 presents the correlation of AHBS subscales with sexual behaviours.
On the negative side, the AHBS was not significantly related to self-reported sexual behaviours and the subscales demonstrated poor internal consistency.
In the present study, scores on the AHBS were not influenced by administration methodology.
Support for discriminant validity was provided by the general lack of a significant correlation between AHBS and social desirability (BIDR scores).
In the present study, convergent validity of the AHBS was supported by the substantial correspondence found between the CAS-A and the AHBS.
Despite support for the convergent validity of the AHBS, the estimate of inter-item homogeneity for the AHBS subscales was cause for concern.
The possibility that inter-item homogeneity is related to race or marital status should be explored in future research on the AHBS.
In terms of gender differences, Zagumny and Brady (1998) report that women had significantly higher total AHBS scores in comparison to men; however, AHBS subscale scores were not reported by gender.
Of the 30 correlations and three t-tests performed examining the relationship between AHBS scores and sexual behaviours, only one was significant.
The AHBS appears to be measuring constructs related to sexual attitudes and beliefs, which in the current study were not related to reported sexual practices.
The results of this study suggest the need for caution in the use of the AHBS in clinical settings.