Using the circumplex model to examine the relationship between gender, gender-role identity, and attachment style

The relationship between gender, gender-role identity, and attachment style was examined using the circumplex model. Participants included 490 college students who completed the Bem Inventory, the Relationship Questionnaire, and the Interpersonal Adjective Scales. Primary hypotheses were formulated to explore parallels within the gender-role and attachment literature. Secondary hypotheses were formulated to: (a) explore, by gender, the masculinity effect in gender-role research along the dimensions of dominance and affiliation; (b) differentiate cross-sex gender-role-types (i.e., masculine females and feminine males) from same-sex gender-role-types (i.e., masculine males and feminine females); and, (c) distinguish, by gender, undifferentiated gender-role types from the remaining gender-role types. Circulinear grouped frequency distributions were prepared and 95% confidence intervals were calculated. Polar vectors among hypothesized types were compared using a circulinear test of significance. As predicted, androgynous gender-role and secure attachment types endorsed dominant and affiliative traits with parity. Masculine gender-role types endorsed more dominant traits than remaining gender-role and attachment types. Also as predicted, preoccupied attachment types and feminine gender-role types endorsed more affiliative traits than remaining types. However, contrary to the hypothesis regarding fearful avoidant attachment and undifferentiated gender-role types, these types endorsed a trait pattern similar to preoccupied and feminine types. Dismissing avoidant attachment types endorsed a trait pattern similar to androgynous and secure types. The unusual grouping of dismissing avoidant types with more socially adept androgynous and secure types was attributed to an artifact in assessment methodology. When trait endorsement patterns were examined by gender, it was found that: (a) masculine males endorsed more dominant and fewer affiliative traits than remaining types; (b) masculine females, androgynous males, and androgynous females endorsed both dominant and affiliative traits with parity; (c) and feminine males, feminine females, and undifferentiated females endorsed more affiliative traits and fewer dominant traits than remaining types. Data regarding undifferentiated males was not tested since the maximum likelihood estimators for this group were too low to be reliable. Since gender-role research has traditionally been analyzed using factor analytic methods, a multivariate factorial MANOVA was used concurrently to test all hypotheses regarding gender, gender-role identity, attachment style, and dispositional traits. Results of the MANOVA and follow-up discriminant function tests support circulinear significance test results.