| This thesis describes a new methodology that allows for the collection and analysis of attributes associated with sex role stereotypes. Two studies provide convergent validity of a theory of seven dimensions that underlie sex roles, three of which could be considered biological, social normative, or sex role phenotypes. In the first study, a free-response methodology for collecting the attributes is combined with objective, statistical criteria to select the attributes most strongly associated with typical men, women, gay men, and lesbians. Examinations of the types of people subjects think best represent the typical man, woman, gay man, or lesbian suggest differences in subjects' personal knowledge. A test of the distributions of attributes of real people versus media related people suggests similar stereotypes are associated with both populations. An analysis of the distributions of the frequencies of the attributes becomes the basis for predicted associations by dimension along with suggesting contrast weights to be used in the second study. The second study collected typicality ratings for attributes selected in the first study and uses contrast t-tests to validate their predicted associations. Contrast analyses of these ratings suggest some important differences in the assessment of typical attributes for straight versus gay subjects, and for white subjects versus people of color. |
