Judgments about sexism: A policy capturing approach

Different definitions of sexism have been used by different experimenters a nd little attention has been given To the possibility that participants' de finitions of sexism might vary across individuals and across situations, ju st as the definitions of investigators vary. Judgments about sexism are ine xtricably based on individual perceptions, values, and beliefs. It would se em reasonable, therefore to examine individual conceptions of what constitu tes sexism. A total of 95 primarily middle-class Caucasian participants (68 females, 27 males) participated in three studies investigating what inform ation people use in deciding whether or not a male actor is sexist. Six spe cific categories of information (or cues) were examined: (I) comments (or l ack thereof) by the actor on a female target's physical appearance, (2) unw anted or inappropriate physical contact (or lack thereof), (3) assumptions (or lack thereof) about the target's work, personality, etc., based on her gender, (4) the target being interrupted (or not) by the actor, (5) an appa rent power differential (or lack thereof) between the actor and the target, and (6) assumptions by the actor about appropriate gender role behaviors. A policy capturing methodology was employed in which three different sets o f scenarios, each describing a number of interactions (called profiles) bet ween a target and an actor, were presented to male and female participants who rated the degree of sexism exhibited by the actor in each scenario. Pol icy capturing represents an ideographic approach to research in which the p rimary focus is on establishing statistical parameter estimates that descri be each individual's behavior in a variety of environmental situations. Onl y when that is accomplished does the researcher examine the possible nomoth etic aggregations across participants. Thus many fewer participants are req uired than in traditional nomothetic approaches. Each profile set used a di fferent subset of the ales which contained either subtle (Study I), overt ( Study 2), or very overt (Study 3) levels of the cues. Multiple regression a nalysis revealed that when relatively overt acts of sexism were described m ost participants had reliable policies; that is, their judgments were predi ctable from the cues. However, when relatively subtle acts of sexism were d escribed most participants did not have reliable policies; that is, their j udgments were nor very predictable from the cues. There were large individu al differences in how participants weighted the importance of various cues, especially with subtle cue levels, and in participants' mean sexism rating s.