Abstract
We report on an experiment that utilizes an application-files design to recreate several features of a hiring decision for junior-engineer positions. The critical situation under study involved an assessor (either male or female) and a male and a female candidate with highly similar professional qualifications but different self-presentation styles. We investigate effects from sex category of assessor, sex category and self-presentation of pair of candidates, and type of assessment decision (namely, choice between the two applicants, and ratings of their competence and suitability). Our hypotheses consider both gender as status and as social identity, and predict different outcomes depending on assessment decision. We found no evidence of bias against the female applicant. The question about competence elicited effects from self-presentation only, from both men and women; the choice of applicant and suitability measures show no bias against the female candidate by men and a bias in her favor by women.
Introduction
The formal assessment of job applicants and their qualifications is a ubiquitous feature of many contemporary societies. Although such contexts usually have explicit rules mandating the fair treatment of all applicants, it is often the case that biases (such as those based on gender and ethnicity) persist nonetheless. Under some circumstances, however, the effects from those biases can, in turn, be modified and sometimes even blocked by other factors, such as level of performance. Here we examine the relative roles of four variables in a hiring decision concerning a professional job: the sex category of the applicants and of their assessors, the applicants’ self-presentation style, and the type of assessment decision.1 We also discuss the ability of our method to elicit the participants’ gender beliefs (rather than, for example, socially desirable answers).
7. Conclusions
In this study we proposed alternative, scope-bound hypotheses on the effects of sex of applicant and self-presentation style on assessments of competence. Our sample included men and women, and we also investigated effects on applicant choice and suitability. Results indicate that, for the three measures, gender was not treated as an activated status characteristic whereby men hold the superior state. This was found in spite of a setting and instructions that allowed for status generalization to occur. As we report in our literature review, other recent studies have also shown no sex-of-performer effects in comparable task-settings. Our study thus highlights the importance of carrying out manipulation checks on gender-as-status in research in this area.
We therefore assessed the hypotheses on effects from self-presentation alone (i.e., not as a gender-based status cue) on competence, and found these effects to be as predicted: regardless of sex category of applicant, men and women considered the self-confident person to be the more competent of the two applicants. It is worth noting that these effects emerged even though the differences in self-presentation between the “confident” and the “neutral” candidates were not large, and these levels were conveyed on paper only, through a brief letter. The choice and suitability results revealed other considerations, as expected. Both measures showed effects from sex of subject whereby the women preferred the female candidate regardless of self-presentation, while the men responded more in line with that factor. We interpret these results in terms of the conditional links we had proposed, in the hypotheses section, between social status and social identity.
We should also note that our results showing no bias in favor of the male applicant across conditions and measures were obtained under specific circumstances. The following are worth noting. (1) We studied the case of no explicit relevance between sex category and task, which constitutes a stricter test of gender effects. (2) Our sample consisted of undergraduates at a research university. In our experience, they are less likely to treat gender as a cue to competence, and to define jobs in gender terms. (3) Our participants were not employees holding jobs that involved the hiring of others. For such employees, hiring decisions have consequences for their own employment. That was clearly not the case for our subjects. All three factors likely contributed to the lack of observed bias against women. In the future, it would be worthwhile to systematically vary these factors over a series of studies.
More generally, it is important to view our study in the context of the following considerations. (a) Gender effects (as well as other status effects) do not emerge in all situations, but are dependent on gender beliefs and the circumstances of their activation. This study illustrates that those circumstances also include the type of question asked. (b) As mentioned earlier, there are now several studies showing either a reduced status value for gender, or even no gender effects at all. It is important to realize that, at least for some contemporary populations and settings, the social climate has been changing towards equality regarding views of men’s and women’s competence. This is particularly the case in experiments with university students as subjects, where recent work reveals that decisions do not always favor men. Our study contributes to a fuller understanding of the operation and expression of gender biases.
