When and how does anxiety influence interview outcomes? The moderating role of gender, interviewer anxiety and impression management.

Contribution/Originality: Test anxiety has received limited attention in personnel selection research, although it is well known that anxiety can impair selection outcomes (McCarthy & Goffin, 2004; Proost, Derous, Schreurs, Hagtvet, & De Witte, 2008). This paper focuses on the selection interview and studies when and why anxious applicants are less attractive to employers. As such, this topic fits into the topic of ‘biases in selection’ of the SGM. We build on studies on gender role stereotypes and the related backlash effect (Rudman & Glick, 2001) to suggest that the negative influence of anxiety on interview performance differs for male and female candidates. Since anxiety fits more with the female gender role and is more contrastereotypical for men (Moss-Racusin, Phelan, & Rudman, 2010), we expect that anxious men are the least attractive, compared to anxious women and non-anxious applicants. Design: To test this assumption, we conducted a 2 (anxiety) x 2 (gender) randomized between subjects experiment on 199 respondents (107 men, aged 18 to 65 years). Participants received a job description of a Business Project Manager and were instructed to hire a candidate for this job. They further received a written transcript of a selection interview and were asked to indicate the job suitability of that candidate (i.e., Would you invite this candidate for a second interview; Would you offer the candidate this job?; would you reject this candidate (reverse scored), alpha = .88). Anxiety was manipulated in two ways: participants received the result of a personality questionnaire, stating that the candidate appeared anxious in social settings vs. that the candidate is self-confident in social settings and read a written transcript of the interview in which notifications of the non-verbal behavior of the candidate were added in terms of ‘the candidate appears calm’ vs. ‘the candidate appears somewhat nervous’. Results: The results of a PROCESS analysis (model 1) showed a main effect of anxiety on interview outcomes, B = -1.10, SE = .12, p = .00, and a significant interaction effect between gender and anxiety, B = .37, SE = .17, p = .03. As expected, the negative effect of anxiety on interview outcome was stronger for the male applicant compared to the female applicant. Limitations and Future research: A second study will be conducted in the spring in which we test two follow-up hypotheses. First, we test whether anxiety of the recruiter influences the relationship between anxiety and interview outcome. We suggest that when the recruiter experiences a similarity with the candidate in terms of anxiety, this will positively influence the interview outcome. Second, we test whether self-promotion can reduce the negative effect of anxiety on interview outcome and whether this effect is especially positive for the anxious male candidate. We base our suggestion on the idea that self-promotion is stereotypical male behavior and may thus especially advantage the (anxious) male candidate. Implications: This study has important implications for the validity of selection interviews and adds to the debate on the subjectivity of personnel selection decisions.