Dark Motives and Elective Use of Brainteaser Interview Questions

The question

Brainteaser interview questions, such as “How many golf balls fit in a school bus” or “Why are manhole covers round,” have no track record of predicting job performance and are known to unsettle candidates. So why do some interviewers keep reaching for them? Highhouse, Nye, and Zhang set out to find whether the answer is a defensible business reason or a trait of the interviewer.

The method

The paper reports two studies. Study 1 presented a representative sample of working adults (n = 736) with a mix of traditional, behavioural, and brainteaser interview questions and used multiple regression, controlling for interviewing experience and sex, to see which personality traits predicted rating brainteasers as appropriate to use [1]. Study 2 followed up with MTurk workers who held final hiring authority, across a three-wave longitudinal design spanning more than a year, to test whether a lack of perspective-taking explained why callous interviewers rated brainteasers as more helpful [2].

The findings

Narcissism and sadism, not interviewing experience or job-relevant judgment, explained the likelihood that a respondent would rate brainteaser questions as appropriate to use in an interview, holding after controlling for interviewing experience and sex [2].

A bifactor analysis showed the two traits share a general callousness factor. Once that shared factor was accounted for, neither narcissism nor sadism added anything further: “narcissism and sadism do not provide significant incremental validity beyond the general factor for predicting the perceived appropriateness of brainteaser interview questions” [3]. In the second study, the trait that broke the pattern was empathy, not experience: interviewers who were better able to see the situation from the candidate’s point of view were more likely to call brainteasers abusive and less likely to call them helpful [2]. Even so, the whole model was a modest predictor. All the personality and attitude variables together explained only 13 percent of the variance in how appropriate people rated brainteasers to be [3], so most of what drives an individual interviewer’s taste for them remains unexplained.

The limits

This paper does not itself test whether brainteasers predict job performance. Its opening framing of them as lacking “evidence for validity” rests on prior literature and Google’s own internal hiring data, not on new data collected here. Both samples are US Amazon Mechanical Turk workers, not a field sample of practicing recruiters, and the effect explained by dark traits is correlational and modest (13 percent of variance): enough to identify a real pattern in who reaches for brainteasers, not enough to say every interviewer who asks one is high in narcissism or sadism.

• A team member wants to add “creative” oddball questions to the interview loop: without this paper, the instinct is to treat it as harmless color. The finding reframes it — the questions carry no demonstrated business justification, and their use tracks the asker’s own callousness more than any signal about the candidate. This bears directly on Running Interviews That Predict Performance: a structured, validated question set is the alternative these findings argue for.
• You’re building or auditing an interview loop and want to know what to cut: the paper gives you a concrete standard — if a question exists to make the interviewer look clever rather than to assess something job-relevant, it fails the test this study applies to brainteasers.
• An interviewer defends a brainteaser as revealing “how someone thinks on their feet”: the paper’s own data shows the people most drawn to these questions score higher on narcissism and sadism and lower on perspective-taking — the confidence in the method is itself part of what the study measured, not independent evidence the method works.