Personality questionnaires in recruitment: what the research shows

Personality questionnaires divide opinion like no other selection tool: dismissed as horoscopes by some, treated as X-rays by others. The research supports neither view. Used well, they add real predictive value; used badly, they measure little more than a candidate's theory of what the employer wants to hear.

What personality actually predicts

The evidence base rests on the Big Five model — conscientiousness, emotional stability, extraversion, agreeableness and openness. Barrick and Mount's landmark meta-analysis established that conscientiousness predicts job performance across essentially all occupations and criteria, a finding replicated for three decades.^[1] Schmidt and Hunter's review of 85 years of selection research placed conscientiousness and the closely related integrity tests among the predictors that add meaningful validity beyond ability measures.^[2] The most recent authoritative re-analysis, by Sackett and colleagues, revised most validity estimates downward under more conservative statistical corrections but confirmed the ranking: personality measures — conscientiousness in particular — retain incremental validity over cognitive ability.^[3]

Two qualifications matter. First, the traits that matter vary by role: conscientiousness generalises, but extraversion predicts in sales and management, not everywhere.^[1] Second, the instrument matters. Sackett and colleagues found that contextualised personality measures — items framed as "at work" rather than "in general" — show meaningfully higher validity than decontextualised ones.^[3] Type-based instruments are a different story: the MBTI's own publisher states it should not be used for selection, and a National Academy of Sciences review found no adequate research supporting such use.^[4]

The faking problem — real, but manageable

The standard objection to personality questionnaires is that candidates fake. The research says the objection is partly right: there is compelling evidence that a substantial share of applicants distort responses in their favour, and that faking reduces validity — faked profiles correlate less with how peers describe the same person.^[5] But the research also shows the distortion is smaller and less strategic than assumed: qualitative work with real applicants finds most strive for a good result while remaining recognisably themselves, and even instructed fakers tend to produce inflated versions of their true profile rather than a different one.^[6]

Design choices manage the rest. Warnings that responses will be checked reduce faking without damaging how candidates feel about the process.^[7] Forced-choice (ipsative) formats constrain impression management.^[6] And scoring at the scale level rather than reacting to single answers means no hiring decision hinges on one flattering response.^[8]

The candidate experience question

Meta-analytic work on applicant reactions finds personality questionnaires rate less favourably than interviews and work samples on perceived job relevance and opportunity to perform.^[9] That matters commercially: candidate reactions predict whether people accept offers and recommend the employer.^[6] The remedies the research supports are procedural — explain why the questionnaire is used, make its relevance to the role visible, and offer feedback. Candidates who receive a feedback conversation report markedly more positive experiences.^[6]

How to use them well

The research points to a consistent recipe. Use trait-based (Big Five) instruments, not type indicators. Weight the traits the role evidence supports rather than a generic profile. Treat personality as one signal in a battery — its value is incremental, alongside ability, skills and structured interviews, not instead of them.^[2,3] And never use personality scores as a knockout: they describe tendencies, not competence.

Where personality questionnaires earn their place is breadth. Ability tests and skills tests predict whether someone can do the work; personality predicts the how — persistence, reliability, teamwork — including outcomes like counterproductive behaviour that no skills test touches.^[2] That is a different question, and in most hiring decisions it is worth asking.