What Actually Predicts Job Performance

Faced with a resume, a recruiter almost always looks at the same three lines first: the degree obtained, the title of the last position, and the cumulative years of experience. These are, almost exactly, the three signals that personnel psychology research ranks among the least predictive of future performance.

A Meta-Analysis Covering Over 32,000 Employees

This is the central finding of a 1998 meta-analysis by Frank Schmidt and John Hunter, published in Psychological Bulletin, which aggregates data from more than 32,000 employees across 515 different civilian jobs to compare the predictive validity of 19 selection methods used in hiring (Schmidt & Hunter, 1998, Psychological Bulletin, 124(2), 262-274). More than twenty-five years after its publication, the study remains one of the most cited references in the field — enough to have been updated in 2016 by Schmidt, Oh, and Shaffer, without overturning its central conclusion.

What Predicts — And What Doesn't

The most predictive combinations pair a measure of general cognitive ability with a work-sample test or a structured interview: according to the authors, these achieve the highest multivariate validity coefficients among all 19 methods compared.

At the other end of the ranking are precisely the signals a resume highlights first: years of professional experience and level of education, among the weakest predictors in the study.

What This Means for a Hiring Process

This finding is far from anecdotal for how a hiring process is designed. Screening that relies mainly on these three elements — degree, job titles, tenure — prioritizes the least robust signals in the entire dataset studied. The most predictive signals, by contrast, require a structured, comparable evaluation across candidates: they can't be read off a resume, they have to be built into the process.

This makes the case for documenting, consistently, the criteria observed at each stage — rather than leaving each evaluator to reconstruct their own judgment from scattered pieces of information.

A Necessary Caveat

One caveat is necessary, however: a meta-analysis of this scale synthesizes aggregated trends across tens of thousands of jobs and sectors. It guarantees the prediction of no individual case, and doesn't eliminate the uncertainty inherent to any human decision. Structuring doesn't mean mechanizing the decision: it means giving the recruiter a reliable basis for comparison, on which they continue to exercise their own judgment.

Clarentia's Position

That's what our agents aim to document at Clarentia — never making the call in the recruiter's place.

Cited Source

Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124(2), 262-274.