Structured Policy Analysis
Do Structured Interviews Actually Predict Job Performance?
A narrow look at what an interview score is worth, why the famous numbers fell in 2022, and what 'structure' actually changes.
Key Findings
The claim that an interview predicts job performance as well as an IQ test came from numbers that were revised downward in 2022 . The revision was statistical: a widely used correction for range restriction had inflated nearly every estimate, and when it was removed cognitive ability fell faster than interviews did . Structure still helps, but research suggests it works mainly by removing interviewer error rather than by adding new insight . Meanwhile unstructured interviewing tends to raise confidence more than accuracy .
These are corrected meta-analytic averages, not guarantees for any one hire. Validity coefficients describe correlations across many people; they say little about whether a specific candidate will succeed. Estimates still vary by job, criterion, and study design.
The famous parity stat rested on a math artifact
Schmidt and Hunter reported structured interviews and cognitive ability near r = .51. After correcting the range-restriction overcorrection, structured interviews held near .42 while cognitive ability fell to about .31, so the interview 'won' partly because the comparison fell.
Structure works by subtraction, not addition
Standardizing questions and rating scales raises reliability and shrinks idiosyncratic interviewer error. Several lines of evidence suggest the validity gain comes from that error reduction more than from extracting deeper signal.
Unstructured interviews can feel useful while hurting accuracy
People form confident impressions even from random answers, and added non-diagnostic talk can dilute valid information. Confidence rises even when predictive accuracy does not.
Even a strong interview explains a minority of the variance
An operational validity near .42 corresponds to roughly 18 percent of variance in performance. That leaves most of the outcome to factors an interview does not capture.
Research Findings
Sources
What this means in practice
Work related to structured hiring often involves repetitive manual tasks: writing the same job-relevant questions for every candidate, transcribing answers, applying a fixed rating scale, and combining scores by a consistent rule instead of a hallway impression. These are typically handled with systems that capture each interviewer's ratings in one place, apply the scoring rubric the same way every time, and roll the numbers up into a comparable summary across candidates.
- Ingest interview notes, rating-scale scores, and candidate records from forms or spreadsheets into one structured place.
- Automate the scoring rule: apply the same rubric and weighting to every candidate and flag missing or inconsistent ratings.
- Generate side-by-side candidate summaries and reports that show how each score was combined rather than relying on memory.