Structured interviews outrank cognitive ability tests as a predictor of job performance. Structure roughly doubles an interview's predictive power and is the cheapest fix available.
Structure beats intuition, and the size of the gap is measurable
Hiring is one of the few management activities where the research is unusually clear and unusually ignored.
Sackett, Zhang, Berry and Lievens published a re-analysis in the Journal of Applied Psychology in 2022 that corrected a systematic overcorrection for range restriction running through decades of earlier selection meta-analyses. The correction changed the rankings. Structured interviews came out at r = .42, above cognitive ability tests at r = .31 — making the structured interview, on this analysis, the strongest single predictor of job performance available.
The older Schmidt and Hunter figures, still widely quoted, put structured interviews at .51 against .38 for unstructured. Both analyses agree on the part that matters in practice: structure roughly doubles the predictive power of an interview, and it is the cheapest improvement available to any hiring process.
Structure means the same questions, in the same order, for every candidate, scored against criteria written before anyone was interviewed. It is not a personality constraint or a bureaucratic imposition. It is the difference between a process that predicts performance and one that predicts how much the interviewer enjoyed the conversation.
That single finding organises this entire pack. Every prompt here exists to move a decision from impression to evidence.
Define the target before you look
Define a Role Scorecard Before You Start Hiring is the load-bearing prompt, and skipping it is what makes everything downstream unreliable. It forces you to state what the person must accomplish in the first year, which competencies actually predict that, and what evidence would demonstrate each one — before any candidate exists to bias the criteria.
The order matters more than it appears. Criteria written after you have met someone impressive tend to describe that person.
Decide Whether You Can Afford to Hire comes even earlier for a small business — total cost against the cash forecast, not against revenue. The business finance pack has the wider context.
Write a Job Description That Attracts the Right Candidates turns the scorecard outward. Its main job is honest filtering: the requirements that are genuinely required, the ones that are preferences, and a real description of the work. Padded requirement lists are the standard way to shrink and homogenise a candidate pool without intending to.
Run a process that produces evidence
Screen Resumes Against a Role Scorecard applies consistent criteria to the stage where inconsistency is least visible and most consequential.
Write a Structured Interview Guide is the direct implementation of the research: the same questions for every candidate, behavioural rather than hypothetical, with follow-ups planned. Hypothetical questions measure how well someone imagines; behavioural questions measure what they have actually done.
Build an Interview Scorecard and Rubric supplies the other half. Structure without a rubric collapses back into impression at the scoring stage — everyone rates a 4 and nobody can say what a 4 means. The rubric describes what each level looks like in terms of observable evidence.
Design a Work-Sample Exercise covers the method that consistently performs well and is consistently designed badly. A good work sample resembles the actual job, takes a bounded and respectful amount of the candidate's time, is paid if it produces anything you would use, and is scored against a rubric written in advance. A five-hour unpaid take-home selects for availability rather than ability.
Run a Reference Check That Tells You Something reworks a stage most organisations conduct as a formality. Specific, behavioural, calibrating questions produce information; would you work with them again
does not.
Decide, then close
Run a Hiring Debrief and Make the Decision protects the evidence you spent the whole process collecting. The standard failure is the loudest or most senior person in the room anchoring everyone else in the first two minutes. The prompt has interviewers submit scores independently before discussion, then focuses the conversation on disagreements — which is where the useful information lives.
Make a Job Offer and Handle the Counter covers the close, including deciding your ceiling before the conversation rather than during it.
Reject a Candidate Without Burning the Relationship matters more than its position in the process suggests. Rejected candidates are future applicants, future customers, and a substantial share of your reputation as an employer. The prompt is also careful about what feedback to give and what not to — specificity is kind, but detailed critique of a protected characteristic or a medical inference is a liability as well as wrong.
Build a 30-60-90 Day Onboarding Plan closes the loop back to the scorecard, so the first ninety days are measured against the same targets you hired for. The management pack covers what comes after.
Two prompts here apply the same discipline outside employment: Recruit a Board That Fills Your Actual Gaps and Build a Volunteer Program That Doesn't Burn People Out, both from the nonprofit pack.
Where this stops
Hiring is among the most heavily regulated things a business does, and the rules vary by jurisdiction. These prompts draft scorecards, guides, rubrics and messages — they do not know your local law on protected characteristics, permissible questions, salary history, pay transparency, background checks, accommodation, or record retention, and several of those carry real penalties. Employment counsel should review your process before it runs, not after a complaint.
One thing worth stating plainly: structure is also the best-evidenced way to reduce the influence of bias in selection, because it constrains the space in which unexamined preference operates. That is a reason to adopt it, not a claim that adopting it discharges any legal obligation.
Sources
- Paul R. Sackett, Charlene Zhang, Christopher M. Berry and Filip Lievens, Revisiting Meta-Analytic Estimates of Validity in Personnel Selection: Addressing Systematic Overcorrection for Restriction of Range, Journal of Applied Psychology, 2022 — structured interviews r = .42, cognitive ability r = .31
- Frank L. Schmidt and John E. Hunter, The Validity and Utility of Selection Methods in Personnel Psychology, Psychological Bulletin, 1998 — structured interviews .51 against .38 for unstructured