A hiring manager sits across from two candidates for the same regional sales role. One went to a similar university, uses the same turns of phrase, laughs at the same jokes. The other doesn't. Ninety minutes later, the manager can't quite explain why, but the first candidate just "felt right."
That feeling isn't chemistry. It's affinity bias doing exactly what it evolved to do: rewarding familiarity and mistaking it for fit.
The interview is over before it starts
Prickett, Gada-Jain and Bernieri's 2000 study filmed 59 real job interviews and showed observers nothing but the opening handshake and greeting. Their snap judgements of the candidates lined up closely with the interviewers' own final evaluations, made a full 20 minutes later, after the actual conversation happened. The interview content wasn't doing the work. The first few seconds were.
This is the uncomfortable part for anyone who trusts their read on a room: the remaining 45 minutes of most interviews function less as evaluation and more as confirmation.
The interviewer isn't gathering new evidence. They're building a case for a verdict they already reached.
Structure helps. It doesn't fix everything
The standard advice is to structure the interview - same questions, same order, same scoring rubric. It works, to a point.
Schmidt and Hunter's 1998 meta-analysis of 85 years of personnel selection research found unstructured interviews predict job performance at a validity of .38.
Structured interviews reach .51. Combine a structured interview with a validated cognitive ability assessment and predictive validity climbs to .63. That’s one of the strongest combinations in the entire selection literature.
A 2022 re-analysis by Sackett and colleagues revisited these numbers using more conservative statistical corrections and reached a similar conclusion from a different angle: structured interviews, not gut-led ones, sit at the top of what actually predicts performance.
The direction of the evidence hasn't changed in over two decades. What's changed is how few organisations have built it into how they actually hire.
Structure removes some of the noise. It doesn't remove the interviewer. A well-run structured interview still asks a human being to score another human being on suitability and that's precisely the point at which affinity bias, halo effects and confidence-as-competence creep back in.
What South Africa's numbers are actually telling us
This isn't a theoretical problem for employers working against transformation targets.
Women hold 19.1% of board seats at JSE-listed companies. A figure that's barely moved in five years and sits below the global average of 23.3%, according to the Stellenbosch Business School's 2025 Women's Report.
Something in the pipeline that feeds senior roles isn't correcting itself, no matter how many bias-awareness workshops sit on the training calendar.
Awareness campaigns ask individuals to notice and override their own bias in real time, mid-interview, under social pressure, while trying to also evaluate a stranger's competence. That's a hard ask of anyone. It's also the wrong layer to intervene at.
Move the filter earlier
The fix isn't asking interviewers to be less biased. It's reducing how much interview-room judgement the process depends on in the first place.
That means assessing cognitive ability, personality traits, motivation and learning agility, before a CV, name, or accent ever enters the room.
Define what the role actually needs before the job is posted. Score every applicant against that definition, not against how well they perform in a conversation.
The practical barrier has usually been cost: per-candidate testing fees mean most companies only assess the shortlist, which means the bias has already done its work by the time testing starts.
A fixed-cost, unlimited-assessment model removes that constraint. You can screen the full applicant pool on the same objective criteria, not just the five people who happened to interview well.
None of this means interviews disappear. It means they stop being the first filter and start being the last conversation, confirming fit for a candidate whose capability is already established, rather than deciding capability based on how confidently someone can talk about themselves for forty-five minutes.
What to change this quarter
- Write the scorecard before the job ad. Define the cognitive and behavioural benchmarks the role needs, before anyone starts reviewing applicants.
- Test the full pipeline, not the shortlist. If cost is limiting who gets assessed, that's the bias re-entering through the back door.
- Make hiring managers defend shortlists with data. "Great culture fit" isn't a selection criterion. Ask which benchmark the candidate scored against, and by how much.
The organisations that move transformation numbers aren't the ones running more bias training. They're the ones that changed what decides who gets shortlisted in the first place.
References
Prickett, T., Gada-Jain, N., & Bernieri, F. J. (2000) University of Toledo. The importance of first impressions in a job interview. Paper presented at the Midwestern Psychological Association Convention, Chicago, IL. https://www.researchgate.net/publication/313878823_The_importance_of_first_impressions_in_a_job_interview
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. Summary: https://firstpersonnel.org/wp-content/uploads/2013/10/Summary-Schmidt-Hunter-1998.pdf
Sackett, P. R., Zhang, C., Berry, C. M., & Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Discussed in: Society for Industrial and Organizational Psychology (SIOP). https://www.siop.org/tip-article/is-cognitive-ability-the-best-predictor-of-job-performance-new-research-says-its-time-to-think-again/
Bosch, A. (2025). Women's Report 2025: Women on boards. Stellenbosch Business School, in association with the South African Board for People Practices. https://www.stellenboschbusiness.ac.za/news/2025-09-02-gender-diversity-board-level-severely-lacking-sa
