Send two CVs to the same job, the same degrees, the same experience, the same spelling, and change nothing but the name at the top. In a famous field experiment, the one with a White-sounding name got noticeably more callbacks than the one with a Black-sounding name. Nobody on the hiring side believed they were biased. That gap is the problem inclusive hiring sets out to solve, and the surprising lesson is that you fix it less by changing minds than by changing the process those minds run inside.

The quick version

  • Inclusive hiring means designing a recruitment process so that capable candidates aren't screened out for reasons unrelated to the job, and so the best person is more likely to be seen, fairly, whoever they are.
  • The most reliable lever is structural, not attitudinal: redesign how you write ads, screen CVs, and run interviews. De-bias the process, because de-biasing individual minds turns out to be hard and unreliable.
  • Two structural moves have strong evidence behind them: blind screening (hide demographic cues early) and structured interviews (same questions, scored against the same anchors for every candidate).
  • The trap is treating this as a values badge or a quota. Done as theatre it breeds backlash; done as design it simply makes your hiring more accurate, which is the honest case for it.

The idea in depth

Start with the evidence that there is something to fix, because the case rests on it. In 2004 economists Marianne Bertrand and Sendhil Mullainathan published "Are Emily and Greg More Employable than Lakisha and Jamal?" in the American Economic Review. They sent roughly 4,870 fictitious applications to real job ads in Boston and Chicago, randomly assigning each CV a name that signalled race. Identical résumés with White-sounding names drew about 50% more callbacks than those with Black-sounding names, and a stronger CV helped White applicants far more than it helped Black ones. The discrimination wasn't a story applicants told; it was visible in the response data of employers who would almost all have described themselves as fair.

So the first move is to stop relying on goodwill as a control. Goodwill is real, but it doesn't survive a busy afternoon screening forty CVs in an hour. The practical response isn't a training session reminding everyone to be open-minded; it's to remove the cue that triggers the snap judgment in the first place, which is exactly what blind screening does. Strip the name, photo, address and graduation year from a CV before a human ranks it, and the shortcut has nothing to grab.

An honest limitation. Blind screening protects the early filter, not the whole funnel, the moment a candidate walks into a room or joins a video call, the cues return. It can also be awkward to implement when a CV's narrative gives identity away regardless. Treat it as one reliable gate, not a cure, and pair it with what happens later in the process.

Why structure beats sincerity

The deeper principle comes from behavioural economist Iris Bohnet of the Harvard Kennedy School, whose 2016 book What Works: Gender Equality by Design (Harvard University Press) makes the argument plainly: trying to de-bias individual minds is difficult, expensive, and often disappointing, while redesigning the choice environment around those minds is cheaper and works. Her hiring prescriptions are concrete, evaluate candidates comparatively and in batches rather than one at a time, remove demographic information from applications, and use structured interviews. The reframe is the sticky idea here: de-bias the system, not the person.

The reliable fix isn't a better-intentioned interviewer. It's a process where good intentions matter less.

Structured interviewing isn't just fairer, it's more accurate, which is what makes the case to a sceptical hiring manager. The landmark meta-analysis by Frank Schmidt and John Hunter, "The Validity and Utility of Selection Methods in Personnel Psychology" (Psychological Bulletin, 1998), aggregated decades of research and put the predictive validity of a structured interview far above an unstructured one, the difference between an interview that forecasts job performance and a pleasant chat that mostly forecasts who the interviewer liked. (More recent re-analyses by Sackett and colleagues revise some of those numbers downward, but the direction holds: structure beats free-form.) So the move is to give every candidate the same job-relevant questions, decide the scoring rubric before the interviews, and score each answer as you go rather than forming a gestalt impression and justifying it afterwards.

flowchart TD
  A(["Job opening"]) --> B(["Write the ad:
essential criteria only,
neutral language"])
  B --> C(["Blind screen:
hide name, photo,
address, grad year"])
  C --> D(["Structured interview:
same questions,
rubric set in advance"])
  D --> E(["Score per answer,
then compare
candidates side by side"])
  E --> F(["Decision on
job-relevant evidence"])
Inclusive hiring as a redesigned funnel, each stage closes a door that bias usually walks through. Leaders Loop

A second honest limitation. The most cited blind-hiring result, Claudia Goldin and Cecilia Rouse's study of "blind" orchestra auditions (American Economic Review, 2000), which estimated that screening candidates behind a curtain explained a meaningful share of the rise in women hired by major US orchestras, has been challenged. Statistician Andrew Gelman and others have argued the sample is small and the headline estimates fragile. The honest position: the orchestra study is a vivid illustration, not a settled proof, and the stronger evidence for process redesign rests on the broader body of work on structured selection, not on that single result. Cite it as a story, lean on the meta-analysis for the claim.

A worked example

Take a 40-person software company, call it Harbor, hiring its first three customer-success managers. (Illustrative throughout; this is a teaching example, not a real firm.) The old process: the founder posts an ad asking for a "rockstar self-starter," skims CVs over coffee, and interviews whoever "feels like a fit." The last four hires all went to the same university and, by coincidence the founder hasn't noticed, all sound alike on a call.

Harbor redesigns it without spending a cent. First, the ad: out goes the vibe language and the decorative "degree from a top university"; in comes a short list of what the job actually needs, handling an upset customer, reading a usage dashboard, writing a clear follow-up email. Second, the early screen goes blind: an assistant removes names, photos and graduation years before the hiring panel ranks the written applications, so the first cut is made on relevant experience alone. Third, the interview is structured: three job-relevant scenarios, the same for every candidate, with a 1–4 rubric agreed in the room before anyone walks in. Each panellist scores independently, then they compare.

flowchart LR
  A(["Old way:
vibe ad, coffee skim,
'feels like a fit'"]) --> B(["Same-looking
shortlist"])
  C(["New way:
criteria ad, blind screen,
scored scenarios"]) --> D(["Wider, stronger
shortlist"])
  D --> E(["Hire on evidence,
and defensible
if challenged"])
Same role, two processes, the redesign widens the field before it narrows it. Leaders Loop

The outcome isn't that Harbor "lowers the bar", the opposite. A candidate who'd have been filtered out at the coffee stage for an unfamiliar name and a non-target university turns out to score highest on the customer-recovery scenario, because she ran a busy café complaints desk for three years. The structured process surfaced job-relevant evidence the old one threw away. That is the whole argument in one hire: inclusive hiring isn't charity toward the candidate, it's accuracy for the company.

Frequently asked questions

Isn't this just lowering standards to hit a diversity number?

It's the reverse. A structured interview predicts job performance better than an unstructured one (Schmidt & Hunter, 1998), and blind screening stops you discarding strong candidates for irrelevant reasons. You're raising the standard of evidence you decide on. Quotas are a separate, contested policy question; the process moves here improve accuracy whether or not you ever set a target.

We're tiny and have no recruiter. Is any of this realistic?

The high-value moves are free. Writing the ad around essential criteria, hiding names on the first CV read, and agreeing three questions plus a scoring rubric before interviews cost nothing but a little discipline. Small teams arguably benefit most, because a single biased gut call carries more weight when there are only a handful of hires a year.

Does unconscious-bias training fix this instead?

On its own, rarely. The premise of the design approach, see Bohnet's What Works, is precisely that awareness training has a weak and short-lived track record, which is why redesigning the process is the more dependable lever. Training can support a culture, but don't mistake a workshop for a control on the actual decision.

Won't structured interviews feel robotic and put candidates off?

Structure governs the questions and scoring, not your warmth. You can open with rapport, explain why the format is consistent (most candidates appreciate knowing everyone gets the same shot), and still have a human conversation. What you give up is the freewheeling tangent that mostly measures chemistry, and that's the part you want to give up.

Is inclusive hiring legally required, and does it expose us to risk?

Anti-discrimination obligations vary by country and jurisdiction, so check your local law or a qualified professional rather than treating any single practice as compliance. As a general principle, a documented, criteria-based, consistently scored process is easier to defend than "we just had a good feeling", it leaves an evidence trail showing decisions were made on job-relevant grounds.

Related in the Toolkit

Inclusive hiring is one thread in a larger talent system: it starts upstream with the reputation that decides who applies at all (employer brand & talent attraction), and it shares its core mechanism, structure over gut feel, with the discipline of interviewing & selection.

Where to go next