How Simple Hiring Changes Can Close Gender Gaps

How Simple Hiring Changes Can Close Gender Gaps

Here is a paradox from the world of hiring: you can make a selection process more fair by adding a blind test, but you cannot make it fairer by simply removing the subjective parts. This is not obvious. Most of us assume that bias lives in human judgment, and that the way to fix bias is to eliminate human judgment. But a new study of tens of millions of hiring records from Brazil’s public sector suggests the opposite is true.

Tatiana Mocanu, an economist at the University of Illinois, analyzed what happened when a federal policy forced Brazilian government agencies to use more impartial hiring practices. The results, published as a working paper on the Social Science Research Network (Mocanu, 2024), show that the design of a hiring process matters far more than anyone had measured before. And the most effective changes are surprisingly simple.

What Happened When Brazil Made Hiring More Impartial

In 2012, Brazil passed a law requiring that all public sector hiring processes use at least one “impartial” screening tool. That meant written tests, blind evaluations, or any method where the evaluator did not know the candidate’s identity. Before the reform, many agencies relied entirely on subjective methods like interviews and CV reviews. After the reform, they had to add at least one objective filter.

Mocanu had access to a staggering dataset: tens of millions of records from Brazil’s public sector, covering every step of the hiring process. She could see who applied, how they scored, who evaluated them, what tools were used, and who got hired. This is the kind of data that most hiring studies only dream of. It allowed her to compare the same types of jobs before and after the reform, and to compare agencies that complied in different ways.

The headline finding: increasing screening impartiality improved women’s evaluation scores, application rates, and probability of being hired (Mocanu, 2024). The effect was not tiny. Women were more likely to apply when they knew the process was fairer. And when they did apply, they scored higher and got hired more often.

But here is where the story gets interesting. Not all impartiality is created equal.

Why Removing Interviews Does Not Work

You might think the simplest fix is to get rid of interviews entirely. Interviews are famous for bias. They reward confidence over competence, and they favor people who look and sound like the evaluator. If you want to close the gender gap, just kill the interview.

Mocanu’s data says no. When agencies removed subjective stages entirely and replaced them with only blind written tests, the gender hiring gap did not shrink. It stayed the same (Mocanu, 2024). This is the counterintuitive result that makes the paper worth reading.

Why would eliminating bias not help? Because subjective evaluations are not just sources of bias. They also provide information. An interview, for all its flaws, can reveal things a written test cannot: communication skills, problem solving under pressure, cultural fit. When you remove that information entirely, you lose screening precision. And that loss of precision can hurt women just as much as the bias did.

Think of it this way. Imagine a hiring process that uses only a multiple choice test. The test is blind and objective, but it measures a narrow set of skills. Women might score well on it, but if the test does not capture the full range of abilities needed for the job, the process still favors candidates who happen to have those specific test-taking skills. The gender gap does not disappear; it just shifts.

Mocanu’s model of hiring helps explain this. She shows that two factors determine gender outcomes: evaluator bias (how much the evaluator favors one group) and screening precision (how accurately the process measures true ability). Removing subjective tools reduces bias, but it also reduces precision. If the precision loss is large enough, the net effect on the gender gap is zero.

The Two Changes That Actually Worked

So if you cannot just remove interviews, what can you do? Mocanu found two specific changes that reliably increased women’s hiring odds (Mocanu, 2024).

First: add a blind written test to a process that already uses subjective methods. This is the easiest intervention. You keep the interviews and the CV reviews, but you also require candidates to take a standardized, anonymous test. The test provides objective information that counterbalances whatever bias might creep into the subjective rounds. Women’s scores go up, and they are more likely to advance.

Second: convert subjective rounds into only blind written tests. This is more aggressive. Instead of having both an interview and a test, you make the entire evaluation based on a blind test. But here is the catch: this only works if you keep the subjective rounds for other stages. If you convert the entire process to blind tests, you get the same null result as removing interviews entirely.

The key is balance. You want enough objective information to correct for bias, but you also want enough subjective information to maintain screening precision. The sweet spot seems to be a hybrid process where both types of evaluation coexist.

What About the Evaluators Themselves?

Mocanu also looked at the people doing the evaluating. She found that the composition of hiring committees matters. When committees were more gender balanced, male evaluators became more favorable toward female candidates in subjective stages (Mocanu, 2024).

This is a subtle but important effect. It is not that female evaluators are inherently less biased. It is that the presence of women on the committee changes how men behave. Perhaps they are more self conscious. Perhaps the discussion becomes more balanced. The data does not say exactly why. But the result is clear: if you want fairer subjective evaluations, make sure the evaluators themselves are diverse.

What This Means for Your Organization

The study is about Brazil’s public sector, which is a specific context. But the underlying logic applies anywhere people are hired. Here are the direct implications.

• ▸Do not eliminate subjective evaluations. Supplement them. If your hiring process relies entirely on interviews and CV reviews, add a blind test. A simple written exam or a work sample evaluated anonymously can correct for bias without losing information.

• ▸If you must remove subjective stages, replace them with something that provides equivalent information. A blind test is not a perfect substitute for an interview. If you drop the interview, the test must measure the same kinds of skills. Otherwise, you will just swap one form of bias for another.

• ▸Diversify your hiring committees. Even if you cannot change the evaluation tools, changing who does the evaluating can help. More women on the committee means male evaluators become less biased. This is not about quotas. It is about changing the social dynamics of evaluation.

• ▸Measure your screening precision. Most organizations have no idea how well their hiring process predicts job performance. That is a problem. If you do not know what you are measuring, you cannot know if a change helps or hurts. Mocanu’s model shows that precision matters as much as bias. Track it.

• ▸Do not assume blind processes are always better. Blind auditions helped orchestras hire more women, but that does not mean blind hiring works everywhere. In some contexts, the loss of information outweighs the reduction in bias. Test your specific process before making changes.

What the Research Does Not Prove

This study is based on observational data, not a randomized experiment. Mocanu exploits a policy reform that affected all agencies at the same time, which is a strong research design, but it is not perfect. Agencies that complied more aggressively with the reform might be different in other ways. They might have been more progressive, or better managed, or more willing to experiment. Those differences could influence the results.

Also, the study measures what happens in Brazil’s public sector, which has its own norms and incentives. Private sector hiring in other countries might respond differently. The specific effects might not replicate exactly.

But the core logic is not specific to Brazil. The trade off between bias and precision is a fundamental feature of any selection process. If you remove a source of information, you lose information. That is a mathematical truth, not a cultural one.

What This Actually Means

• ▸Adding a blind test to a subjective process is the single most effective change you can make. It reduces bias without sacrificing information. Start there.

• ▸Removing subjective stages is a trap. It feels like progress, but it often does nothing for the gender gap. Do not do it unless you are certain the replacement tool captures the same skills.

• ▸Diversity on hiring committees changes male evaluator behavior. The effect is real and measurable. If your committees are all male, diversify them. You do not need a majority; even a few women can shift the dynamics.

• ▸Precision matters as much as fairness. A hiring process that is fair but inaccurate will still produce unequal outcomes. Measure both.

• ▸The goal is not to eliminate human judgment. It is to structure it. Bias is not a bug that can be patched out. It is a feature of how humans evaluate each other. The best you can do is design a process that accounts for it. Mocanu’s paper shows exactly how to do that.