Bertrand-Mullainathan, the 2023 NBER replication, age discrimination audits, and what the field-experiment literature actually shows about hiring bias.
Resume audit studies — where researchers send paired fictitious resumes that differ only on a single attribute and measure callback rates — are one of the cleanest tools social scientists have to measure hiring bias. The literature now spans more than two decades, dozens of countries, and tens of thousands of submitted applications. The findings are consistent enough to be uncomfortable: bias is real, persistent, and measurable across race, gender, age, name, and address.
In a field experiment sending nearly 5,000 fictitious resumes to Boston and Chicago employers, identical resumes with names commonly perceived as African-American (Lakisha, Jamal) received roughly half the interview callbacks of identical resumes with names perceived as White (Emily, Greg). Published in the American Economic Review.
Sources: [1]
A more recent large-scale field experiment, sending over 80,000 applications to Fortune 500 employers, found that aggregate racial gaps in callback rates persist, but that the bias is concentrated in a relatively small subset of firms. Some employers showed essentially no measurable bias; others showed large effects.
Sources: [2]
Neumark, Burn, and Button conducted resume audits across multiple US states for retail, administrative, and customer service roles. Older applicants — especially women in retail — received significantly fewer callbacks than younger applicants with otherwise-identical resumes.
Sources: [3]
Quadlin (2018) found that "high-achieving" female applicants received fewer callbacks than equivalent male applicants for STEM-adjacent roles, suggesting the bias is conditional on field, not universal across all hiring.
Sources: [4]
Audit studies replicated across Europe (Heath & Di Stasio in the UK, Petit et al. in France, Kaas & Manger in Germany) find similar magnitudes of name-based bias, though specific groups affected vary by country's demographic context.
Sources: [5]
Audit studies are the gold standard for measuring labor market discrimination because they control for everything except the variable being tested. The studies cited here are peer-reviewed or working papers from leading economics departments and bureaus (NBER, journals AER and ASR). Effect sizes are reported as published.
Hiring bias is real and measurable. For job seekers, this is not advice to misrepresent yourself — it's context for understanding why the same resume can perform differently depending on who reads it. Tactically: referrals (which bypass resume-stage screening), employer audits and inclusion data, and applying through skills-first hiring programmes are all paths that demonstrably reduce the bias surface. For employers: anonymised resume review and structured interviews have published evidence of reducing the gap.
How many resumes really get rejected before a human sees them, who's running the systems, and what the published data does — and doesn't — say.
IBM, Google, and 50+ companies announced commitments to skills-first hiring. The published data on what actually changed is more modest than the headlines.
The data shows what works. Build the resume that benefits from it.