The Gender Pay Gap, Revisited: What the New Data Shows

On a Tuesday in late January, a recruiter in Minneapolis walked a finalist through an offer for a senior operations role: $118,000 base, an 8 percent bonus, and a start date the candidate could set herself. The candidate, a woman with 11 years of experience, asked for $125,000. The recruiter, who has placed more than 300 operations and supply-chain hires over 15 years and asked not to be identified because her clients keep offer details confidential, said she was caught off guard. "She was the first finalist in that search to ask for more," the recruiter said. "The others took the first number." Most of her candidates never ask, she added, and the ones who do almost always get something. The offer closed at $124,500, with the difference framed as a signing bonus.

That exchange sits inside a larger pattern. The gender pay gap — the difference between what men and women earn across the workforce — has barely moved in two decades, and the latest Census Bureau and Bureau of Labor Statistics figures put women's median earnings at about 82 to 84 cents for every dollar men earn. Beneath that familiar number, though, the story has changed. Economists increasingly argue that the largest drivers are no longer overt pay discrimination but the occupations women work in, the caregiving they take on, and the negotiations they never have. This article lays out the new data, the argument over what the numbers mean, and what salary transparency laws have changed so far.

The distinction that matters most is between two different measurements. The unadjusted gap compares all men and all women. The adjusted gap compares men and women in the same job with similar experience and education, and it is far smaller — in some analyses, close to zero. Neither number is wrong; they answer different questions, and the debate over which one describes reality has shaped everything from corporate pay policy to state law.

The gap, in the latest numbers

Start with the headline figure. In the Census Bureau's American Community Survey, the median woman working full time and year-round earned about 83 cents for every dollar earned by the median man in the most recent full year of data, and the Bureau of Labor Statistics, which tracks median weekly earnings of full-time workers, puts the ratio near 84 cents. Pew Research Center's analysis of Census data found the gap had narrowed by only about 2 cents in the previous 20 years — from roughly 80 cents on the dollar in 2002 to about 82 cents in 2022.

The two surveys measure slightly different populations, which is why they differ by a cent or two. The Census figure covers full-time, year-round workers age 15 and older. The BLS figure covers anyone who worked at least 35 hours in the surveyed week. What both agree on: the gap is real, it stands at roughly 16 to 18 percent, and it has been flat since the early 2000s after narrowing sharply through the 1980s and 1990s.

The flatness is the story. The equal-pay laws of the 1960s and the movement of women into college and professional work closed the early gap. What remains is not the residue of an old problem but something newer — and the age breakdown shows where it lives. The pattern holds across income groups, education levels, and regions: the older the worker, the wider the gap.

Women's median weekly earnings as a share of men's, by age. BLS usual weekly earnings, 2025. Figures approximate.

Age group	Share of men's earnings
16–24	about 93 cents
25–34	about 90 cents
35–44	about 84 cents
45–54	about 80 cents
55–64	about 79 cents

Women start near parity and lose ground with each decade. That shape — flat at 25, a widening gap through the 30s and 40s — is the single most important fact in the new data, and it points away from hiring discrimination and toward the years when careers and family collide.

The adjusted gap: same job, same experience

Compensation firms that control for job title, experience, education, and hours get a very different picture. Payscale's most recent analysis found that when men and women with the same job, level, education, and years of experience are compared, women earn about 99 cents for every dollar men earn. Glassdoor's earlier analysis of its own salary data found a controlled gap of roughly 5 percent — 95 cents on the dollar — after accounting for job, seniority, and employer.

The two approaches bracket the truth. The unadjusted number measures the world as it is: women earn less because they are overrepresented in lower-paid fields, more likely to step out of the labor force, and less likely to sit at the top of the pay scale. The adjusted number measures what happens inside a single job — and finds that most of the gap disappears. For a worker deciding whether to ask for a raise, the adjusted number is the more useful one. For a society deciding whether a problem exists, the unadjusted number is the one that matters.

The Economic Policy Institute, which studies the gap from the left, cautions against reading the adjusted number as proof that discrimination has ended. Its researchers argue that controlling for occupation can hide the sorting itself — the ways schools, families, and hiring managers channel women into lower-paying work before the first paycheck is signed. If the sorting is the problem, then the adjusted gap is not the absence of a problem; it is the problem wearing different clothes.

"The unadjusted number is a social fact. The adjusted number is a labor-market fact. Both are true, and they describe different problems." — a compensation consultant who advises about 40 employers a year on pay equity

Where the money is: occupational sorting

The sorting is easy to see in BLS occupational data. Among the 20 highest-paying occupations the bureau tracks — surgeons, anesthesiologists, chief executives, airline pilots — men hold the large majority of jobs. Among the lowest-paying — childcare workers, home health aides, cashiers, restaurant servers — women do. The median pay for childcare workers is about $31,000 a year; for surgeons, roughly $250,000. The distance between those numbers is a large part of the gender gap, before any individual paycheck is signed.

Researchers have argued for decades about whether female-dominated fields pay less because of the work or because of the gender of the people doing it. Studies of comparable worth — the idea that jobs requiring similar skill and responsibility should pay similar wages — have found that occupations lose pay as they become majority female. A field that starts out male-dominated and flips to female-dominated pays less afterward, even when the work is unchanged. Economists call it occupational devaluation; the practical effect is that a woman who takes over a job a man once held often earns less than he did.

Hours matter too. Women are about twice as likely as men to work part-time, and part-time work pays less per hour in most occupations, so the part-time years widen the gap measured over a career even when each year's full-time wage is fair. That is also why the lifetime earnings gap is wider than any single year's figure. The compounding is the point: a few part-time years in one's 30s, when earnings curves are steepest, cost more than the same years later in a career.

The motherhood penalty

The single largest factor behind the remaining gap, in the view of most labor economists, is children. The research on the motherhood penalty is decades old and remarkably consistent: women's earnings dip after a first child and never fully recover, while men's earnings rise slightly after fatherhood. A landmark study in the American Sociological Review estimated the penalty at about 4 percent per child after controlling for experience and education.

More recent work using tax records has sharpened the picture. A 2019 study in the American Economic Review, based on Danish administrative data, found that the arrival of a first child created an immediate, persistent earnings gap of about 20 percent between otherwise similar men and women, driven almost entirely by women's reduced hours and time out of the workforce. Follow-up studies using American data found a similar child penalty in the United States — on the order of 20 to 30 percent in the years after a first birth — and showed that it narrowed only slightly over a decade. The penalty is largest for women with graduate degrees, who face the steepest earnings trajectories to lose.

The fatherhood side is its own finding: men's earnings tend to rise modestly after they have children, a pattern researchers attribute to employers' perception of fathers as more committed and to men's own increased hours. The two effects move in opposite directions, which is why the gap between mothers and fathers is wider than the gap between women and men broadly. For a mother of two and a father of two, the difference in pay approaches 25 percent — a chasm no negotiation script can close.

The ask: negotiation and the first offer

Negotiation differences play a smaller but real role. The classic evidence comes from studies of graduating students: in a widely cited experiment at a large American university, about 7 percent of women negotiated their initial job offer, compared with 57 percent of men — and the women who asked improved their offers by about 7 percent on average, roughly the same gain the men who asked received. The difference was not in the outcome of asking; it was in who asked.

Experimental research on backlash has shown why the ask is harder for women: when women negotiate, evaluators often rate them as less likable and less desirable to work with, a penalty that does not attach to men who make the same request. Recruiters describe the effect in plainer terms. "Candidates assume the first number is the number," said a recruiter in Austin who has negotiated more than 200 offers. "It almost never is. The range is real, and nobody checks it." Pay data support the point: Payscale's surveys find that women negotiate less often than men at the same level, and the gap in who asks is widest early in careers, when the stakes of anchoring are highest.

The first salary matters disproportionately because every later raise and offer is built on it. Research on anchoring finds that workers who accept an offer at the bottom of a range tend to stay below the midpoint of their pay band for years, even after strong reviews. The compounding effect is why economists describe the ask gap as an amplifier: a small difference at the start becomes a persistent one. That is why pay-equity audits so often find the widest gaps not in hiring but in the slow drift of annual raises, which are percentage-based and preserve whatever difference existed at the start.

What transparency laws have changed

The newest lever is the law. Beginning with Colorado in 2021, more than a dozen states and several cities have required employers to post salary ranges on job listings. New York City followed in 2022, California and Washington state in 2023, and a second wave of states, including Massachusetts and Illinois, brought the rules into effect in 2025. Pew surveys conducted as the laws spread find that about three in ten workers now say salary ranges appear in most or all of the postings they see — a share that has grown steadily.

Early research suggests the laws are moving the numbers, if slowly. Studies of Colorado's posting requirement, which took effect in January 2021, found that advertised wages rose and that the gap between advertised pay for men and women narrowed. Analyses of New York City postings found uneven compliance at first — some employers posted ranges wider than $100,000 — and narrower ranges as enforcement began. The Economic Policy Institute's review of the research concludes that transparency reduces the pay gap among newly hired workers, though the effect on the median remains small.

The loopholes are equally visible. Ranges as wide as $70,000 to $200,000 satisfy the letter of many laws while disclosing almost nothing. Bonuses, equity, and commissions — the parts of pay where gaps are often widest — sit outside most posting requirements. And the laws govern what employers advertise, not what they pay: a company can post an honest range and still place every woman at its bottom. That is why the next frontier is pay-data reporting. California and Colorado now require large employers to file pay reports by gender, race, and ethnicity, and California's first filings showed women overrepresented in the lowest pay bands at most large companies.

Where the gap has narrowed

Amid the flat national number, one demographic stands out. Among workers ages 25 to 34, women earn about 92 cents for every dollar men earn, according to Pew's analysis, and in more than 20 of the 250 largest U.S. metros, young women earn the same as or more than young men. The edge reflects an educational advantage: young women are more likely than young men to hold a four-year degree, and at the start of careers the college wage premium outweighs the gender penalty.

The gap then widens with age — the motherhood penalty arriving on schedule. The trajectory, from near parity at 25 to a peak gap in the 40s and 50s, is so consistent across countries that labor economists treat it as a stylized fact of modern labor markets. It also suggests where the next decade of policy and negotiation will concentrate: not on the first offer, but on the years around a first child.

What to watch, and what to do

For the worker reading this, the practical question is what the data mean for one salary. The honest answer: the small adjusted gap is good news for the individual negotiation, because it means that in most fields the same job with the same experience pays about the same regardless of gender. The large unadjusted gap is good news for the individual ask, because it means the system still carries assumptions about who earns what — and a woman who negotiates with data is not fighting the market, she is correcting it.

Concretely: if you live in a state with a posting law, the range is public information; use it. Before a negotiation, find the market rate for your role and experience — the salary calculators on this site show how those figures are built, and the methodology page explains the sources — and bring a specific number to the conversation. Ask for the range when it is not posted. And if you manage people, the cheapest fix for the gap may be the one the Minneapolis recruiter watched happen: a candidate asked, the employer met her, and a $6,500 difference disappeared. The data suggest thousands of such differences are sitting in payroll files, uncorrected.

The next few years will test whether the laws catch up with the research. Watch the pay-data filings from California and Colorado, which will soon publish multi-year trends. Watch the second wave of posting laws, which will show whether ranges hold up under enforcement. And watch the child-penalty studies now testing whether the parental-leave expansions adopted by a dozen states move the earnings curve after a first birth. None of it will move the 16 percent overnight. But the history of the gap is a history of slow, compounding corrections — the 1960s laws, the 1980s educational shift, the 2020s transparency wave — and the next correction is already visible in the data.