Gender Pay Gap Calculator — unadjusted and adjusted, side by side

The gender pay gap is a published statistical measure with a precise definition. Eurostat defines the unadjusted version as the difference between the average gross hourly earnings of male and female employees expressed as a percentage of the average gross hourly earnings of male employees. This gender pay gap calculator applies that definition to figures you supply, on both a median and a mean basis, and separately computes an adjusted gap if you have like-for-like inputs.

Arb Digital publishes free calculators for measures that are widely quoted and frequently miscomputed. This page sets out the arithmetic and the definitions. It does not interpret the result, attribute a cause, or make any recommendation about pay practice — those are questions the arithmetic cannot answer.

What This Gender Pay Gap Calculator Does

It performs the same subtraction and division three times over. Given men's and women's median hourly pay, it returns the unadjusted median gap. Given the two means, it returns the unadjusted mean gap. Given adjusted figures from a like-for-like comparison, it returns the adjusted gap. All three are expressed as a percentage of men's pay, which is the convention used by Eurostat, the UK Office for National Statistics and the OECD.

Two supporting figures make the result easier to read. The pay ratio expresses women's median pay per one unit of men's median pay, which some readers find clearer than a percentage gap. And the cumulative difference multiplies the hourly gap by paid hours per year and years of working life, showing what the same hourly difference sums to over a career if it persisted unchanged. That last figure is arithmetic on a fixed assumption, not a forecast.

How to Use It

1. Enter median hourly pay for each group. Use gross hourly pay, excluding overtime, which is the basis the ONS uses for its headline figure.
2. Enter mean hourly pay for each group. Both measures are usually published together, and the difference between them is informative in itself.
3. Enter adjusted figures if you have them. These must come from a like-for-like comparison — the same role, level, contracted hours, location and experience — otherwise the third figure is not an adjusted gap.
4. Set annual hours and working years to see what the hourly difference sums to over a career at constant rates.
5. Read all three gaps together. They answer different questions, and reporting only the one that suits an argument is exactly the error this page is built to avoid.

The Formula and How It Is Calculated

Every gap on this page uses one expression:

Gap % = (men's pay − women's pay) ÷ men's pay × 100

Working the defaults through. Median: (22.50 − 20.93) ÷ 22.50 = 1.57 ÷ 22.50 = 6.98%. Mean: (24.80 − 22.10) ÷ 24.80 = 2.70 ÷ 24.80 = 10.89%. Adjusted: (22.50 − 21.83) ÷ 22.50 = 0.67 ÷ 22.50 = 2.98%.

The pay ratio is the simple quotient: 20.93 ÷ 22.50 = 0.9302, so women's median hourly pay is 93.02 per 100 of men's. The cumulative figure multiplies the median hourly difference by hours and years: 1.57 × 1,800 × 35 = 98,910. Note that a positive gap means men's pay is higher and a negative gap means women's pay is higher; the sign is part of the result, not a formatting detail.

Unadjusted and Adjusted Measure Different Things

This is the distinction that matters most, and conflating the two is the most common error in reporting on this subject.

The unadjusted gap compares all men to all women across a population, with no controls at all. It reflects the whole composition of who is employed in what: the mix of occupations, seniority levels, sectors, contracted hours and working patterns across the two groups, together with any difference in pay for the same work. Eurostat notes explicitly that the unadjusted measure encompasses factors beyond direct pay differences, including differences in employee characteristics and job types.

The adjusted gap compares men and women who match on a defined set of characteristics — typically role, level, contracted hours, location, tenure and experience. Because those controls remove composition from the comparison, the adjusted figure is normally much smaller than the unadjusted one, as it is in the defaults above.

Neither is a corrected version of the other. The unadjusted gap answers "across this whole population, how do typical earnings compare?" The adjusted gap answers "among people matched on these specific characteristics, how does pay compare?" Both are legitimate questions. What is not legitimate is quoting one number while describing the other, or presenting an adjusted figure without saying which variables were controlled for — the choice of controls determines the answer, and two analysts controlling for different things will produce different adjusted gaps from identical payroll data.

Median or Mean: Why Both Are Published

The mean is the arithmetic average and the median is the midpoint of the distribution. They differ whenever pay is skewed, and pay is always skewed, because a small number of very high earners pull the mean upward while leaving the median where it is. The NIST/SEMATECH statistics handbook's discussion of measures of location makes the general point: for skewed and heavy-tailed distributions the mean and median diverge, and the mean is the measure more sensitive to extreme values.

In pay data this has a concrete consequence. If one group holds more of the highest-paid positions, the mean gap will exceed the median gap, as it does in the default figures here. The median is therefore less affected by a handful of very large salaries, while the mean captures what is happening at the top of the distribution. Statistical offices publish both for that reason. The UK's mandatory employer reporting requires both a mean and a median figure precisely so that neither can be quoted alone.

Published National Figures, and Why Yours Will Differ

The ONS reported that the gender pay gap among full-time employees in the UK stood at 7.0% in April 2024, down from 7.5% the previous year, measured as the difference between median gross hourly earnings excluding overtime as a proportion of men's. That figure covers full-time employees only; the gap across all employees, including part-time work, is published separately and is substantially larger, because part-time working is unevenly distributed between the groups.

Any figure you calculate for a single organisation will differ from a national one, and the difference is not evidence of anything on its own. National statistics cover an entire economy with its full mix of industries; a single employer covers one industry, one set of roles and one grade structure. Eurostat's figures apply to enterprises with ten or more employees and are computed on gross hourly earnings, so even a like-for-like national comparison requires checking the population, the earnings definition and the reference period all match.

Getting the Inputs Right

Three input choices change the answer more than anything else. First, hourly versus annual pay: comparing annual salaries across groups with different contracted hours mixes a pay difference with an hours difference. Hourly pay is the standard basis for exactly that reason. Second, what counts as pay: gross pay before deductions is the norm, and whether bonuses, overtime, shift premia and salary sacrifice arrangements are included must be stated, since each can move the figure by a percentage point or more. Third, the reference period: a single pay period is the usual snapshot, and a period containing an annual bonus round produces a very different mean from one that does not.

If you need to convert between pay bases before using this tool, the hourly to salary calculator and the salary to hourly calculator handle the conversion consistently, and the mean, median and mode calculator will produce both averages from a list of individual pay figures. For the underlying percentage arithmetic on any two values, the percentage difference calculator covers the general case.

What This Page Deliberately Does Not Do

It does not attribute the gap to any cause. A percentage difference between two averages is consistent with many different underlying situations, and distinguishing between them requires individual-level data, a specified model and a clear statement of what has been controlled for. Nothing on this page constitutes such an analysis.

It also does not tell any employer what to do. Reporting obligations, and what follows from a published figure, are legal and policy questions that vary by jurisdiction, employer size and sector. If you are subject to a statutory reporting requirement, the governing regulations and the relevant government guidance define the calculation you must perform, and those definitions take precedence over the general arithmetic here.

Common Mistakes to Avoid

• Describing an unadjusted gap as a like-for-like pay difference — the unadjusted figure includes the whole composition of who works in which role, and says nothing about pay for identical work.
• Quoting an adjusted gap without listing the controls — the choice of variables determines the result, so the figure is uninterpretable without them.
• Comparing annual salaries instead of hourly pay — this folds any difference in contracted hours into what looks like a pay difference.
• Reporting only the mean or only the median — they diverge whenever pay is skewed, and publishing both is standard practice for that reason.
• Comparing an organisation's figure to a national one without checking that population, earnings definition and reference period all match.

Related Free Tools From Arb Digital

Use the mean, median and mode calculator to derive both averages from raw pay lists, the hourly to salary calculator and salary to hourly calculator to put figures on a common basis, the percentage difference calculator for the general percentage arithmetic, and the pay raise calculator for individual pay changes. The full free online tools hub lists everything else.

Frequently Asked Questions

This tool performs the published percentage arithmetic on the figures you enter. It does not attribute any result to a cause, does not constitute legal, statistical or employment advice, and does not determine any reporting obligation. Where a statutory reporting requirement applies, the governing regulations define the calculation and take precedence.