Compensation Equity Analysis

Pay equity is no longer just a moral imperative — it's a legal requirement. The EU Pay Transparency Directive takes effect in June 2026, requiring companies with 100+ employees to report gender pay gaps and take corrective action if the gap exceeds 5%. In the U.S., California, New York, Colorado, Washington, and Illinois have enacted pay transparency laws. Most HR teams still compare average salaries by gender in a spreadsheet and call it an audit. This analysis does what a proper audit requires: compare compensation across groups while controlling for legitimate pay factors, test whether the remaining gaps are statistically significant, and produce results that stand up to regulatory scrutiny.

Why a Spreadsheet Average Is Not a Pay Equity Audit

The raw pay gap is the simple difference in average pay between two groups. In the U.S., women earn approximately 84 cents for every dollar men earn. In the EU, the average gender pay gap is around 12.7% (EU Council, 2026). These numbers are real but misleading for your organization's internal audit, because they don't account for legitimate pay factors.

A proper compensation equity analysis asks a more specific question: after controlling for job role, experience, education, performance, and location, do statistically significant pay gaps remain? The adjusted gap is almost always smaller than the raw gap. But it's the adjusted gap that matters legally and operationally, because it reveals whether the company is paying differently for the same work.

Enterprise pay equity tools — Syndio, Trusaic, PayScale — start at $50,000+ per year and require long implementation cycles. Mid-market companies (200 to 10,000 employees) face a gap: they need a proper statistical analysis but can't justify the enterprise price tag. This analysis fills that gap. You upload a CSV from your HRIS, specify the compensation column and the grouping variables, and get a statistical comparison in under 60 seconds.

What the Law Requires

The regulatory landscape is tightening rapidly:

The common thread: governments are moving from "we encourage fair pay" to "prove it with data." A spreadsheet of averages does not meet that standard. A statistical analysis with documented methodology, significance testing, and effect sizes does.

How a Proper Pay Equity Analysis Works

A defensible audit has two layers:

Layer 1: Unadjusted group comparison — compare mean and median compensation between groups (male vs. female, or across ethnic groups) without any controls. This is the "headline number" that matches the publicly reported pay gap. You produce it with a t-test (two groups) or ANOVA (three or more groups). Even though it doesn't control for pay factors, this number matters: it's what regulators report, it's what the press cites, and it's what your employees will compare their experience against.

Layer 2: Controlled comparison — the same analysis, but run within comparable groups. Instead of comparing all men to all women, compare men and women within the same job level, department, and experience band. If you have enough granularity in your data (30+ employees per comparable group), this approach controls for legitimate pay factors without requiring multivariate regression. The groups should differ only by the protected characteristic, so any remaining gap is unexplained.

The first layer tells you what the gap looks like. The second layer tells you whether the gap is fair. Both are necessary for a complete audit.

What Data Do You Need?

A CSV export from your HRIS (Workday, ADP, BambooHR, SAP SuccessFactors) combining compensation and demographic data. Most systems can produce this as a custom report. Some companies need to merge payroll data with demographics from a separate HRIS export.

Required columns

Columns for a controlled analysis

Sample size guidance

How to Read the Report

Group means and medians — the raw numbers. Mean salary for men, mean salary for women, and the difference in both absolute dollars and percentage terms. The median is often more informative than the mean for compensation data, because a few highly paid executives can pull the mean up without reflecting the typical employee's experience.

Box plots — visual comparison of the salary distributions by group. Overlapping boxes suggest similar compensation. Separated boxes with visible gaps confirm the statistical result visually. Look for outliers (the dots beyond the whiskers) — a few extremely high or low salaries can distort the means.

ANOVA F-test (for 3+ groups) — if you're comparing pay across multiple ethnic groups, departments, or job levels, the F-statistic and p-value tell you whether at least one group differs significantly. A p-value below 0.05 means the gap is real. But significance alone isn't enough — look at the effect size (eta-squared). An effect size below 0.01 means the gap is statistically real but practically tiny.

Tukey HSD post-hoc comparisons — after ANOVA finds a significant overall difference, Tukey identifies which specific pairs differ. This is critical for pay equity: "Compensation differs significantly across ethnic groups" is a finding. "Hispanic employees earn significantly less than White employees (difference = $4,200, p = 0.008), while Black-White and Asian-White differences are not significant" is an actionable finding.

Confidence intervals — for each comparison, the report shows the estimated size of the gap with upper and lower bounds. A confidence interval that doesn't cross zero confirms significance. The width tells you about precision: a narrow interval ($3,800 to $5,200) gives you a specific remediation target. A wide interval ($500 to $9,000) means you need more data or tighter comparable groups.

Effect size (Cohen's d) — measures practical significance, separate from statistical significance. A d of 0.2 is a small effect, 0.5 is medium, and 0.8 is large. In compensation analysis, even a small effect size can translate to meaningful dollar amounts when multiplied across the workforce.

What to Do With the Results

Immediate

Strategic

When to Use Something Else

References