Pay Gap Analysis
A statistical examination of compensation data that measures the difference in pay between employee groups (typically by gender or ethnicity) after accounting for legitimate factors like job level, tenure, and location.
What Is a Pay Gap Analysis?
Key Takeaways
- ✓A pay gap analysis is a data-driven comparison of compensation between employee groups, designed to surface pay differences that can't be explained by job-related factors.
- ✓The unadjusted pay gap measures the raw difference in median or mean pay between groups (e.g., women earn 84 cents per dollar men earn in the US). The adjusted gap controls for job level, tenure, performance, education, and geography.
- ✓Most organizations find an adjusted pay gap of 2-8% when they analyze their data for the first time (Payscale, 2024). Gaps below 2% are often within normal variation. Gaps above 5% typically signal systemic issues.
- ✓At the current rate of progress, the global gender pay gap won't close for 134 years (WEF, 2024). Proactive company-level analysis is the fastest path to closing it within your organization.
- ✓Pay gap analysis is now legally required in the UK (for gender), the EU (directive effective 2026), and several US states. Even where it isn't mandatory, it's becoming a standard risk management practice.
A pay gap analysis answers a straightforward question: are we paying people fairly, or are there gaps that correlate with gender, race, ethnicity, or other demographic characteristics? The answer usually isn't what leadership expects. Most organizations believe they pay fairly because they have salary bands and a compensation philosophy. But bands have ranges. Starting salaries differ. Raises compound differently over time. Promotions happen at different rates. After five years, two people hired into the same role at the same level can have a $15,000 pay gap with no documented justification. The analysis works in two layers. First, the unadjusted (or raw) gap: what's the difference in overall median pay between men and women, or between white employees and employees of color? This number is usually large because it reflects occupational segregation, level distribution, and other structural factors. Second, the adjusted gap: after you control for job family, level, tenure, location, performance, and education, what gap remains? This adjusted number is what matters most for legal compliance and internal remediation, because it represents unexplained pay disparity.
Types of Pay Gaps
Understanding which type of gap you're measuring changes both the methodology and the interpretation of results.
| Gap Type | What It Measures | Typical US Figure | What It Tells You |
|---|---|---|---|
| Unadjusted gender gap | Median pay difference between all men and all women in the organization | 16-20% | Reflects both occupational segregation and potential bias in pay decisions |
| Adjusted gender gap | Pay difference after controlling for job level, tenure, location, performance | 2-8% | Isolates unexplained disparity likely driven by bias in pay practices |
| Unadjusted racial/ethnic gap | Median pay difference between white employees and employees of color | 10-25% depending on group | Shows structural inequality in who holds which roles and levels |
| Adjusted racial/ethnic gap | Pay difference after controlling for legitimate factors | 2-7% | Reveals bias in compensation decisions for equivalent work |
| Intersectional gap | Pay difference for cross-demographic groups (e.g., Black women vs white men) | Often 15-35% unadjusted | Captures compounding effects of multiple forms of bias |
| Opportunity gap | Difference in representation at senior/high-paying levels by demographic group | Varies widely | Explains the structural factors behind the unadjusted gap |
How to Conduct a Pay Gap Analysis
A credible pay gap analysis requires clean data, sound methodology, and the right statistical approach for your organization's size and complexity.
Step 1: Prepare your data
Pull compensation data (base salary, total cash compensation, and if applicable, equity grants and bonus payouts) for all employees. Join it with demographic data (gender, race/ethnicity), job data (family, level, function), and employment data (hire date, tenure, location, performance ratings). Clean the data: standardize job titles and levels, remove duplicates, resolve missing values. Data preparation typically takes 40-50% of the total project time. If your HRIS data is messy, the analysis results won't be reliable no matter how good the statistics are.
Step 2: Calculate unadjusted gaps
Start with the simple median comparison. What's the median pay for men vs women? For white employees vs each racial/ethnic group? Use median rather than mean because median isn't skewed by executive outliers. Calculate the gap as a percentage: (median male pay - median female pay) / median male pay x 100. Do this at the total organization level and by division. The unadjusted gap gives leadership the big picture and often creates the urgency needed to invest in deeper analysis.
Step 3: Run regression analysis for adjusted gaps
Use multiple linear regression with base salary or total compensation as the dependent variable. Independent variables should include: job family, job level, years of experience (or tenure), geographic market, most recent performance rating, and highest education level. Then add demographic variables (gender, race/ethnicity). If gender or race is a statistically significant predictor of pay (p < 0.05) after all legitimate factors are controlled, you have an unexplained gap. The regression coefficient tells you the dollar amount of the gap. For organizations with fewer than 200 employees, regression may not produce reliable results due to small sample sizes. In that case, use matched-pair cohort analysis instead.
Step 4: Analyze by segment
Don't rely solely on the company-wide regression. Break the analysis down by job family, level, and location. A company-wide gap of 3% might mask a 12% gap in engineering and no gap in operations. Segment analysis tells you where the problems are, which helps you target remediation. Also run the analysis for total cash compensation and equity separately, because gaps in base pay, bonuses, and equity grants may have different root causes.
Step 5: Identify root causes
Statistical analysis tells you the gap exists. Root cause analysis tells you why. Common causes include: lower starting salaries for women (negotiation gaps or anchoring to salary history), smaller annual raises for underrepresented groups, slower promotion velocity that delays movement into higher pay bands, and market adjustments or retention raises that disproportionately benefit certain groups. Trace each identified gap back to the process that created it so your remediation targets the cause, not just the symptom.
Unadjusted vs Adjusted Pay Gap: Why Both Matter
The unadjusted and adjusted gaps tell different stories. Organizations that focus on only one miss half the picture.
The adjusted gap and equal pay for equal work
The adjusted gap answers: do we pay people fairly for the same or similar work? This is the legal standard under the Equal Pay Act and Title VII in the US, the Equality Act in the UK, and equivalent laws in most countries. When the adjusted gap is 5% or more, it usually means your pay-setting and pay-progression processes have bias embedded in them. Starting salaries, raise percentages, and promotion-linked increases are the typical culprits.
The unadjusted gap and structural equity
The unadjusted gap answers: do all groups have equal access to high-paying roles and levels? Even if your adjusted gap is zero (everyone in the same job gets the same pay), a large unadjusted gap means women or minorities are concentrated in lower-paying roles. That's an opportunity gap, and it points to issues in hiring, promotion, development, and retention. Investors, regulators, and the public increasingly look at the unadjusted gap because it reflects the full picture of economic inequality within the organization.
Closing the Pay Gap: Remediation Strategies
Finding the gap is the analytical challenge. Closing it is the operational one. Here's how organizations actually fix pay inequities once they've been identified.
Immediate pay corrections
For employees whose pay falls below the predicted range after controlling for legitimate factors, calculate the adjustment needed to bring them to parity. Budget 0.5% to 2% of total payroll for corrections. Make the adjustments in a single cycle rather than spreading them over multiple years. Notify affected employees individually, explaining that a proactive review identified their pay was below the expected level for their role and qualifications, and the company is correcting it. Don't call it a raise. It's a correction.
Fix the processes that created the gaps
Pay corrections without process changes are temporary fixes. Common process changes include: implementing structured offer calculations that base starting pay on job level and experience rather than salary history, narrowing salary band widths to limit pay dispersion, requiring compensation committee review for all offers below or above the midpoint, equalizing raise and bonus allocation by requiring managers to justify any difference in increase percentages between employees at the same level, and auditing promotion criteria to ensure they're applied consistently across groups.
Monitor continuously
Run the analysis annually at minimum. Some organizations run it quarterly as part of their compensation planning process. Set alerts in your HRIS or compensation platform to flag new hires, promotions, or raises that fall outside expected ranges by demographic group. Continuous monitoring catches new gaps before they compound. It's much cheaper to prevent a gap than to remediate one that's been growing for five years.
Pay Gap Statistics [2026]
Current data on pay gaps globally, nationally, and at the organizational level.
Legal Requirements by Region
Pay gap reporting is mandatory in a growing number of jurisdictions. Understanding your obligations prevents compliance surprises.
| Jurisdiction | Requirement | Who Must Comply | Reporting Frequency |
|---|---|---|---|
| UK | Publish mean and median gender pay gap, bonus gap, and pay quartile distribution | Employers with 250+ employees | Annually by April 4 (private sector) |
| EU (Directive 2023/970) | Report gender pay gap; joint assessment required if gap exceeds 5% | Employers with 100+ employees (phased rollout) | Annually for 250+; every 3 years for 100-249 |
| Iceland | Obtain Equal Pay Certification proving pay equity | Employers with 25+ employees | Every 3 years (certification renewal) |
| California (SB 1162) | Submit pay data report with median and mean hourly rates by demographic group | Employers with 100+ employees or 100+ contract workers | Annually by second Wednesday of May |
| Illinois (SB 1480) | Submit EEO-1 equivalent data and equal pay compliance statement | Employers with 100+ employees | Annually (with business registration renewal) |
| Australia | Report gender pay gap publicly | Employers with 100+ employees | Annually |
Tools for Pay Gap Analysis
The right tool depends on your organization's size, complexity, and analytical capability.
Spreadsheet-based analysis
For organizations under 500 employees with straightforward job structures, Excel or Google Sheets with pivot tables and basic statistical functions can produce a credible pay gap analysis. You'll need comfort with MEDIAN, PERCENTILE, and basic regression (using the Analysis ToolPak in Excel or LINEST functions). The limitation is scalability: once you exceed 500 employees or need intersectional analysis across multiple dimensions, spreadsheets become unwieldy and error-prone.
Statistical software
R, Python (with pandas and statsmodels), or SPSS provide the statistical power for regression analysis and handle large datasets well. These require someone with data analysis skills on staff or consulting support. The advantage is flexibility: you can customize the model, run sensitivity analyses, and produce publication-quality outputs. Many compensation consulting firms use R or Python as their primary analysis tool.
Dedicated pay equity platforms
Syndio, Trusaic, Pihr, and PayParity offer purpose-built platforms for pay equity analysis. They connect to your HRIS, automate data preparation, run the statistical models, and generate reports ready for compliance filing or executive presentation. Costs range from $20,000 to $100,000+ per year depending on employee count and features. These platforms are worth it for organizations with 1,000+ employees, complex job architectures, or multi-country operations where manual analysis would be prohibitively time-consuming.
Frequently Asked Questions
What's the difference between the gender pay gap and the equal pay gap?
How often should we run a pay gap analysis?
Is a 3% adjusted gap acceptable?
Does pay gap analysis have to account for negotiation differences?
Can we conduct a pay gap analysis without employee demographic data?
What do we do if the pay gap analysis reveals no significant gaps?
Written by Adithyan RK