Evenpay calculates two pay gap metrics: the unadjusted (raw) gap and the adjusted gap. Understanding the difference between them — and knowing what drives each one — is the foundation of effective pay equity work. This article explains how the analysis works, what the numbers mean, and what to pay attention to when reading your results.

Every pay equity analysis in Evenpay produces two numbers. They tell you different things, and both matter.

Unadjusted (raw) pay gap — the straightforward difference in average pay between men and women. No explanation of why the gap exists — just the number. This is the figure required for EU Pay Transparency Directive reporting.

Adjusted pay gap — the portion of the gap that remains after controlling for legitimate pay factors like job grade, experience, and education. This is the number that tells you what actually needs investigating.

A raw gap of 12% might sound alarming — but if 8 percentage points are explained by differences in job grade, experience, and tenure, the unexplained gap is 4%. That 4% is what requires action. The adjusted gap separates the signal from the noise.

Evenpay uses both mean-based and median-based calculations depending on the context. The right formula depends on what you're comparing:

A statistical significance test (t-test) is run for all gaps to confirm whether the observed difference is real or could be due to chance.

Evenpay calculates unadjusted pay gaps at five levels. Each serves a different purpose — use them together for a complete picture.

Grade-based gaps are the most important for compliance purposes. Job levels are mapped to grades to define what "work of equal value" means in your organization — for example, IC3, M2, and P2 might all map to the same grade.

The adjusted gap uses the Blinder-Oaxaca decomposition — an internationally recognized standard for pay equity analysis. It works in three steps.

Step 1: Salary transformation

Salaries are log-transformed before analysis. This is standard practice in labor economics because salary differences are proportional (percentage-based), not absolute. It also means the model's outputs can be read directly as approximate percentage effects.

Step 2: Regression model

Evenpay runs a multiple regression using your enabled pay factors — job grade, experience, tenure, and so on — to model how salaries are distributed across the organization. Gender is deliberately not included in the model. It's only used afterwards to compare outcomes between groups.

Step 3: Decomposition

The total pay gap is split into two parts:

The decomposition trains three separate models — one on men, one on women, and a pooled reference on all employees — to calculate explained and unexplained portions fairly.

Good to know: Factor contributions can be negative. If women in your organization have more experience on average than men, experience actually counteracts the gap rather than contributing to it.

Beyond the overall split between explained and unexplained, Evenpay uses Shapley values to show exactly how much each individual factor contributes to the difference between the raw and adjusted gap.

Think of it this way: if your raw gap is 12% and your adjusted gap is 4%, something accounts for that 8 percentage point difference. The factor contribution breakdown tells you how much of that 8pp comes from job grade, how much from experience, how much from education, and so on. The values always sum to the total adjustment, giving you a complete and consistent breakdown.

This is where the analysis becomes actionable — you can see exactly which structural differences in your workforce are explaining pay differences, and decide whether each one is a genuine business reason or something worth addressing.

The Factor Statistics panel shows each pay factor's coefficient, p-value, impact on the gap, and data quality rating:

Organizations choose which factors count as legitimate pay determinants through the Pay Philosophy settings in Evenpay. Each factor can be independently enabled or disabled. Only enable factors that genuinely drive pay decisions in your organization.

You can also add custom factors based on your organization's own fields if your compensation philosophy includes criteria beyond the defaults.

Categorical factors (like organization unit and job family) are encoded as separate comparison groups in the model. Small categories are automatically merged to keep the model stable. Continuous and ordinal factors go directly into the regression. If a factor has too much missing data, it's automatically disabled for that analysis run — you'll see which factors were excluded and why.

You control exactly what goes into the salary figure used for analysis:

Employees with zero or invalid salary data are automatically excluded. Evenpay supports two salary modes: Agreed salary (current contractual salary — the default for standard reporting) and Historical salary (for tracking how gaps evolve over time using stored snapshots).

Evenpay offers two modes depending on how much context you want to bring into the analysis.

Statistical Mode (default) — Pure Blinder-Oaxaca regression based on your data and enabled pay factors only. No manual adjustments. This is the right mode for standard compliance reporting — it gives you a clean, objective baseline.

Justification-Aware Mode — This mode incorporates documented justifications: approved business reasons for individual pay differences recorded by managers or HR. The system adjusts salaries by justified amounts before running the regression, then shows what gap remains after both statistical factors and documented reasons are accounted for. It tracks documentation coverage, addressed vs. unaddressed gaps, and overall compliance readiness.

Here's how that looks in practice. Say you have an employee whose salary sits above the expected range for their level, contributing to the pay gap. Once you document a justified business reason — for example, a specialized certification — that amount shifts from unexplained to explained:

The raw gap never changes — it's always the real number. What changes is how much of it is accounted for.

At the grade level, Evenpay also calculates a justification-adjusted gap: the raw gap minus documented justifications, floored at zero. This gives you a combined operational view that's useful for tracking remediation progress over time.

Evenpay flags gaps automatically at configurable severity levels: low, medium, high, and critical. You set the percentage thresholds for each level and define the minimum group sizes required for a comparison to be flagged. This means you're always aware of the most pressing issues without having to check the analysis manually.

Every analysis includes quality indicators so you can assess how reliable the results are. You don't need to be a statistician to use these — the key question is simply: does this model explain pay well enough to trust the adjusted gap?

The model quality panel is accessible by expanding Show model transparency at the bottom of your analysis results:

This opens the full Model Performance breakdown:

Each individual factor also reports its own coefficient, significance level, and confidence interval — so you can see exactly which factors are carrying weight in the model and which are not.

If your R² is low, the most common cause is a missing factor. Job grade is almost always the biggest driver of pay — if it's not included or not assigned consistently, the model won't explain much.

For the analysis to run, you need:

If you don't have enough data for the full decomposition, Evenpay will tell you why and continue with the prediction model alone. Factors with too much missing data are automatically disabled for that run — you'll always see which factors were excluded and how many employees were omitted from the analysis.

Evenpay's methodology is designed to meet and exceed the requirements of the FPI Universal Fair Pay Check, which requires an adjusted gender pay gap calculation based on multiple regression using the Oaxaca-Blinder method. FPI does not prescribe a specific formula, transformation, or variable set — Evenpay's approach is fully compatible.

For the EU Pay Transparency Directive, Evenpay supports the required reporting on gender pay gaps and comparison of pay between workers doing the same work or work of equal value. The adjusted analysis is an additional method that supports deeper investigation of underlying drivers.

Features like factor contribution attribution, cost-to-fix estimation, model diagnostics, and justification-aware analysis are Evenpay's own analytical enhancements that go beyond what any standard or directive requires.

Now that you understand how the analysis works, here's where to go next:

Which number should I focus on — the unadjusted or adjusted gap?

Both, but for different purposes. The unadjusted gap is required for EU Pay Transparency Directive reporting and gives you the headline number. The adjusted gap tells you what actually needs to be addressed — it's the part of the gap that your legitimate pay factors can't explain. If your adjusted gap is close to zero, your pay structure is working equitably. If it's significant, that's where to focus.

What if my adjusted gap is higher than my unadjusted gap?

This can happen. It means that your pay factors are actually masking a larger underlying gap — for example, if women in your organization tend to be in higher job grades on average than men at similar pay levels. It's unusual but not impossible, and it's important information worth investigating.

My R² is below 0.50. What should I do?

A low R² usually means the model is missing an important factor. The first thing to check is whether job grade is enabled and whether all employees have a job level assigned — it's almost always the strongest predictor. After that, consider whether experience, performance, or another factor is genuinely driving pay in your organization and isn't yet reflected in the model.

When should I use Justification-Aware Mode vs. Statistical Mode?

Use Statistical Mode for compliance reporting — it gives you a clean, objective baseline with no manual inputs. Use Justification-Aware Mode when you've already documented business reasons for specific pay differences and want to see what gap remains after those justifications are accounted for. It's particularly useful for tracking progress as you document and address gaps over time.

How do I reduce my adjusted gap?

There are two levers. The first is to address actual pay inequities — reviewing and correcting salaries where the gap is unjustified. The second is to improve data completeness: if key factors like job grade, experience, or performance are missing for many employees, the model can't explain pay differences that are actually justified, which inflates the adjusted gap. Start by checking factor coverage and fixing data gaps before making pay adjustments.

Why are some employees excluded from my analysis?

Employees are excluded when they're missing data the analysis requires: no salary on record, no gender assigned, or no job level. You'll always see a count of excluded employees and the specific reason for each. The two most common gaps are missing job levels and missing salary data — fixing those will maximize your analysis coverage.