library / lib78f019157e8257df
Compensating Employees Fairly
In a sentence
A practical, statistically grounded guide to measuring and achieving internal pay equity so employers can detect, prevent, and remedy compensation discrimination before it becomes litigation.
Written by econometrician and pay-equity consultant Stephanie R. Thomas, this book demystifies the statistical analysis of compensation for fairness, showing HR professionals, business leaders, and legal counsel how to examine whether employees are paid equitably irrespective of gender, race, or other protected characteristics. It walks readers through the legal landscape of equal pay law, the two theories of discrimination (disparate treatment and disparate impact), and the full toolkit of statistical methods — multiple regression, t-tests, cohort analysis, four-fifths rule, chi-square, Fisher's exact, Mantel-Haenszel, and logistic regression — while emphasizing rigorous data preparation, correct employee groupings, and disciplined follow-up. Crucially, it reframes the conversation away from the misleading '77 cents' raw wage gap toward properly specified models that account for legitimate, nondiscriminatory factors, and it makes the business case for proactive self-analysis as a litigation-avoidance and competitive-advantage strategy. Readers come away able to recognize the difference between statistical and practical significance, between correlation and causation, and between merely achieving statistical insignificance and achieving true equity.
The four lenses
- Science
- Statistics
- Systems
- Strategy
The model
A causal/path model in which compensation design levers and data conditions shape the validity of statistical analysis, which in turn determines detected pay disparities, perceived fairness, and outcomes such as litigation risk and equity.
Compensation Policy Clarity and Consistencydesign lever
The degree to which compensation decisions are based on a consistent, well-articulated, documented set of objective and measurable criteria rather than arbitrary or discretionary judgment.
Compensation Decision Documentation and Retentiondesign lever
The extent to which the rationale, metrics, surveys, performance reviews, and special circumstances behind each compensation decision are recorded contemporaneously and retained for future justification and analysis.
Similarly Situated Employee Grouping Accuracydesign lever
The correctness with which employees who perform similar work at similar responsibility levels requiring similar skills and qualifications are grouped together for comparison purposes, the most important determinant of valid analysis.
Data Completeness, Accuracy, and Consistencycontextual condition
The degree to which the compensation data set is comprehensive, machine-readable, internally consistent in units, free of measurement error, and captures full cradle-to-grave employment histories for all relevant employees.
Statistical Model Specification Qualitybehavioral pattern
The extent to which the regression or statistical model correctly approximates the actual compensation decision process by including all relevant legitimate factors, excluding irrelevant ones, and satisfying the assumptions of the classical linear regression model.
Detected Pay Disparity by Protected or Overall Statusoutcome metric
The estimated, statistically and practically evaluated difference in compensation associated with protected group status or relative to similarly situated comparators after accounting for legitimate factors.
Analysis Follow-Up Rigorbehavioral pattern
The thoroughness with which flagged disparities are investigated through deeper data review, file examination, and discussions with managers to identify legitimate explanations before any remediation.
Employee Perceived Fairnesspsychological state
Employees' perception of the fairness of compensation outcomes, the procedures used, and the transparency of information communicated, encompassing distributive, procedural, interactional, and informational justice.
Internal Pay Equityoutcome metric
The state in which similarly situated employees performing similar work at similar responsibility levels requiring similar skills are paid similarly irrespective of protected characteristics, the central outcome the book seeks.
Litigation and Regulatory Investigation Riskoutcome metric
The organization's exposure to compensation discrimination lawsuits, regulatory audits, damages, penalties, and legal costs arising from unaddressed pay inequities.
Talent Retention, Engagement, and Productivityoutcome metric
The competitive-advantage outcomes attributed to fair compensation, including attraction and retention of top talent, reduced turnover and absenteeism, and improved engagement, motivation, and productivity.
How they connect
- compensation policy clarity → influences internal pay equity
- compensation documentation − influences litigation and regulatory risk
- employee grouping accuracy → influences model specification quality
- data quality → influences model specification quality
- model specification quality → predicts detected pay disparity
- analysis followup rigor − moderates detected pay disparity
- detected pay disparity − predicts internal pay equity
- detected pay disparity → predicts litigation and regulatory risk
- internal pay equity → influences perceived fairness
- internal pay equity − predicts litigation and regulatory risk
- internal pay equity → influences workforce outcomes
- perceived fairness → influences workforce outcomes
A candidate measure
Compensating Employees Fairly — derived measurement candidates
Compensation Policy Clarity and Consistency
policy documentation completeness score; percent of decisions tied to defined criteria; manager training completion rate
self-report suitability: medium
Compensation Decision Documentation and Retention
percent of decisions with supporting documentation; retention compliance rate; retrieval time per record
self-report suitability: medium
Similarly Situated Employee Grouping Accuracy
inter-rater agreement on groupings; percent of groupings validated by job analysis
self-report suitability: low
Data Completeness, Accuracy, and Consistency
missing-data percentage; unit-consistency error count; proportion of truncated histories
self-report suitability: low
Statistical Model Specification Quality
adjusted R-squared; diagnostic test outcomes; specification rationale documentation
self-report suitability: none
Detected Pay Disparity by Protected or Overall Status
coefficient magnitude and sign; p-value/t-statistic; standard deviation units; four-fifths ratio
self-report suitability: none
Analysis Follow-Up Rigor
percent of flagged groupings investigated; number of follow-up steps documented per flag
self-report suitability: medium
Employee Perceived Fairness
distributive justice perception; procedural justice perception; informational justice/transparency perception
self-report suitability: high
Internal Pay Equity
share of groupings without unexplained disparity; aggregate residual dispersion
self-report suitability: low
Litigation and Regulatory Investigation Risk
number of claims/audits per period; settlement and penalty dollars; legal defense expenditures
self-report suitability: low
Talent Retention, Engagement, and Productivity
turnover rate; absenteeism rate; engagement survey index; output/productivity metrics
self-report suitability: medium
The story
The reader An employer, HR/compensation professional, or in-house counsel who wants to ensure employees are paid fairly and to protect the organization from compensation discrimination claims.
External problem
Compensation practices may contain undetected pay inequities that expose the organization to litigation and regulatory investigation.
Internal problem
The reader feels uncertain and anxious about whether their pay decisions are defensible and fears costly surprises they cannot see coming.
Philosophical problem
It is simply wrong to pay similarly situated people unequally, and equally wrong to assume fairness without ever measuring it.
The plan
- Understand the legal theories and dimensions of fairness.
- Learn multiple regression and other statistical tools.
- Build correct similarly situated employee groupings and clean data.
- Estimate appropriate models and interpret results carefully.
- Follow up on flagged disparities and remediate under legal counsel.
- Institutionalize proactive self-analysis as ongoing risk management.
Success
- Defensible, transparent pay practices and reduced litigation and regulatory exposure.
- Genuine internal equity that improves retention, engagement, and competitive advantage.
- Confidence that pay decisions can be explained and justified.
At stake
- Undetected inequities surface as costly lawsuits, settlements, damages, and legal fees.
- Reputational harm and regulatory penalties from unaddressed discrimination.
- Loss of talent and competitive disadvantage from unfair compensation.
Related in the literature
The measurement literature behind this signal — sourced, so you can defend it.
“A review of the pay increases granted to the remaining people in this similarly situated employee grouping reveals a similar pattern.This, coupled with the views he expressed regarding women in the warehouse, indicates that the likely cause of the observed disparity in hourly…”
— Compensating Employees Fairlymatch 63%
“The difference of $2,080 annually ($2,080 = $30,888 – $28,808) is 7.2% of the woman’s annual salary (0.072 = $2,080 / $28,808). 17Systemic disparate treatment is discussed in Chapter 2. 18For example, the similarly situated employee grouping may contain too few employees to…”
— Compensating Employees Fairlymatch 61%
“For smaller companies, a pragmatic way of reviewing performance and linking it to compensation decisions is for the direct manager to suggest increases and then discuss and approve them in a “compensation committee.” CEO, CFO, and CPO should be permanent members. Invite…”
— Scaling Up Compensationmatch 61%
Resources: Compensating Employees Fairly · Scaling Up Compensation