Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus convallis sem tellus, vitae egestas felis vestibule ut.

Error message details.

Reuse Permissions

Request permission to republish or redistribute SHRM content and materials.

Using Data to Help Close the Gender Wage Gap

Two business people looking at a tablet in an office.

Remedying pay disparities between men and women has become a priority for businesses and lawmakers alike. How can HR professionals find out if unlawful pay discrepancies exist in their workplace?

Data analytics can use an organization's employment information to identify wage disparities and, from there, companies can look at their hiring practices and other policies to resolve unwanted pay gaps, according to Zev Eigen, J.D., Ph.D., the global director of data analytics for law firm Littler Mendelson. He noted that some employers may not know that women in their workforce are earning less than their male counterparts until they conduct this type of analysis.

[SHRM members-only toolkit: Managing Pay Equity]

Much of the information employers need for a pay audit is already housed in their human resource information system (HRIS) or other electronic systems, said Lara de Leon, an attorney with Ogletree Deakins in Orange County, Calif., and San Antonio.

She said that the first step in the analysis is for employers to make sure they have all the data they need and that the information is up to date and accurate. Some key data fields are:

  • Basic employment status and historical employment information.
  • Demographic information on gender, race and other categories.
  • Job title, job level, overtime exemption status, full-time or part-time status, and base pay. 
  • Geographic work location and business unit information. 
  • Quantity or quality of work, other performance measures, or a seniority system.

De Leon said it's also helpful to have pay ranges and any other compensation data on hand when it comes time to analyze the results.

Complex Analysis

A statistical analysis can help employers find out if pay differences are random or caused by some other factor, Eigen said. He noted, however, that not all intentional pay discrepancies are discriminatory. There are justifiable reasons for paying one worker more than another in the same job grouping—such as seniority, education level or years of experience.

Data can help employers figure out the root cause of a pay discrepancy, he said, noting that such an evaluation may be easier for smaller employers that have fewer factors and job groupings to consider.

Large employers may want to perform a multiple regression analysis, which is a statistical method that can account for a variety of factors that influence pay. Employers can build a model that looks at workers' geographic location, performance ratings, time at the job, seniority with the company, education and certification data, and other factors that may be relevant for the particular workplace or job grouping.

Although education can be a factor that justifies a pay discrepancy, Eigen cautioned that it isn't always relevant to the job. For example, he said, having a Ph.D. in molecular gastronomy might make sense for a scientist in a food research lab, but it might not matter for a fast-food chain manager.

It's also important for employers to identify objective measures of skills, effort and accountability for each job, he noted. "If sales people need to be good communicators, how is that measured objectively?"

Some employees may earn a higher salary because they have more direct reports or interact with more team members and business partners. These factors should also be identified.

"It takes a little bit of work on the front end to make sure you're measuring the right groupings and comparing apples to apples," Eigen said.

After the appropriate variables are identified and taken into consideration, the analysis will point to whether a difference in pay may be justified.

That's why employers need to be sure the data in their systems are accurate and updated. "As with anything, if your data is not sound, your results won't be accurate," de Leon said. "An audit with a significant amount of data inconsistencies could yield erroneous results—either fail to uncover an issue, or indicate that there is an issue where one really does not exist."

Outside Help

Eigen urged employers to avoid assigning the analysis to someone within the company, unless that employee has experience with this type of statistical evaluation and knowledge of the relevant pay equity laws that apply to the employer's workforce.

"You could pull a tooth out yourself, but there are obvious advantages to going to a dentist," he said.

Multiple regression analyses are complex, and many employers do not have someone in-house who can perform them, de Leon said. "You will want to work with legal counsel and, most likely, an external consultant or statistician to help complete the work."

De Leon and Eigen both noted that conducting the analysis under attorney-client privilege can potentially protect the information from discovery during a lawsuit.

HR's Role

HR professionals should understand that a multiple regression analysis is different than other analyses performed for pay reporting or auditing obligations—such as for the new EEO-1 reporting requirements, de Leon said.

Since different analyses can yield different results, employers should perform a separate audit specifically to address potential pay discrimination. However, businesses shouldn't conduct an audit unless they are prepared to correct any unjustifiable disparities that are discovered.

Some fixes might be costly, but others could be as simple as reclassifying a star employee. "If Joe is earning more than other employees in his workgroup, it's possible that Joe isn't correctly assigned to that group," Eigen said. "Maybe he is doing a lot of other tasks that he took over as people left the company and a change to his job title and reporting structure is appropriate."

It's better to fix those discrepancies before a claim is filed, he added. If the change is made after litigation commences, it might look like the employer is just trying to escape liability.

It's also a good practice for employers to implement policies that address pay discrimination and establish guidelines that set forth their compensation philosophy, de Leon said. For example, employer guidelines should say whether compensation decisions are based on tenure, the quantity or quality of work, or other factors. 

"Employers should also look at how they set starting pay and implement guidelines for that," she noted. Some states have recently passed laws restricting inquiries into a job applicant's salary history—and more states are expected to follow suit.


Was this article useful? SHRM offers thousands of tools, templates and other exclusive member benefits, including compliance updates, sample policies, HR expert advice, education discounts, a growing online member community and much more. Join/Renew Now and let SHRM help you work smarter.


​An organization run by AI is not a futuristic concept. Such technology is already a part of many workplaces and will continue to shape the labor market and HR. Here's how employers and employees can successfully manage generative AI and other AI-powered systems.