Strah et al. (Reference Strah, Rupp and Morris2021) did a nice job of discussing the case law related to the similarity of two jobs as well as discussing how job analysis methods can be used to compare the similarity of two or more jobs. As demonstrated by Strah et al., most of the Equal Pay Act (EPA) and Title VII litigation regarding employee pay is a battle between the plaintiff’s claim that two jobs are substantially equal or substantially similar and the defendant’s claim that two jobs are different enough that they are not substantially equal or similar. The standard used by most courts for substantially equal or similar is narrow (Sady & Aamodt, Reference Sady and Aamodt2016). For example, in Finke v. Trustees of Purdue University (2014), deans in different schools were not considered to be substantially similar; in Jacobs v. College of William and Mary (1980), athletic coaches were not considered to be substantially equal; and in Spaulding v. University of Washington (1984), faculty in different schools were not considered to be substantially equal.
In audits conducted by the Office of Federal Contract Compliance Programs (OFCCP), however, it is our experience that the battle in these audits is quite different than in litigation. To create pay analysis groups large enough for regression analysis, OFCCP often combines jobs that are not substantially similar and, based on the reviews of Strah et al. (Reference Strah, Rupp and Morris2021) and Sady and Aamodt (Reference Sady and Aamodt2016), would not be considered substantially similar by the court if the case went to litigation.
In OFCCP audits, there are three main sources of disagreement between the OFCCP and the federal contractor regarding job similarity and sex or race/ethnicity differences in compensation:
-
1. Must the jobs being compared be substantially similar (contractor’s position) or can dissimilar jobs be combined and the areas of dissimilarity (e.g., job title, salary grade, job family) be controlled for in a regression analysis (OFCCP’s position)?
-
2. Are the jobs being combined by OFCCP so dissimilar that combining them not only violates EPA and Title VII concepts of being similarly situated but violates common business and scientific (e.g., job analysis) sense?
-
3. Are the factors that drive pay similar across all the jobs being combined? That is, does having a graduate degree have similar worth for each of the jobs being combined? Does each year of relevant experience have similar worth for each of the jobs being combined?
If only we had the time and journal space to thoroughly discuss each of these three sources of disagreement, but we don’t. Thus, our commentary will focus on the second area: What makes two jobs so dissimilar that they clearly can’t be combined in an audit. It is important to note that this is a very different question than that addressed by Strah et al. (Reference Strah, Rupp and Morris2021). Whereas Strah et al. focus on using job analysis to distinguish in litigation between two jobs that might be substantially similar, we will focus on factors that would clearly indicate that two or more jobs are too dissimilar to be combined in an OFCCP audit.
The OFCCP audit process
During an audit, a federal contractor will provide data related to compensation for each employee. This set of data, referred to as Item 19, includes such fields as an employee’s sex, race, salary, job group, job title, salary grade, job family, hire date, and job date. From these data, OFCCP will create similarly situated employee groupings (SSEGs) that contain a sufficient number of employees to run a regression analysis. Because there are often not enough employees in a particular job title to provide adequate statistical coverage, OFCCP will combine job titles into larger groups such as Equal Employment Opportunity (EEO) job group (e.g., professionals, trades), job family, or related title (e.g., accountants in all accounting titles and grades). Once the SSEGs have been created, OFCCP will run a series of regression analyses to determine whether there are statistically significant sex or race/ethnicity differences after controlling for the factors thought to be related to pay (e.g., experience, performance). If any statistically significant differences are found, OFCCP asks the contractor to respond. It is during this response that the contractor will often challenge the OFCCP’s SSEGs as containing employees who are not similarly situated as required by Title VII. Moreover, during this audit, the agency and contractor are cognizant of OFCCP’s definition of similarly situated employees. In OFCCP Directive 2018-05, the agency defines similarly situated employees as “those who would be expected to be paid the same based on: (a) job similarity (e.g., tasks performed, skills required, effort, responsibility, working conditions and complexity); (b) other objective factors such as minimum qualifications or certifications.”
OFCCP is required to implement the audit process described above across a wide range of industries and organizations. To meet this requirement, the aforementioned aggregation of job titles may be used in part because it may be too time consuming to perform highly detailed, individualized employee comparisons at every organization; however, during this aggregation process, the agency often creates SSEGs that appear not to meet the OFCCP similarly situated definition described above. A question ultimately arises: At what point does efficiency for statistical analysis purposes transform into a vice where one is sacrificing job analysis and human resources (HR) best practice in order to create groupings large enough for statistical analyses? To answer that question, it may help if professionals could agree on standards to determine when jobs are clearly not similarly situated. For example, would most professionals agree that even if one has an admirable desire for efficiency, jobs in different job families (e.g., finance, engineering) are too different to ever be considered similarly situated?
Differences between jobs
In creating SSEGs, industrial-organizational (I-O) psychologists and labor economists usually begin with existing job information created by an organization.Footnote 1 This information can include job titles, salary grades, job families, job subfamilies, and job functions. The question is, does the combining of any elements of these types of information automatically violate the substantially similar standard?
Job titles
At first glance, it would seem that employees in the same job title perform substantially similar work. It is important, however, to note that even employees in the same title may not be similarly situated (Sady et al., Reference Sady, Aamodt, Cohen, Hanvey and Sady2015). Take for example, two employees sharing the job title of Software Developer II. One writes code for a new web-based application, whereas the other writes code to increase the security of an existing application. Although both are writing code, the programming languages used, the complexity of the project, and the development versus maintenance aspects may make the two jobs too dissimilar to compare to one another.
Furthermore, organizations vary in their approaches to creating and using job titles. The extent to which employees in the same title perform substantially similar work is in part a function of how the organization creates job titles. For organizations that use broad titles (e.g., accountant, programmer) or that have titles that cross salary grades, the title alone is not sufficient to infer substantially similar. For organizations that use more narrow titles (e.g., cost accountant, tax accountant) that are grade specific (e.g., Cost Accountant I, Cost Accountant II), the title alone may be sufficient to infer substantially similar.
Salary grades
When constructed properly, salary grades (also called levels, tiers, bands) can provide information about the complexity level of the work performed. For example, a Cost Accountant IV in grade D performs substantially more complex work than a Cost Accountant I in grade A. Given that salary grades represent different levels of complexity, would combining titles in different grades automatically violate OFCCP’s definition of similarly situated? Is this especially true when the contractor’s compensation uses broad bands rather than narrow salary grades?
Job families and subfamilies
Job families provide information about the general field of knowledge that the employees perform (e.g., finance, human resources, engineering). Subfamilies further refine that field (e.g., audit and accounting for the finance family). Given that the knowledge base for a family such as finance is very different from a family such as engineering, would combining titles in different job families automatically violate OFCCP’s definition of similarly situated because they don’t share similar knowledge or skill requirements? Are there some job families that could reasonably be combined?
Job functions
Job functions provide general information about the type of work performed by employees (e.g., managers, analysts, clerks). Job functions typically cross job family (e.g., financial analysts, human resource analysts). Many organizations begin with two broad functions—managers and individual contributors—and then develop more narrow functions within these two broad groups. Could managers and individual contributors ever be considered similarly situated? Could analysts and clerks ever be considered similarly situated?
EEO job groups
For affirmative action purposes, federal contractors are required to classify each job title into one of 10 broad EEO job groups. These 10 job groups are
-
1. Executive/senior-level officials and managers
-
2. First/midlevel officials and managers
-
3. Professionals
-
4. Technicians
-
5. Sales workers
-
6. Administrative support workers
-
7. Craft workers
-
8. Operative
-
9. Laborers and helpers
-
10. Service workers
Because these job groups must be submitted during an audit, it is not unusual for OFCCP to use EEO job group for its initial SSEG creation. Because EEO job groups cross title, salary grade, family, and function, could they ever meet the legal standard for similarly situated or substantially similar?
Final thoughts
Reasonable people can disagree about whether two jobs are substantially similar. The question we pose, however, is are there characteristics of jobs (e.g., grades, families, functions) that can’t be crossed if two or more jobs are to be considered from a legal perspective as being similar enough to compare the pay of employees? If so, what are those characteristics and who should identify them: I-O psychologists? Lawyers? Federal agencies? The courts?
I-O psychologists are uniquely qualified to apply the scientific method to measure whatever characteristics end up being of interest in this context. We hope that the field can work toward developing basic standards for studying jobs in the context of pay equity that would provide decision makers with information about job characteristics necessary to evaluate job similarity. Even if professionals are hesitant to define standards indicating what job characteristics cannot be crossed or if they believe that the answers to the questions above will likely depend on context, can the field embrace certain presumptions? For example, can professionals agree on a rebuttable presumption that jobs from different families are not similarly situated? The entity looking to combine the families would have the burden of establishing that there is in fact similarity. If this cannot be achieved, then the presumption of no similarity would hold.
You have
Access
Strah et al. (Reference Strah, Rupp and Morris2021) did a nice job of discussing the case law related to the similarity of two jobs as well as discussing how job analysis methods can be used to compare the similarity of two or more jobs. As demonstrated by Strah et al., most of the Equal Pay Act (EPA) and Title VII litigation regarding employee pay is a battle between the plaintiff’s claim that two jobs are substantially equal or substantially similar and the defendant’s claim that two jobs are different enough that they are not substantially equal or similar. The standard used by most courts for substantially equal or similar is narrow (Sady & Aamodt, Reference Sady and Aamodt2016). For example, in Finke v. Trustees of Purdue University (2014), deans in different schools were not considered to be substantially similar; in Jacobs v. College of William and Mary (1980), athletic coaches were not considered to be substantially equal; and in Spaulding v. University of Washington (1984), faculty in different schools were not considered to be substantially equal.
In audits conducted by the Office of Federal Contract Compliance Programs (OFCCP), however, it is our experience that the battle in these audits is quite different than in litigation. To create pay analysis groups large enough for regression analysis, OFCCP often combines jobs that are not substantially similar and, based on the reviews of Strah et al. (Reference Strah, Rupp and Morris2021) and Sady and Aamodt (Reference Sady and Aamodt2016), would not be considered substantially similar by the court if the case went to litigation.
In OFCCP audits, there are three main sources of disagreement between the OFCCP and the federal contractor regarding job similarity and sex or race/ethnicity differences in compensation:
1. Must the jobs being compared be substantially similar (contractor’s position) or can dissimilar jobs be combined and the areas of dissimilarity (e.g., job title, salary grade, job family) be controlled for in a regression analysis (OFCCP’s position)?
2. Are the jobs being combined by OFCCP so dissimilar that combining them not only violates EPA and Title VII concepts of being similarly situated but violates common business and scientific (e.g., job analysis) sense?
3. Are the factors that drive pay similar across all the jobs being combined? That is, does having a graduate degree have similar worth for each of the jobs being combined? Does each year of relevant experience have similar worth for each of the jobs being combined?
If only we had the time and journal space to thoroughly discuss each of these three sources of disagreement, but we don’t. Thus, our commentary will focus on the second area: What makes two jobs so dissimilar that they clearly can’t be combined in an audit. It is important to note that this is a very different question than that addressed by Strah et al. (Reference Strah, Rupp and Morris2021). Whereas Strah et al. focus on using job analysis to distinguish in litigation between two jobs that might be substantially similar, we will focus on factors that would clearly indicate that two or more jobs are too dissimilar to be combined in an OFCCP audit.
The OFCCP audit process
During an audit, a federal contractor will provide data related to compensation for each employee. This set of data, referred to as Item 19, includes such fields as an employee’s sex, race, salary, job group, job title, salary grade, job family, hire date, and job date. From these data, OFCCP will create similarly situated employee groupings (SSEGs) that contain a sufficient number of employees to run a regression analysis. Because there are often not enough employees in a particular job title to provide adequate statistical coverage, OFCCP will combine job titles into larger groups such as Equal Employment Opportunity (EEO) job group (e.g., professionals, trades), job family, or related title (e.g., accountants in all accounting titles and grades). Once the SSEGs have been created, OFCCP will run a series of regression analyses to determine whether there are statistically significant sex or race/ethnicity differences after controlling for the factors thought to be related to pay (e.g., experience, performance). If any statistically significant differences are found, OFCCP asks the contractor to respond. It is during this response that the contractor will often challenge the OFCCP’s SSEGs as containing employees who are not similarly situated as required by Title VII. Moreover, during this audit, the agency and contractor are cognizant of OFCCP’s definition of similarly situated employees. In OFCCP Directive 2018-05, the agency defines similarly situated employees as “those who would be expected to be paid the same based on: (a) job similarity (e.g., tasks performed, skills required, effort, responsibility, working conditions and complexity); (b) other objective factors such as minimum qualifications or certifications.”
OFCCP is required to implement the audit process described above across a wide range of industries and organizations. To meet this requirement, the aforementioned aggregation of job titles may be used in part because it may be too time consuming to perform highly detailed, individualized employee comparisons at every organization; however, during this aggregation process, the agency often creates SSEGs that appear not to meet the OFCCP similarly situated definition described above. A question ultimately arises: At what point does efficiency for statistical analysis purposes transform into a vice where one is sacrificing job analysis and human resources (HR) best practice in order to create groupings large enough for statistical analyses? To answer that question, it may help if professionals could agree on standards to determine when jobs are clearly not similarly situated. For example, would most professionals agree that even if one has an admirable desire for efficiency, jobs in different job families (e.g., finance, engineering) are too different to ever be considered similarly situated?
Differences between jobs
In creating SSEGs, industrial-organizational (I-O) psychologists and labor economists usually begin with existing job information created by an organization.Footnote 1 This information can include job titles, salary grades, job families, job subfamilies, and job functions. The question is, does the combining of any elements of these types of information automatically violate the substantially similar standard?
Job titles
At first glance, it would seem that employees in the same job title perform substantially similar work. It is important, however, to note that even employees in the same title may not be similarly situated (Sady et al., Reference Sady, Aamodt, Cohen, Hanvey and Sady2015). Take for example, two employees sharing the job title of Software Developer II. One writes code for a new web-based application, whereas the other writes code to increase the security of an existing application. Although both are writing code, the programming languages used, the complexity of the project, and the development versus maintenance aspects may make the two jobs too dissimilar to compare to one another.
Furthermore, organizations vary in their approaches to creating and using job titles. The extent to which employees in the same title perform substantially similar work is in part a function of how the organization creates job titles. For organizations that use broad titles (e.g., accountant, programmer) or that have titles that cross salary grades, the title alone is not sufficient to infer substantially similar. For organizations that use more narrow titles (e.g., cost accountant, tax accountant) that are grade specific (e.g., Cost Accountant I, Cost Accountant II), the title alone may be sufficient to infer substantially similar.
Salary grades
When constructed properly, salary grades (also called levels, tiers, bands) can provide information about the complexity level of the work performed. For example, a Cost Accountant IV in grade D performs substantially more complex work than a Cost Accountant I in grade A. Given that salary grades represent different levels of complexity, would combining titles in different grades automatically violate OFCCP’s definition of similarly situated? Is this especially true when the contractor’s compensation uses broad bands rather than narrow salary grades?
Job families and subfamilies
Job families provide information about the general field of knowledge that the employees perform (e.g., finance, human resources, engineering). Subfamilies further refine that field (e.g., audit and accounting for the finance family). Given that the knowledge base for a family such as finance is very different from a family such as engineering, would combining titles in different job families automatically violate OFCCP’s definition of similarly situated because they don’t share similar knowledge or skill requirements? Are there some job families that could reasonably be combined?
Job functions
Job functions provide general information about the type of work performed by employees (e.g., managers, analysts, clerks). Job functions typically cross job family (e.g., financial analysts, human resource analysts). Many organizations begin with two broad functions—managers and individual contributors—and then develop more narrow functions within these two broad groups. Could managers and individual contributors ever be considered similarly situated? Could analysts and clerks ever be considered similarly situated?
EEO job groups
For affirmative action purposes, federal contractors are required to classify each job title into one of 10 broad EEO job groups. These 10 job groups are
1. Executive/senior-level officials and managers
2. First/midlevel officials and managers
3. Professionals
4. Technicians
5. Sales workers
6. Administrative support workers
7. Craft workers
8. Operative
9. Laborers and helpers
10. Service workers
Because these job groups must be submitted during an audit, it is not unusual for OFCCP to use EEO job group for its initial SSEG creation. Because EEO job groups cross title, salary grade, family, and function, could they ever meet the legal standard for similarly situated or substantially similar?
Final thoughts
Reasonable people can disagree about whether two jobs are substantially similar. The question we pose, however, is are there characteristics of jobs (e.g., grades, families, functions) that can’t be crossed if two or more jobs are to be considered from a legal perspective as being similar enough to compare the pay of employees? If so, what are those characteristics and who should identify them: I-O psychologists? Lawyers? Federal agencies? The courts?
I-O psychologists are uniquely qualified to apply the scientific method to measure whatever characteristics end up being of interest in this context. We hope that the field can work toward developing basic standards for studying jobs in the context of pay equity that would provide decision makers with information about job characteristics necessary to evaluate job similarity. Even if professionals are hesitant to define standards indicating what job characteristics cannot be crossed or if they believe that the answers to the questions above will likely depend on context, can the field embrace certain presumptions? For example, can professionals agree on a rebuttable presumption that jobs from different families are not similarly situated? The entity looking to combine the families would have the burden of establishing that there is in fact similarity. If this cannot be achieved, then the presumption of no similarity would hold.