1

What a Government Agency Wants, What a Government Agency Needs…./�Your Big Data

Whitney R. Brown, Shareholder at Lehr Middlebrooks Vreeland & Thompson, wbrown@lehrmiddlebrooks.com

​

EEOC Investigative Trends

Big Data Requests

About EEOC Big Data Requests

• If a Charge is a strong Priority “A” Charge or otherwise being considered by the EEOC or advocated by a particular investigator for EEOC-led litigation, you will at minimum receive these requests.
• If the Charge contains systemic allegations—or the investigator believes systemic issues may be implicated—expect to receive a follow up request for the underlying data itself.

​

4

Sample EEOC RFI Request – � About Your HRIS/ATS

5

Sample EEOC RFI Request – � About Your Use of AI, Algorithms, etc.

6

Sample EEOC RFI Follow Up Request – � For Your ATS or HRIS Data

7

My Observations and Theories about EEOC Big Data Requests

• I believe the requests are drafted by EEOC e-discovery lawyers in D.C. There are likely templates that investigators can use off-the-shelf, but questions of compliance or interpretation must be resolved from “above.”
• There can also be loss of continuity as local investigators pull follow up (“Step 2”) templates that might refer to, for example, data fields which the Respondent has already discussed not existing with the EEOC.
• In my experience, the local B’ham and its Mobile satellite don’t have even pseudo-statisticians on staff.
• Have recently caught wind of “tech-to-tech” interviews where EEOC requests to interview the “most knowledgeable” person about HRIS. Normally, the EEOC’s interviewer is not actually a separate technician, but the investigator, with a list of questions (likely from D.C.).

8

What Can All This Data Mean?

What Data-Agile Federal Contractors and Subcontractors Already Know

What Federal Contractors Already Know: � Incumbency vs. Availability Analysis

41 C.F.R. §60-2.15

1. The contractor must compare the percentage of minorities and women in each job group determined pursuant to §60-2.13 with the availability for those job groups determined pursuant to §60-2.14.
2. When the percentage of minorities or women employed in a particular job group is less than would reasonably be expected given their availability percentage in that particular job group, the contractor must establish a placement goal in accordance with § 60-2.16.

41 C.F.R. §60-2.14

(a) Purpose: Availability is an estimate of the number of qualified minorities or women available for employment in a given job group…

(c) In determining availability, the contractor must consider at least the following factors:

(1) The percentage of minorities or women with requisite skills in the reasonable recruitment area.

(2) The percentage of minorities or women among those promotable, transferable, and trainable within the contractor's organization.

(g) Where a job group is composed of job titles with different availability rates, a composite availability figure for the job group must be calculated.

10

What Federal Contractors Already Know: � Incumbency vs. Availability Analysis

11

Source: OFCCP Sample AAP

https://www.dol.gov/sites/dolgov/files/ofccp/regs/compliance/AAPs/Sample_EO11246_AAP_final_01.03.18_Contr508.pdf

What Federal Contractors Already Know: � Incumbency vs. Availability Analysis

12

Source: OFCCP Sample AAP

https://www.dol.gov/sites/dolgov/files/ofccp/regs/compliance/AAPs/Sample_EO11246_AAP_final_01.03.18_Contr508.pdf

What Federal Contractors Already Know: � Incumbency vs. Availability Analysis

What an Incumbency vs. Availability Analysis Can Provide:

• Awareness of an oppositional viewpoint.
• Especially in high turnover positions.
• Lack of mutual understanding between HR and management on job functions (discovered by crosswalking your job IDs to ACS data).
• Trendlines.

Where Incumbency vs. Availability Analyses Fall Short:

• Very low-turnover positions.
• Small job groups.
• Very specific job duties or circumstances that don’t line up with ACS jobs.
• Without more, do not identify the stages of employment relationship in apparent need of improvement.

13

What Federal Contractors Already Know: � Impact Ratio Analyses

41 C.F.R. §60-3 (UGESP)

• “These guidelines apply to tests and other selection procedures which are used as a basis for any employment decision.”
• “If the information called for by …this section shows that the total selection process for a job has an adverse impact, the individual components of the selection process should be evaluated for adverse impact.”
• “The use of any selection procedure which has an adverse impact on the hiring, promotion, or other employment or membership opportunities of members of any race, sex, or ethnic group will be considered to be discriminatory and inconsistent with these guidelines, unless the procedure has been validated in accordance with these guidelines…”

41 C.F.R. §60-2.17

(b)Identification of problem areas. The contractor must perform in-depth analyses of its total employment process to determine whether and where impediments to equal employment opportunity exist. At a minimum the contractor must evaluate:

….

(2) Personnel activity (applicant flow, hires, terminations, promotions, and other personnel actions) to determine whether there are selection disparities;

….

14

What Federal Contractors Already Know: � Impact Ratio Analyses

15

Source: https://www.dol.gov/sites/dolgov/files/OASAM/legacy/files/StatisticalAnalysisComponentEEOToolkitModel.ppt

• Repeated by job group.
• If adverse impact on selection shown, OFCCP now takes position contractor should review each phase of the process (e.g., application review by HR, application review by managers, background check, interview 1, interview 2, detailed reference checks) under UGESP.
• Similar calculations for promotions, transfers, and terminations.

What Federal Contractors Already Know: � Impact Ratio Analyses

What Data Agile Federal Contractors Get Right

• Use disposition codes well and consistently.
• Exclude candidates– individuals who apply but lack objective minimum qualifications that are not in any way relative to other candidates.
• One job = one applicant pool.
• Applicants a manager wants to contact for another position/a future position cannot be carted over to that applicant pool (they can be invited to apply).
• No evergreen applications.
• Don’t use a “show me qualified candidates” button.
• Keep very clean phases of the application process.

16

Data-Driven Compensation Analyses

Approaches; EEOC and OFCCP Update

Internal Compensation Analyses:� Considerations

• Grouping comparators:
• The big EEO “Job Groups” are generally too broad to be meaningful; on the other hand, if two jobs have the same minimum qualifications and are set in the same salary range, they could be appropriate comparators.
• Locations: Can be handled by a multiplier, or, in a group large enough for multiple regression analysis, by using location as a factor. In other cases, it’s not so clear cut.
• Managers
• Other exceptions
• What types of pay to compare (ex: if OT is largely a matter of employee choice, it should possibly be excluded).
• Privilege (note, federal contractors need to take additional steps)

18

Internal Compensation Analysis: � Multiple Regression Analysis

The gold standard of compensation analysis.

Needs:

• Sufficiently large group;
• Data that can be reduced to a number.

​

Internal Compensation Analysis: � Multiple Regression Analysis

20

When Multiple Regression Analysis Isn’t Possible

• Calculate mean and median comparable compensation (could be rate, annualized; could include bonus(es), OT, or not, etc.) by race/ethnicity and sex.
• Where disparity exceeds a threshold, look for explanation.
• Threshold – we often use $5,000/5%, but $2,000/2% or other figures could also be valid.

21

When Multiple Regression Analysis Isn’t Possible

Common Explanations for Disparities

• Outliers (with an LNDR for outlier status).
• Seniority correlated positively with favored protected status.
• Performance.
• Book of business.
• Availability.
• Location.
• Special certifications.

​

​

​

22

EEOC Update - Compensation��

23

EEOC Update - Compensation

What Happened to the Component 2 Pay Data From the EEO-1?

• The data was (kinda) collected for years 2017 and 2018.
• It was studied by the National Academy of Sciences from 2020-2022.
• While voluminous and courteously phrased, the NAS review left few viable future case uses for the data, if it were ever collected again.
• The aggregated and anonymized data from 2017-2018 lives here: https://www.eeoc.gov/data/2017-and-2018-pay-data-collection
• It was released in 2023.

​

24

EEOC Update - Compensation

25

A comparison of Hispanic male Operatives and Craft Workers in Alabama vs. Hispanic female Operatives and Craft Workers n Alabama

The median pay band for the Hispanic males was $30.6K-$38.9K, while the median pay band for women was two pay bands lower at $19.2K-$24.4K.

EEOC Update - Compensation

26

A comparison of Hispanic male Operatives and Craft Workers in Alabama vs. Hispanic female Operatives and Craft Workers n Alabama

The median pay band for the Hispanic and Black males was the same: $30.6K-$38.9K.

OFCCP Updates��

OFCCP’s Dreaded Audit Item 19

28

OFCCP Discovers Annual Compensation Evaluation Obligation

• In early 2022, the OFCCP took a new look at obligation that contractors ought to “evaluate…compensation system(s) to determine whether there are gender-, race-, or ethnicity-based disparities,” which had historically been satisfied by a general non-discrimination clause, and decided this meant that contractors were actually required to conduct full, annual assessments of pay each year.
• Initially, OFCCP took the approach that the analysis couldn’t be privileged, even if external counsel conducted it.
• In August 2022, OFCCP withdrew that, but added that the contractor had to produce a non-privileged affidavit, summarizing how the analysis was conducted, and identifying any disparities, how large the disparities were and who was affected, what investigation into the disparity was conducted, and if any action items were implemented.

29

Questions?

Whitney Brown, wbrown@lehrmiddlebrooks.com

205-323-9274