Click for next page ( 8


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 7
1 Background The U.S. Equal Employment Opportunity Commission (EEOC) has a significant and active data collection program, which primarily collects information about employment status. While EEOC currently collects some pay data in its periodic reports from state and local government agencies for antidiscrimination enforcement, the agency has not collected pay data from private-sector employers, except on a case-by-case basis as necessary to support specific investigations. With that exception, the agency has no ex- perience in collecting pay information from the private sector.1 In its annual collection of data from private employers (EEO-1), the EEOC collects only employment classified by job category, gender, race, and national origin. In this chapter, we briefly summarize relevant employment discrimi- nation laws and describe the data that are currently collected in support of EEOC’s enforcement program. We also describe the current roles and responsibilities of the key federal agencies that enforce those laws and that now use the EEOC data. 1 The terms pay, wages, and earnings are used interchangeably in this report, depending on the context. They are taken to mean remuneration for labor or services to a worker on an hourly, daily, weekly, or annual basis or by the piece. The terms salary and compensation are also used in this report: salary is a fixed form of pay, wages, or earnings; compensation is the total amount of the monetary and nonmonetary pay provided to an employee by an employer in return for work performed, including money, benefits, services, and in-kind payments. 7

OCR for page 7
8 COLLECTING COMPENSATION DATA FROM EMPLOYERS LEGISLATION, AUTHORITIES, AND RESPONSIBILITIES Discrimination in pay on the basis of sex has been outlawed by the federal government for almost 50 years, since the Equal Pay Act of 1963. Enacted as an amendment to the Fair Labor Standards Act, the Equal Pay Act’s coverage is very broad. It applies to any employer “engaging in com- merce or in the production of goods for commerce” with an annual gross income of $500,000 or more (29 U.S.C. § 203(s)). Government entities and health and educational institutions are covered irrespective of size. There are narrow exceptions to coverage under the statute for certain kinds of employees (see 29 U.S.C. § 213(a)). The Equal Pay Act requires that men and women in the same work- place be given equal pay for jobs “the performance of which requires equal skill, effort, and responsibility, and which are performed under similar working conditions” (29 U.S.C. § 206(d)(1)). Unequal pay between men and women for jobs that are substantially equal violates the act unless the employer can show that the difference in pay is attributable to a bona fide seniority, merit, or incentive system or another factor other than sex. Al- though the U.S. Department of Labor (DOL) was initially given authority to enforce the act, that authority was transferred to the EEOC in 1978. Originally enacted one year after the Equal Pay Act in 1964, Title VII of the Civil Rights Act (hereafter, Title VII) prohibits a wide range of dis- criminatory employment practices, including discriminatory pay practices, and addresses discrimination based on sex, as well as race, color, religion, and national origin. Title VII covers private-sector employers with 15 or more employees and state and local government employers. Under Title VII, an employee challenging pay discrimination must show that he or she is paid less than another similarly situated employee because of race, color, religion, sex, or national origin. If he or she does so, then the employer must explain the reason for the disparity. The em- ployer may assert any of the defenses in the Equal Pay Act or a different, nondiscriminatory reason for the pay disparity. If the employer is unable to provide a satisfactory explanation for the disparity, the employer will be liable for penalties for pay discrimination. If the employer does provide a satisfactory reason for the disparity, the employee would have to show that the employer’s stated reason is a pretext in order to succeed in proving pay discrimination. Even where an employer does not intend to discriminate, a practice that is, on its face, neutral but that has the effect of disproportionately exclud- ing or adversely impacting members of a protected group can violate Title VII. In such “disparate impact” cases, the individual alleging discrimination must prove—usually through statistical evidence—that the challenged prac- tice has a substantial and significant adverse effect on a protected group.

OCR for page 7
BACKGROUND 9 If the individual proves this, the employer will be liable for discrimination unless it can show that the practice in question is job related and consistent with business necessity. If an employer can demonstrate that a practice is indeed justified, the individual will be given an opportunity to prove that there are other practices that would also serve the employer’s purposes, but with less impact on the protected group. Title VII’s prohibitions on compensation discrimination are broader than those contained in the Equal Pay Act. For example, under Title VII, an employee can challenge not only unequal pay between men and women performing substantially equal work, but also discriminatory practices that lead to unequal compensation, such as steering women to lower paid jobs than men or maintaining “glass ceilings,” artificial barriers to the advance- ment of women. Title VII empowers the EEOC to accept and investigate charges of discrimination from persons who believe they have been subjected to em- ployment discrimination and from those acting on their behalf. Title VII also allows for members of the commission itself to file charges of unlawful employment practices against employers. The EEOC is also empowered to open “directed investigations” under the Equal Pay Act, thereby allowing the EEOC to investigate the possibility of a violation of the act without having received a charge of discrimination from an aggrieved person. Individuals must exhaust their administrative remedies through the EEOC prior to filing a lawsuit under Title VII. But under the Equal Pay Act, while aggrieved persons may file charges of discrimination with the EEOC, they are not required to do so in order to file a lawsuit under the act. Moreover, filing a charge under the act with the EEOC does not suspend the statute of limitations under the Equal Pay Act, as it does under Title VII. For this reason, and in light of the significant time it can take to exhaust administrative remedies through the EEOC, some aggrieved individuals find it preferable to file a lawsuit under the Equal Pay Act without filing a charge with the EEOC. Under both Title VII and the Equal Pay Act, the EEOC investigates charges of discrimination and seeks to resolve them without litigation. However, the EEOC litigates a number of charges in which conciliation has failed each year. Under Title VII, the EEOC can litigate cases against private employers; charges against state and local governmental entities have to be referred to the U.S. Department of Justice (DOJ) for litigation. Under the Equal Pay Act, the EEOC may litigate against any covered employer, private or public. In fiscal 2010, a total of 99,922 charges were filed, many for multiple allegations of discrimination (U.S. Equal Opportunity Commission, 2010). Relatively few of those multiple allegation charges included wage discrimi- nation; far fewer were for wage discrimination only: see Table 1-1. The

OCR for page 7
10 COLLECTING COMPENSATION DATA FROM EMPLOYERS TABLE 1-1  Charges Filed with U.S. Equal Employment Opportunity Commission, by Issue: October 1, 2009, to September 30, 2010 Basis for Charge Issue Total Charges Race/National Origin Gender Total charges in which 4,478 2,314 2,164 wage discrimination was an issue Charges alleging only 638 282 356 wage discrimination Percent of wage 14.3% 12.2% 16.5% discrimination charges in which wage discrimination was the only allegation SOURCE: Data from U.S. Equal Opportunity Commission. majority of wage charges also involved other issues, most commonly terms and conditions of employment, termination, promotions, or discharges. The Employment Litigation Section of the DOJ’s Civil Rights Division is also charged with the enforcement of Title VII of the Civil Rights Act. Specifically, DOJ has jurisdiction to enforce Title VII against state and local government employers nationwide. DOJ can initiate litigation under Title VII in two ways: (1) DOJ has independent authority to bring suit against a state or local government employer when there is reason to believe that a “pattern or practice” of discrimination exists; or (2) DOJ may investigate and file suit against a state or local government employer based on an in- dividual charge of discrimination referred by the EEOC. DOJ can initiate such a suit if the EEOC has found reasonable cause to believe that discrimi- nation occurred, the EEOC’s efforts to obtain voluntary compliance have been unsuccessful, and EEOC has referred the charge to DOJ. The Office of Federal Contract Compliance Programs (OFCCP) in DOL is responsible for making certain federal contractors follow require- ments in the Executive Order 11246 (issued in 1965) to practice equal opportunity and take affirmative action on issues of race and gender.2 In addition, OFCCP is responsible for enforcing Section 503 of the Rehabili­ ta­ion Act of 1973, covering persons with disabilities, and the Vietnam t Era Veterans’ Readjustment Assistance Act of 1974 (VEVRAA), covering ­ veterans and disabled veterans. Under these laws, federal contractors must provide equal employment opportunities and take affirmative action to 2  n I addition to race and sex, Executive Order 11246 (originally implemented in 1965) ad- dresses equal opportunity on the basis of religion, color, and national origin.

OCR for page 7
BACKGROUND 11 employ and advance applicants and employees; provide reasonable accom- modations to disabled employees and applicants; prepare Affirmative Ac- tion Plans (AAPs); permit OFCCP access during compliance reviews; and file an annual report with the EEOC.3 OFCCP regulations require contractors to maintain records on em- ployee compensation and provide them on request (41 C.F.R. § 60-1.12(a), covering records on “rates of pay or other terms of compensation”). The regulations also require contractors to “regularly” monitor their compensa- tion systems for potential pay disparities based on race and gender, develop and implement appropriate corrections to any problem areas they identify, and report the results of their internal monitoring to management (41 C.F.R. § 60-2.17). This language apparently requires federal contractors to maintain data on earnings by demographic characteristics. EEOC DATA COLLECTION AND REPORTS The various laws and regulations to enforce antidiscrimination laws are accompanied by laws and regulations for the federal government to collect data that can be used in their enforcement. The EEOC uses its authority under Section 2000e-8(c) of Title VII to collect workforce data from em- ployers. The statute requires employers to preserve “records relevant to the determinations of whether unlawful employment practices have been or are being committed,” and to “make such reports therefrom as the Commission shall prescribe by regulation or order, after public hearing, as reasonable, necessary, or appropriate for the enforcement of [Title VII] or the regula- tions or orders thereunder.” The EEOC currently collects workforce data from private-sector em- ployers with more than 100 employees, from federal contractors with 50 or more employees, and from all state and local government employers. Employers that meet the reporting thresholds have a legal obligation to provide the data; it is not voluntary. The data are collected through several equal employment opportunity (EEO) reports. There are four versions of the required EEO reports, each addressed to different employer groups. Each of the versions collects employment data about gender and race/ethnicity by some type of job grouping; each provides, in essence, a snapshot of the demographics of the workplace by job category. Copies of these report forms are provided in Appendix A. 3  he T application of each of these requirements may vary on the basis of contract size and number of employees.

OCR for page 7
12 COLLECTING COMPENSATION DATA FROM EMPLOYERS TABLE 1-2  EEO-1 Reports by Number of Employees Covered and Percentage Female and Minority, 2010 Number of Number of Percentage Percentage Size of Firm Firms Employees Female Minority TOTAL 67,422 59,128,582 5th percentile: 3,443 191,965 38.6 26.8 1–67 employeesa 5th–25th percentile: 13,511 1,312,297 41.7 29.6 68–120 employees 25th–50th percentile: 16,875 2,587,008 45.6 31.1 121–194 employees 50th–75th percentile: 16,767 4,615,048 46.6 32.6 195–407 employees 75th–90th percentile: 10,090 6,541,695 47.4 33.6 408–1,118 employees 90th percentile and higher: 6,736 43,880,569 50.0 34.8 more than 1,118 employees aIncludes only firms with at least 50 employees. SOURCE: Data from U.S. Equal Employment Opportunity Commission (2010 EEO-1 Ag- gregate Report of U.S.). EEO-1 Report The EEO-1 report is required from private employers with: 100 or more employees or 50 or more employees and a federal contract. Firms must file a separate report for each facility with 50 or more employees. Approximately 67,000 firms filed more than 1.3 million EEO-1 reports (Standard Form 100) in 2009 and 2010. For 2010, the reports covered 59 million employees, which is almost one-half of the 108 million employees for all firms in the pri- vate sector. The largest 10 percent of covered firms represented about 75 per- cent of covered employment, and covered firms with 120 employees or less represented only about 2.5 percent of covered employment: see Table 1-2. Employers are required to file the EEO-1 report annually to the EEO-1 Joint Reporting Committee (joint between EEOC and OFCCP) (due on September 30). The data elements that are collected include 7 race/ethnicity categories and 10 job groups, by gender.4 Employers may use employment 4  he race/ethnicity categories are Hispanic or Latino, and—under not Hispanic or Latino— T white, black or African American, Native Hawaiian or Other Pacific Islander, Asian, American Indian or Alaska Native, and two or more races. The 10 job groups are executive/senior level officials and managers; first- or mid-level officials and managers; professionals; techni- cians; sales workers; administrative support workers (formerly, office and clerical workers);

OCR for page 7
BACKGROUND 13 figures from any pay period in July through September. Employers may sub- mit their EEO-1 reports on paper forms, as data files5 by electronic transfer, or by keying the data online through the EEO-1 online filing system. About 99 percent of the reports are received electronically. There are different types of reports for single establishment employers and multiple establishment employers. Multiple establishment reports must include a consolidated form that includes all employment for the company, one for headquarters locations, and one for each location with 50 or more employees. Locations with fewer than 50 employees are required to report only the address and total number of employees at that establishment, rather than a complete matrix. EEO-3 Report The EEO-3 report form is used for referral unions, which are generally unions with exclusive hiring arrangements with an employer. The report is required in even-numbered years with a due date of December 31. The EEO-3 form collects data on membership and referrals by race/ethnicity and gender. In 2010, there were about 1,200 reporting unions. The reports are used for enforcement and provide basic membership and referral data for investigators. They also allow statistical analyses to examine equity in membership and referrals. EEO-4 Report The EEO-4 report form is used for state and local governments. It is required in odd-numbered years and is due on September 30. Approxi- mately 6,000 jurisdictions filed EEO-4 reports in 2009. The reports that year covered 5,980,305 employees. This is the only EEO report that now collects employment data by job group and salary ranges for race/ethnicity and gender, with separate reports by function. Data are also collected separately for part-time employees and new hires. (See Chapter 3 for discussion of the wage data that are collected in this survey.) craft workers (formerly, craft workers, skilled); operatives (formerly, operatives, semiskilled); l ­aborers and helpers (formerly, laborers, unskilled); and service workers. 5  he files are sent as ASCII/text files, a simple data transfer that does not use developing T techniques such as XML.

OCR for page 7
14 COLLECTING COMPENSATION DATA FROM EMPLOYERS EEO-5 Report The EEO-5 report form is used for primary and secondary public school districts. It is required in even-numbered years with a due date of November 30. For 2010, the EEOC received more than 5,800 of these reports. The data are collected from each school district with 100 or more employees by race/ethnicity and gender for relatively detailed job groups.6 EEO-5 data are also collected for part-time employees and for new hires.7 WHITE HOUSE TASK FORCE REPORT AND PANEL CHARGE Following President Obama’s pledge in the 2010 State of the Union address to increase enforcement of equal pay laws, the administration es- tablished the National Equal Pay Enforcement Task Force in 2010, bring- ing together EEOC, DOJ, DOL, and the Office of Personnel Management (OPM). The task force identified several challenges to successful enforce- ment of compensation discrimination laws and made recommendations to address each challenge. Three of the five challenges identified by the task force have implications for this report: • Three different federal agencies have distinct responsibilities to enforce the laws prohibiting pay discrimination, and the agencies do not consistently coordinate these responsibilities. • The government’s ability to understand the full scope of the wage gap and to identify and combat wage discrimination is limited by the data that are currently available. As the task force report says, “this lack of data makes identifying wage discrimination dif- ficult and undercuts enforcement efforts. We must identify ways to collect wage data from employers that are useful to enforcement agencies but do not create unnecessary burdens on employers” (National Equal Pay Task Force, 2010, p. 5). • Existing laws do not always provide federal officials with ad- equate tools to fight wage discrimination. The task force report noted the administration’s strong support for the Paycheck Fairness Act, which would have required EEOC to use its data collection 6  The job groups are officials, administrators, and managers; principals; teaching assistant principals; nonteaching assistant principals; elementary classroom teachers; secondary class- room teachers; other classroom teachers; guidance staff; psychological staff; librarians and audiovisual staff; consultants and supervisors of instruction; other professional staff; teacher aides; technicians; clerical and secretarial staff; service workers; skilled crafts; and unskilled laborers. 7  art-time employees are grouped by professional, instructional, and all other; new hires are P grouped by officials, administrators, managers, principals and assistant principals, classroom teachers, other professional staff, and nonprofessional staff.

OCR for page 7
BACKGROUND 15 BOX 1-1 Statement of Task The National Research Council through its Committee on National Statistics (CNSTAT) will convene a panel of experts to review methods for measuring and collecting pay information from U.S. employers for the purpose of administering Section 709 of the Civil Rights Act of 1964, as amended. The panel will evaluate currently available and potential data sources, methodological requirements, and appropriate statistical techniques for the measurement and collection of employer pay data. The panel will consider suitable data collection instruments, procedures for reducing reporting burdens on employers, and confidentiality, disclosure, and data access issues. It will issue a report with findings and recommendations on what data the EEOC should collect to enhance wage discrimination law enforce- ment efforts, which will assist the Equal Employment Opportunity Commission (EEOC) in formulating regulations at the conclusion of an 18-month study. authority to implement a pay data collection program within 18 months of its enactment. Specifically, the bill text would require EEOC to “consider factors including the imposition of burdens on employers, the frequency of required data collection reports (including which employers should be required to prepare reports), appropriate protections for maintaining data confidentiality, and the most effective format for the data collection reports.”8 The Paycheck Fairness Act would also have amended the Equal Pay Act to prohibit employers from retaliating against employees for discussing their pay.9 The EEOC charge to the panel stressed that it is important for the panel to bear in mind the key considerations about the balance between enforcement utility and burdens on employers. Regardless of the fate of the Paycheck Fairness Act, the EEOC wants to ensure that any effort to col- lect wages takes into full account the considerations expressed in the Act regarding burden on employers, confidentiality, and appropriate format for collection. The complete statement of task is in Box 1-1. 8  his T text is from the 112th Congress version of the bill, S. 3220. 9  he T legislation passed the House of Representatives in 2009 but then failed in a cloture vote in the Senate in 2010. It has since been reintroduced in both chambers in the 112th Congress, with the Senate version (S. 3220) failing a cloture vote in June 2012.

OCR for page 7
16 COLLECTING COMPENSATION DATA FROM EMPLOYERS PAY RATE INFORMATION The employment data collected by EEOC are currently used for a vari- ety of purposes, including enforcement, self-assessment by employers, and research. The EEOC’s current collection of employment data contributes significantly to the efficiency of EEOC investigations and particularly to the development of systemic investigations. However, in a statement submitted to the panel, EEOC chair Jacqueline A. Berrien stated that the agency sees the absence of “employer-specific pay data broken down by demographic category” as a “significant barrier” to the agency’s work to eradicate pay discrimination. Berrien contrasted pay discrimination, a form of discrimina- tion she described as “largely invisible,” with other forms of discrimination that are easier to detect and that EEOC can more easily confirm or refute through the use of its current data collections. Many workplaces explicitly prohibit employees from discussing pay, and even in the absence of an explicit prohibition, employees in the United States rarely discuss their pay with one another. Because very few people know what their coworkers are paid, few people file complaints with the EEOC alleging that they are being paid in a discriminatory manner. In his testimony to the panel, EEOC commissioner Stuart Ishimaru pointed out that sex-based wage charges have made up a surprisingly small portion of the charges EEOC has received—about 2.5 percent. Berrien contended that, in addition to strengthening the EEOC en- forcement program under Title VII and the Equal Pay Act, better pay data collection would also assist employers in monitoring their compliance with federal, state, and local laws prohibiting wage discrimination. By maintain- ing accurate pay data, Berrien said, employers will be able “to compare and identify pay differentials that deserve closer scrutiny and to detect other patterns that may suggest departures from the standard of equal pay for equal work.” EARNINGS INFORMATION Use by OFCCP10 OFCCP officials similarly argued for the collection of earnings in- formation in a presentation to the panel. Under the authorities discussed above, federal contractors must provide equal employment opportunities, take affirmative action to employ and advance their employees, and make reasonable accommodations to employees and applicants. 10  This section summarizes a presentation to the panel by Pamela Coukos, senior program advisor, OFCCP.

OCR for page 7
BACKGROUND 17 A major requirement imposed on certain covered federal contractors is to develop an AAP. To meet this requirement, contractors must maintain ap- propriate records by establishment or function. The AAP data requirements cover the following topics: an organizational profile; a job group analysis; and information on placement of incumbents, determining availability, and comparing incumbency to availability. The AAP should spell out placement goals and designate an individual responsible for implementation. Problem areas need to be identified and action-oriented programs specified, and the plans need to be audited periodically. The AAP instructions call on employers to group jobs by similar pay and work content and to classify them into an appropriate EEO category based on similar duties and responsibilities, as well as similar opportuni- ties for training, transfer, pay, and promotion, and similar jobs in lines of progression. An example of an AAP workforce analysis is shown in Table 1-3. The OFCCP has minimum employee and contract size requirements for federal contractors11 and different rules for construction contractors. For example, construction contractors with federal contracts or subcontracts valued at more than $10,000 in any 12-month period are covered by Execu- tive Order 11246 at all construction worksites in the United States (Office of Federal Contract Compliance Programs, 2009). The enforcement activities of OFCCP primarily involve full compliance reviews. These reviews begin with desk audits of information submitted by a contractor in response to a scheduling letter, and they may also include an onsite review. Contractors are identified as being subject to enforcement activities based, in part, on a system called the Federal Contractor Selection System (FCSS). This system draws information from the universe of EEO-1 reports and federal contractor databases. Using these data sources, OFCCP selects contractors based on threshold requirements, sampling procedures, and mathematical modeling. An OFCCP compensation analysis consists of an initial review of aver- age pay differences for job categories. The agency then performs a statisti- cal or individual analysis as appropriate (depending on sample size and available data) and further review and analysis based on contractor pay practices and data. These data are used to assess the company’s practices. The investigation is designed to answer some basic questions: Are there pay differences between employees in a protected class and otherwise similar 11  asically, B all federal contracts and subcontracts are covered under Executive Order 11246 unless specifically exempted. Contracts and subcontracts of less than $10,000 generally are ex- empt, though some contracts under that amount are covered (e.g., bills of lading). Also exempt is work performed outside the United States; certain contracts with state or local governments; contracts with religious corporations, associations, and educational institutions; and contracts involving work on or near an Indian reservation. See 41 C.F.R. § 60-1.5.

OCR for page 7
TABLE 1-3  Example of an Employer’s Workforce Analysis for an Affirmative Action Plan 18 Workforce Analysis DEPARTMENT/WORK UNIT: Administration MALES FEMALES EEO-1 Category Wage (EEO-1 Job Total Job Title Rate Form or Group Employees OFCCP regulations) White Hispanic Total Black/African American Asian/Pacific Islander American Indian/Alaskan Native Total White Black/African American Asian/Pacific Islander American Indian/Alaskan Native Hispanic S-A 1 1 1 1 1 General Manager S-D 1 1 1 1 1 Personnel Manager S-J 5 5 1 1 1 Executive Assistant Administrative H-8 5 5 1 1 1 Assistant File Clerk H-11 5 5 2 1 1 1 1 DEPARTMENT TOTAL 6 3 2 1 3 1 1 1 SOURCE: U.S. Department of Labor, available: http://www.dol.gov/ofccp/regs/compliance/pdf/sampleaap.pdf [July 2012]. SAMPLE AAP Page - 4

OCR for page 7
BACKGROUND 19 employees? Are there differences in salary/hourly rate, promotions, job assignment, and access to earning opportunities? Are there legitimate ex- planations for any differences? At the time this report was being prepared, OFCCP was considering a new compensation reporting tool that would proactively allow the agency to more effectively identify potential violations of Executive Order 11246. The agency has requested public input on the kind of instrument that could be used for this purpose. This initiative is discussed in Chapter 2. Use by DOJ12 According to Jocelyn Samuels, senior counselor to the assistant at- torney general for civil rights of DOJ, the department uses data, including pay data, gleaned from the EEO-4 reports to fulfill its responsibilities under antidiscrimination statutes. The “pattern or practice” cases initiated based on the department’s independent authority under Title VII, Samuels told the panel, “are factually and legally complex cases that seek systemic injunc- tive relief to alter unlawful employment practices—such as discriminatory recruitment, hiring, assignment or promotion policies—which have the purpose or the unjustified effect of denying employment or promotional opportunities to a class of individuals.” DOJ may also investigate and file suit against a state or local government employer based on an individual charge of discrimination referred by the EEOC, as described above. The department routinely consults and relies on the information in- cluded in the EEO-4 reports regarding workforce composition and new hires, in combination with other information, to determine whether or not to use its enforcement jurisdiction to investigate a specific state or local government employer. Specifically, the department relies on EEO-4 reports for data on the demographics of different job categories in an employer’s workforce to assist in deciding whether to pursue investigations of allega- tions that may constitute a “pattern or practice” of discrimination. The EEO-4 information enables comparisons of an employer’s workforce in a particular job category to an applicable benchmark—such as civilian labor force data in the relevant geographic area taken from census and survey sources—to determine whether a particular group appears to be underrep- resented in that job category or in the employer’s workforce as a whole. The comparison provides a basis to estimate whether there is a disparity in representation in the workforce and to make an initial assessment of the sig- nificance of the disparity, which is one factor that informs the department’s 12  tatement S of Jocelyn Samuels to the Panel on Measuring and Collecting Pay Information from U.S. Employers by Gender, Race, and National Origin Workshop, May 24, 2011.

OCR for page 7
20 COLLECTING COMPENSATION DATA FROM EMPLOYERS evaluation of whether to open an investigation in order to gather more detailed information from an employer. In her presentation to the panel, Samuels stated that the demographic data collected on the EEO-4 reports are invaluable for enforcement pur- poses, but the wage data on the form are currently less useful. The job categories and the wage bands reported on the EEO-4 form are too broad, and the current EEO-4 form does not include any other information, such as longevity (years of service), which can be a key determinant of salary in the public sector. In order to allow meaningful analysis, the department needs salary in- formation in narrower job classes and information about years of service in the job class. In addition, according to Samuels, salary information should be collected in narrower bands, and should, to the extent possible, reflect the entire amount earned, not solely base pay. State reports suggest that these data are readily available in many states.13 In addition, DOJ has recently executed a memorandum of understand- ing with the EEOC in order to obtain access to EEO-1 data for private employers. DOJ anticipates that it will use these data in enforcement efforts for comparison purposes in job categories that exist in both the public and private workforce. Use for Analysis and Research In their presentations to the panel, the representatives of the EEOC, OFCCP, and DOJ emphasized the enforcement purposes behind the collec- tion of data from employers and unions. However, by virtue of their depth and coverage, these data also have statistical, analytical, and research uses. EEOC publishes annual statistical summaries of employment data from the EEO-1 and EEO-4 reports, as well as information received from federal government departments and agencies, on its website in three series: Job Patterns for Minorities and Women in Private Industry (EEO-1); Job Pat- terns for Minorities and Women in State and Local Government (EEO-4); and Federal Sector Reports. The employment data by race/ethnicity and sex are published by industry, geographic area (state and local areas), and job category. As part of an emphasis on proactive prevention, EEOC’s Office of Re- search, Information, and Planning has produced a series of reports based on EEO-1 data. The reports over the past decade have focused on industries and sectors (the finance industry, retail distribution centers, the media, high- end department stores, investment banking, broadcasting, and law firms) 13  or F example, see information from the Florida Bureau of State Payrolls, available at: http:// www.archive.org/details/StateOfFloridaPayrollDatabase2008 [July 2012].

OCR for page 7
BACKGROUND 21 as well as on particular labor market topics, including How New Business Processes Impact Minority Labor Markets; Women of Color: Their Em- ployment in the Private Sector; Glass Ceilings: The Status of Women as Officials and Managers in the Private Sector; and Characteristics of Private Sector Employment Report. A major use of the employment data is in the context of charge-based investigations, in which the data are used to assist EEOC in identifying em- ployers that warrant statistical comparisons, which could, in turn, trigger further investigation of their EEO practices. For example, using the EEO-1 reports of the numbers of employees in the establishment(s) in a certain job group and gender, race, and ethnic category, EEOC staff calculate a number of indicators that are designed to assess the EEO status of the firm. Those indicators include • Actual number: The reported number of employees in a particular job group and gender, race, and ethnic category. • Expected number: The number of employees that would be ex- pected to exist in that certain job group and gender, race, and ethnic category according to the percentage employed by com- parison establishments that have been selected based on specified geographic and industrial scope. • Difference: The difference between the actual number and expected number of employees in a certain job group and gender, race, and ethnic category. If the difference is positive, the establishment is over the expected number; if it is negative, the number of employ- ees in that category is below the expected number—a difference that is often referred to as a “shortfall.” • Actual percent: The percentage of employees in a certain job group and gender, race, and ethnic category. • Expected percent: The percentage of employees that would be expected in that certain occupational and gender, race, and ethnic category based on that percentage in comparison establishments. • Two-tail probability: A binomial statistical significance test, which is used to determine if the differences between the actual and ex- pected numbers are statistically significant. Administratively, EEOC primarily uses the EEO-1 data to identify potential discriminatory practices in the context of an investigation of a charge and to otherwise support investigations. The EEO-1 data are used in different ways at different stages of the investigation, and the analysis becomes more refined as the investigation progresses. In a presentation to the panel, Bliss Cartwright of the EEOC Office of Research, Information, and Planning discussed these uses, selecting as a

OCR for page 7
22 COLLECTING COMPENSATION DATA FROM EMPLOYERS hypothetical example a comparison of gender employment in one firm to employment in similar firms in the labor market. In his example, the firm had 180 female professionals of 624 total professionals, about 29 percent: in contrast, the proportion of female professionals in the labor market was 40 percent. He assumed that the labor market percentage was estimated by aggregate EEO-1 data on other firms in similar industries and locations, and he applied a one-sample binomial test of statistical significance. The main characteristics of this hypothetical example can be summarized as follows: • Total Professionals: 624 • Female Professionals: 180 • Observed Proportion: 0.2885 • Labor Market Proportion: 0.4067 • Null Hypothesis: No Difference • Two-Tailed Probability: <0.0000 (less than one chance out of 10,000) • Conclusion: Substantial Evidence Against Null Hypothesis of No Difference in Proportions Other situations may require more refined analyses. For example, sometimes a national firm has many facilities, hiring workers for the same job in different local labor markets. Alternatively, a single firm may recruit executives from a national market, midlevel managers from a regional mar- ket, and operatives from a local market. The issue is that there are multiple units of analysis, each with different employee counts and labor market estimates. In these situations, other statistical methods might be more ap- propriate. For example, Cartwright illustrated one approach commonly known as a pooled binomial (Gastwirth and Greenhouse, 1987), which provides an estimate of the overall shortfalls giving a single probability value. It also allows examination of homogeneity, the extent to which the units of analysis differ from each other. The next step in an analysis is to seek additional information from an employer through a request for information (RFI) that is tailored to the potential infraction alleged in the charge. For hiring issues, for example, EEOC typically requests files with demographic information, applicant flow data, and job history records. The requested data may be extensive. The job history information typically contains the effective date of the hire or the action that distinguishes initial hires from rehires or returns by use of employee identification numbers. The requested records also include specific job titles, divisions, and salary grades. At this stage, a wide variety of statistical methods would be considered—including linear regression,

OCR for page 7
BACKGROUND 23 survival analysis, and stratified contingency tables—depending on the facts and issues in a particular case. Understanding the Labor Market Since collection of information about employment by gender, race, national origin, and job category was initiated on a regular basis in the 1970s, there has been intense interest by the academic community in using the data to understand labor markets, especially the effect of governmental programs and corporate human resource practices on employment dis- crimination. EEO-1 reports and enforcement data from the OFCCP have been used to examine the effect of affirmative action and other factors on the employ­ ent of minorities and women across different sectors of the m economy. Selden (2006) assessed a variety of studies that transcended disciplines,14 pointing out that most use the EEO-1 survey data to examine the impact of affirmative action on minority and female employment shares among firms with or without federal contracts in the private sector. Selden summarized work by Leonard (1990) that concluded that affirmative action led to em- ployment gains among women and minorities for the period 1974–1980 and rose more significantly for federal contractors than for noncontractors. Selden (2006, p. 915) concluded that “overall, studies using EEO-1 data have shown that affirmative action has significantly and positively influ- enced the minority employment share in the private sector, particularly in unskilled positions.” Although there have been difficulties in obtaining access to EEOC’s sur- vey data, the agency has made significant efforts to increase the access that researchers have to these data. Since 1996 the EEOC has entered into agree- ments with more than 35 researchers to allow access to these confidential databases. Much of this work has been published in peer-reviewed articles and books, which in many cases have raised new questions and topics for academic research. In economics, for example, Donohoe and Levitt (2001), McCrary (2007), and Miller and Segal (2011) examined the relationship between diversity and crime rates using EEO-4 data. In sociology, Dobbin, Kalev, and Kelly (2006) examined how personnel practices impact a firm’s work force diversity, particularly in management. These researchers also ex- amined the impact of OFCCP compliance reviews and Title VII lawsuits on employment profiles (Dobbin, Kalev, and Kelly, 2007; Kalev and Dobbin, 14  elden’s S assessment covered Ashenfelter and Heckman (1976); Chay (1998); Goldstein and Smith (1976); Holzer and Neumark (2000a, 2000b); Kellough (1990a, 1990b); Leonard (1984a, 1984b, 1990); Naff (2001); Naylor and Rosenbloom (2004); Rodgers and Spriggs (1996); Stephanopoulos and Edley (1995); and U.S. General Accounting Office (1991).

OCR for page 7
24 COLLECTING COMPENSATION DATA FROM EMPLOYERS 2009), and Kalev (2009) examined how work restructuring impacts occu- pational segregation based on race and gender. A wide range of other work has also been done. Several research- ers compared firm-level and sector-level changes in segregation by race, ethnicity, and sex (Stainback and Tomaskovic-Devey, 2009; Stainback, Robinson, and Tomaskovic-Devey, 2005). Huffman, Cohen, and Pearlman (2010) studied the impact of women managers on firm gender integration for the period 1975–1990. Skaggs (2008) studied how government ac- tion, including court decisions, affected female employment in food stores. Several other researchers explored the impact of various factors, including EEOC charge processing, on the employment of women and nonwhites (Hirsh, 2008, 2009; Hirsh and Kmec, 2009; Hirsh and Kornrich, 2008). Yet another group of researchers used EEO-4 data for a series of articles examining diversity in state and local governments including an examina- tion of glass ceilings among those employers (Kerr, Reid, and Miller, 1999, 2000a, 2000b, 2002, 2003, 2004). All of this research has been done even with the difficulties of obtaining access to the data, which are discussed in Chapter 5, and in the absence of compensation data. In the absence of employer-based earnings data by job category and demographics, however, the research community largely turned to house- hold data to support analysis of the extent and effect of compensation discrimination in the labor market. The Current Population Survey and, more recently, the American Community Survey have emerged as powerful sources of data on earnings, industry groups, occupations, and demograph- ics. However, these sources are limited because they do not associate the indicators of discrimination with actual employer situations and practices, nor can they be directly linked to measures of enforcement. There is clearly a strong research and analytical interest in having an earnings dimension to establishment, occupation, and demographic data (see, e.g., Consad Research Corporation, 2009, p. 2). It is expected that there would be significant pressure on agencies that held data enriched with earnings information to make them available for analytical uses by private sector researchers. Such data could quickly become a primary source for new analytic work on equal employment and compensation issues. Auditing the Effectiveness and Efficiency of Antidiscrimination Programs Over the years, Congress and a number of government agencies have used data collected on EEO-1 forms to assess the effectiveness of govern- ment antidiscrimination programs. Just as the research community would benefit from the availability of earnings data, these agencies would be ex- pected to take advantage of earnings information to sharpen their auditing reports.

OCR for page 7
BACKGROUND 25 The U.S. Government Accountability Office (GAO), in particular, has been at the forefront in terms of using employment data by job category and demographics. In the past two decades, GAO has published seven major studies that have been based in part on the EEOC employment data: • Sharing Promising Practices and Fully Implementing Strategic Human Capital Planning Can Improve Management of Growing Workload (2008); • Financial Services Industry: Overall Trends in Management-Level Diversity and Diversity Initiatives (2006); • Equal Employment Opportunity: The Policy Framework in the Federal Workplace and the Roles of EEOC and OPM (2005); • Women’s Earnings: Work Patterns Partially Explain Difference between Men’s and Women’s Earnings (2003); • Equal Employment Opportunity: Discrimination Complaint Case- loads and Underlying Causes Require EEOC’s Sustained Attention (2000); • Equal Employment Opportunity: DOL Contract Compliance Re- views Could Better Target Federal Contractors (1995); and • EEOC: An Overview (1993). Cross-Checking the Integrity of EEO Data An additional justification for the collection of pay data is that they may help to improve the integrity of EEO employment data. Smith and Welch (1984) found some evidence that the number of minorities and women reported to be in high-level occupations by their employers on EEO-1 forms exceeded the number who reported themselves to be in those occupations in the Current Population Survey. To the extent that some employers of minority- or female-intensive occupations systematically up- grade (or misclassify) them, it would cause unusual pay compression across EEO-1 job categories and unusual pay dispersion within the higher level occupations. Being able to make such assessments by using pay data would be valuable for evaluation purposes.