National Academies Press: OpenBook

Evaluation of Compensation Data Collected Through the EEO-1 Form (2023)

Chapter: 4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?

« Previous: 3 Utility of Current Concepts and Alternatives
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

4

Do All Eligible Employers Receive and Respond to the Component 2 Instrument?

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

INTRODUCTION

The EEO-1 Component 21 data collection intends to collect data from a census of employers that meet the eligibility criteria for filing. With a census of all required employers, EEOC would have initial information for its regulatory functions, would be able to provide external data to firms for self-assessment of pay discrepancies, and would facilitate use of full-population data for researchers studying contributors to pay gaps. The Component 2 data are the only tool that EEOC has for investigating potential pay disparities in the absence of in-depth investigations subsequent to employee complaints. Thus, missing firms and establishments mean that EEOC lacks information necessary for fulfilling its regulatory duty with respect to those missing firms and establishments. In addition, firms, researchers, and advocacy groups using data missing firms and establishments may draw incorrect conclusions about pay gaps within and across organizations or industries.

The target population for the Component 2 data collection constitutes the group of employers (firms) with 100 or more employees across all locations during a pay period between October 1 and December 31 of the selected reporting year. For consistent language, we will refer to employers as firms in this report. A firm may consist of a single establishment (i.e., have one location) or have multiple establishments. Establishments with fewer than 100 employees but that are affiliated with multi-establishment firms

___________________

1 In 2018 and 2019, EEO-1 data collections occurred in two components. The customary EEO-1 instrument containing composition data (Component 1) was collected in 2018 for reporting year 2017, and in 2019 for reporting year 2018. In 2019 and 2020, pay data (Component 2) were collected for reporting years 2017 and 2018. In this report, we refer to historical EEO-1 data collections as EEO-1, and the 2017 and 2018 EEO-1 Component 1 and Component 2 instruments as “Component 1” and “Component 2,” respectively. All information collected by the Component 1 instrument is also collected by the Component 2 instrument.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

with at least 100 employees are eligible to file the Component 2 instrument. Federal contractors with at least 100 employees are included in the target population, which is a larger size cutoff than for Component 1. State and local governments, public schools, institutions of higher learning, American Indian or Alaska Native tribes, and private-membership clubs are excluded, as are firms in Puerto Rico, the Virgin Islands, and other American Protectorates. Information must be reported by (1) all firms, (2) all establishments within the firm, and (3) all full-time and part-time employees on the payroll during the employer-selected reference period.

The target population includes both single-establishment firms and multi-establishment firms. Single-establishment firms are those firms that conduct business at only one location; multi-establishment firms are those that have more than one business location. For EEOC’s enforcement purposes, the unit of analysis is at the establishment level (EEOC, 2021i). In multi-establishment firms, however, the Component 2 data-collection respondent is the firm or its servicer (EEOC, 2020h). Thus, the panel examined representation of both firms and establishments.

Errors in representation may occur because of errors in the frame and failure to respond to the collection request. At the firm level, such errors may be further exacerbated by errors at the item level if information requested by the instrument is missing or in error (see Chapter 5).

This chapter begins with a brief discussion of the Total Survey Error and Total Quality Frameworks as the organizing frameworks for the analyses presented here and in Chapter 5. The focus then turns to the representation of establishments and firms, starting with internal comparisons (Groves, 2006) of the Component 2 data collection—including creation of the frame, response rates, and evaluation of the respondent pool over the data-collection period. External comparisons of Component 2 data to information about firms and establishments present in business registers maintained by federal statistical agencies are then discussed.

The completeness of the data depend on both the completeness of the frame used to contact eligible firms and the extent of nonresponse. The panel found that, among firms identified for the frame (prior Component 1 reporters and self-identified firms), firm-level response rates were around 90 percent, exceeding the court-mandated response rate of 82 percent for firms of all sizes, states, and industries. Data collection hit the target response rate by the first extended deadline. However, even with high response rates, there are substantial issues related to births, deaths, eligibility, and coverage across the two reporting years. Many firms not on the initial frame were brought into the filing universe during the Component 2 data collection; other firms previously identified as eligible were found to no longer be eligible. Due to the high response rate and additional firms that filed the Component 1 instrument, the degree to which Component 2 respondents

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

reflect the coverage of Component 1 respondents is quite high at the firm and establishment level. Component 2 cases often exceed the number of firms and establishments in Component 1 data across size categories and industry in both years.

Even with this churn, it appears that the number of new establishments reported in 2018 or the number of exiting establishments reported only in 2017 exceeds numbers expected from national establishment birth and death estimates. High-quality national benchmark estimates of the number of firms and establishments that match EEOC’s eligibility criteria are somewhat difficult to obtain, so the panel examined multiple benchmarks. Measurement differences in reporting size and industry further complicate comparisons between Component 2 data and national benchmarks.

Nevertheless, the panel estimates that about two in five firms with 100 or more employees did not file Component 2 instruments, and about the same proportion of the establishments in these firms were not included in the Component 2 data collection—a sizeable undercoverage rate. This undercoverage appears to be more severe for smaller firms, although precise estimates by firm and establishment size are difficult due to measurement errors in the reported numbers of employees. Undercoverage varies across industries.

This chapter examines coverage of firms and establishments but does not address coverage of employees. Due to large reporting errors in numbers of employees, the total estimated number of employees in firms and establishments is highly implausible, making comparisons to Component 1 data or to external benchmarks nonsensical without further restrictions on outlying reports (see Chapter 5). This misreporting introduces measurement error into the estimated size bins in the analyses reported below. In particular, the panel believes that the number of large firms and large establishments (1,000+ employees) is too high when using only Component 2 data and thus firms and establishments belonging in smaller size categories were miscategorized as larger firms. To address this, firm and establishment size from Component 1 data were used (if available) in the analyses below. A detailed examination of the quality of the reported information on employees, pay bands, and total hours worked is addressed in Chapter 5.

METHOD FOR EVALUATING DATA QUALITY

The Total Survey Error framework (Groves, 1989) and the more general total quality framework (Biemer and Lyberg, 2003) are useful organizing structures for understanding breakdowns in data quality. The total survey error framework divides errors into those generated by representation (i.e.,

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

the process of obtaining observations or rows in an analytic dataset; here, firms and/or establishments) and those by measurement (i.e., the process of obtaining variables or columns in an analytic dataset; here, reports of employee pay and total hours worked) (Groves et al., 2009). The total quality framework expands on the total survey error framework, identifying accuracy (or biasing errors) and precision (or variance) as one component of quality, but quality is also indicated by timeliness and accessibility.

Errors of representation occur due to coverage, sampling, nonresponse, and post-collection processing errors. Given the fluid nature of the Component 2 data collection, the sources of the errors of representation in Component 2 data are difficult to disentangle.

First, coverage errors occur when: (1) eligible units are not part of the frame for the study; and (2) covered units differ on characteristics of interest (e.g., pay discrepancies across groups of interest) from those who were not covered. In the Component 2 data collection, coverage errors may occur because an employer may not have appeared in EEOC’s Component 1 data collection in the years used to construct the Component 2 frame and thus was not on the contact list. Firms could remedy this coverage error themselves: firms that had not previously submitted Component 1 data could opt into the Component 2 data collection by contacting the data collector and requesting access. To facilitate reporting for both existing and new filers, a help desk and public website were established in June 2019, prior to the start of the July 2019 data collection. The website featured answers to frequently asked questions and reference materials to assist filers, and reminders were sent throughout the collection period to encourage prompt and complete responses (EEOC, 2020f).

Additionally, the population of eligible employers was not static over the two years of Component 2 data collection. For example, an employer may have had 100 or more employees in 2017 but not in 2018 (or vice versa); establishments may have gone out of business; multi-establishment firms may have closed some of their establishments; or new establishments may have been created.

Second, sampling errors are not relevant for the Component 2 data collection, as the intention was to conduct a census.

Third, nonresponse errors occur when: (1) the response rate is not 100 percent; and (2) those who responded differ from those who did not respond on the characteristics of interest. In the Component 2 data collection, nonresponse errors occur when surveyed employers do not participate at all or participated in one year but not both years, even when eligibility did not change. Responding multi-establishment firms may have failed to provide information on all their establishments. Responding firms may have failed

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

to include information for all eligible employees. Additionally, firms may have reported twice—once through self-report and again through a PEO.2

Finally, post-survey processing errors may occur at the representation stage when data are edited (possibly affecting eligibility of sample units) or when incorrect adjustments are created through weighting or imputation. Weighting and imputation do not apply to the Component 2 data collection, but data editing is needed to address inconsistent reports and/or outliers. Because Component 2 data were collected under unusual circumstances, less editing was performed during and after submission than would be performed in a more routine data-collection effort.

Errors due to measurement occur because of problems with construct validity, issues relating to instrument design or modes of data collection, or mistakes in editing responses during post-survey processing. Chapter 3 discusses the construct validity of the Component 2 data-collection approach. Chapter 5 discusses errors due to questions and modes, as well as potential editing errors.

The total quality framework emphasizes that data need to be “fit for use.” In other words, various uses and users of data may value unique aspects of quality. Total survey error, comprising both accuracy (bias) and precision (variance) of estimates, is one component of quality. Timeliness (i.e., minimal lag between data collection and availability of data for analysis) and accessibility (i.e., access to data by intended users) are two additional parts of the total quality framework (Biemer and Lyberg, 2003). Component 2 data for 2017 and 2018 were collected in 2019 to early 2020; at the end of 2021, more than three years had elapsed since the last year of data eligibility. The data are not currently accessible to EEOC investigators, complaint filers, scholars, or other groups.

COMPLETENESS OF EEOC’S BUSINESS LIST

EEOC’s frame, or business list, is updated in three ways: when firms self-identify as belonging in the list (all employers with at least 100 employees and federal contractors with at least 50 employees are required to submit an EEO-1 report),3 when a firm is identified through a field investigation, and (potentially) through comparisons with commercial data.

___________________

2 A PEO is an outsourcing firm that provides services to small and medium-sized businesses. Typically, PEOs offer human resource consulting, payroll processing, employer payroll tax filing, workers’ compensation insurance, health benefits, and other services. PEOs enter into contractual co-employment agreements with their clients. Through co-employment, the PEO becomes the employer of record for tax purposes through filing payroll taxes under its own tax identification numbers (Wikipedia, 2021).

3https://www.eeoc.gov/employers/small-business/legal-requirements#:~:text=Employers%20who%20have%20at%20least,Department%20of%20Labor%20every%20year

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

EEOC recognizes that its current list may be incomplete, and a September 2020 contract with Westat included an optional task to improve the existing list. All firms in the list are required to participate in the Component 1 data collection.

The Component 1 business list became the basis for the Component 2 data collection, except that federal contractors with 50–99 employees were not required to participate in Component 2. Also, the list is subject to change even during the process of data collection, such as through mergers, acquisitions, closures, and newly self-identified firms. The National Opinion Research Center (NORC) at the University of Chicago, the organization responsible for collecting Component 2 data, started with a list of eligible 2017 and 2018 filers of the Component 1 instrument. Initial delivery of this list was on June 17, 2019; data collection started on July 15, 2019; and a “final” list was provided to NORC on August 1, 2019. This August 1 delivery resulted in an initial sample frame of n = 71,798 employers across both years (EEOC, 2020e). Details of the development of the Component 2 frame are provided in Table 4-1.

Eligible firms that did not complete the Component 1 instrument were missing from the frame. These firms were eligible for inclusion if they met eligibility criteria for the Component 2 data collection. There were three processes by which the Component 2 frame could be updated.

First, EEOC provided historic (2015 and 2016) Component 1 data, allowing those firms to be included if they requested access to complete the Component 2 instrument. This process made 34,540 employers eligible to complete the Component 2 instrument (EEOC, 2020e), and more than 1,100 participated for both 2017 and 2018 as past EEO-1 reporters.

Second, some firms used PEOs to file their Component 2 data. PEOs may have reported for one or both years for any firm; additionally, PEOs may have uploaded joint reports for multiple firms, even if those firms had not completed the Component 1 instrument. Almost 6,600 firms were added to the sampling frame through PEO reports in 2017, and more than 7,400 firms were added this way in 2018.

Finally, employers may have voluntarily asked to complete the Component 2 instrument although they did not file Component 1 in either 2015 or 2016; EEOC reviewed and approved these requests, adding about 1,730 employers in each year. Across all three of these processes to augment the frame, a total 9,464 (2017) and 10,298 (2018) firms were added to the filing universe through these additions.

During data collection, some firms in the filing universe were deemed ineligible and thus deleted. Ineligibility was mostly due to having fewer than 100 employees (8,924 in 2017 and 7,431 in 2018). Other deletions (a total of 604 for 2017 and 548 for 2018) arose due to mergers or closures, status as an educational institution or government entity that was not required to

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-1 EEO-1 Component 2 Filing Universe Frame Development

Filing Universe Status 2017 2018
1 2017 or 2018 Component 1 EEO-1 certified filers 71,798 71,798
2 Additions to the filing universe: 9,464 10,298
3 Employers reported within PEO files 6,598 7,422
4 Employers self-identified as eligible in Component 1 2,866 2,876
5 Past EEO-1 reporter 1,132 1,143
6 New to EEO-1 1,734 1,733
7 Deletions to the filing universe: 9,916 8,377
8 <100 employees, self-reported 5,760 3,760
9 Mergers / acquisitions / closures 520 510
10 Ineligible business type 36 31
11 Hardship exemption request approved 48 7
12 Not required to file 2017 Component 1, reported to EEOC 274 274
13 <100 employees, PEO or parent reported 3,164 3,671
14 From initial sample 22 40
15 From PEO files 2,984 3,457
16 From self-identified employers 158 174
17 Duplicate employer 114 114
18 Current filing universe (line 1+ line 2 – line 7) 71,346 73,719

SOURCE: Component 2 EEO-1 Pay Data Key Indicators Report Universe Calculation, Data as of August 14, 2020.

file, or receipt of a hardship exemption due the burden of reporting. Across both years, 114 firms were identified as duplicates, so one observation was dropped. In total, 9,916 employers were deleted in 2017 and 8,377 in 2018. The final filing universe consisted of 71,346 employers in 2017, a decrease in 452 firms (0.6%) from the initial filing universe. The final filing universe in 2018 was 73,719, an increase of 1,921 firms (2.7%) from the initial filing universe.

These shifts in the eligible filing universe occur because EEOC updates the “master list” of eligible filers of Component 1 data only when a filer self-identifies as needing to submit a Component 1 instrument, or if a discrimination charge is filed but no Component 1 data are available for that employer (EEOC, 2021g). Historically, EEOC does not update the list of possible filers with any commercial list or other government agency list.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

Filers can self-report changes to their companies (e.g., mergers, acquisitions, spinoffs, closures, or changes in size), but EEOC has not traditionally solicited this information from filers.

The panel notes that EEOC’s ongoing data-modernization efforts involve contracts for internal evaluations of master list quality (EEOC, 2021g). The need for this effort is further explained below.

RESPONSE RATES

Because EEOC defines eligibility based on employer (firm-level) size, the eligible respondents are firms. As such, response rates for firms, not establishments, can be evaluated (Liao, 2021a). If the frame adequately covers the universe of eligible firms, the response rate and variation in response rate across subgroups indicate the quality of the data-collection effort and the resultant respondent pool (Groves et al., 2008). Differential failure of eligible employers to respond means that EEOC’s data may be incomplete in comparisons of similar employers for enforcement purposes.

The Component 2 data collection had a court-mandated minimum response rate:

It is FURTHER ORDERED that the EEO-1 Component 2 data collection(s) will not be deemed complete, for the purpose of this Order, until the percentage of EEO-1 reporters that have submitted their required EEO-1 Component 2 reports equals or exceeds the mean percentage of EEO-1 reporters that actually submitted EEO-1 reports in each of the past four reporting years. (NWLC et al. v. OMB et al., 2020, p. 2)

Following this instruction, the average expected response rate was targeted at 82 percent or greater (Grigorian, 2021). Thus, the panel evaluated whether this target response rate was met across subgroups.

Unfortunately, because firms could self-select into the Component 2 data collection, evaluating response rates and variation in response rates is complicated. As noted above, almost 10,000 firms were added to the frame during data collection. By definition, those added firms were respondents—they were identified as part of the target population through completing Component 2 instruments. Since these added firms were not part of the 2017 or 2018 Component 1 data collection, information on employer size, federal contractor status, or sex/race/ethnicity composition separate from Component 2 data was not available. Additionally, the panel was not provided with the original sampling frame, and IDs were not designed to facilitate merges of Component 1 and 2 data, further lessening the information available for calculating response rates. Thus, the panel did not have adequate information to evaluate variation in response rates across

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

employers of various sizes, for federal contractors versus non-federal contractors, or across firms with distinct compositions of their workforce. The panel did have information on the source of the firm’s submissions, North American Industry Classification System (NAICS) code, and state; NAICS industry codes and states are identified by firms at the time of filing.

Nonresponse bias is a concern in survey estimates. Nonresponse bias occurs when those who respond to the survey are systematically different in important ways from those who do not respond. Nonresponse bias in totals occurs whenever sample members fail to respond and post-survey adjustments are not made. This is the case for Component 2 data—as no weights or imputation methods were calculated, totals (e.g., the total number of establishments in a particular industry) will be systematically underestimated due to nonparticipation.

Overall, as shown below, most nonparticipation for the Component 2 data collection was produced by noncoverage of the initial master list coupled with failure to self-nominate; less nonresponse is of the traditional type, arising due to noncompletion of the Component 2 instrument after (presumed) contact by the data-collection contractor.

By Firm Reporting Method

In total, filings were collected from 64,100 firms in 2017 and 66,860 firms in 2018, yielding an overall response rate of 89.8 percent for 2017 and 90.7 percent for 2018. These high response rates are notable for a new data collection with a difficult response task. In this section, the full set of firms (n = 64,100 for 2017 and n = 66,860 for 2018) is used to evaluate variation in response rates.

Table 4-2 shows response rates by source of firm for the Component 2 filing frame. The initial list of 2017 and 2018 Component 1 filers constitutes about 92 percent of the full set of firms in each year; about 8 percent of eligible and responding firms in each year were identified through other means. The response rate among this initial list of firms eligible for filing the 2017 and/or 2018 Component 1 instrument was about 90 percent in each year. In contrast, self-identified volunteers were less likely to participate in the Component 2 data collection, with a response rate of 67 percent in 2017 and 73 percent in 2018.

By definition, all firms reported by PEOs are respondents and thus have a 100 percent response rate. Similarly, firms identified as part of the 2015 or 2016 Component 1 data collection but absent from the original frame for 2017 and 2018 were identified by their request for access to complete the Component 2 instrument; unsurprisingly, the response rate among these self-identified respondents meets or exceeds 99 percent.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-2 Component 2 Firm Response Rates, by Source of Firm

2017 2018
# Eligible Firms # Responding Firms Response Rate % # Eligible Firms # Responding Firms Response Rate %
Overall Response Rate (AAPOR RR2)a 71,346 64,100 89.8 73,719 66,860 90.7
Source of Firm for Component 2 Filing Universe
Firm eligible for 2017 and/or 2018 Component 1 65,715 58,878 89.6 67,614 61,131 90.4
Firm eligible for 2015 or 2016 Component 1 but absent from Component 1 790 787 99.6 776 768 99.0
Firm from Component 1 PEO filing 1,583 1,583 100.0 1,866 1,866 100.0
Self-identified volunteer firm new to EEO-1 filing 1,241 835 67.3 1,373 1,005 73.2
Firm from PEO new to EEO-1 filing 2,017 2,017 100.0 2,090 2,090 100.0

SOURCE: Panel generated from Component 2 employer files for 2017 and 2018.

NOTE: No exclusion rules were applied.

a American Association for Public Opinion Research response rate calculation formula 2.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

This variation in response rates by sample source provides two pieces of information. First, a firm that had previous contact with EEOC was highly likely to file—about 90 percent of firms that were on the initial list and were contacted by EEOC did so; firms without previous EEOC contact but who instigated the instrument-completion process were much less likely to file Component 2 instruments. To the extent that the data-modernization efforts (EEOC, 2021g) expand the filing universe list, EEOC may see an initial decline in filing response rates. Second, firms that use PEOs are added even when the firms themselves are not directly contacted by EEOC. In fact, about 2,000 firms were added to the filing universe in each year via PEO reports. Future Component 2 data collections would do well to consider whether and how PEO reports should be obtained.

By Industry

Response rates across all NAICS sectors exceeded 80 percent and thus exceeded the court-mandated response rates.

Figure 4-1 illustrates firm-level response rates across NAICS sectors. The highest response rates are among utilities (NAICS 22) and finance and insurance (NAICS 52) firms, with a response rate of approximately 96 percent for each year. That is, only four percent of the utilities and finance and insurance firms identified through Component 1 data or through other sources were missing from the Component 2 filings. Other than missing NAICS codes, accommodation and food services (NAICS 72), waste management and remediation services (NAICS 56), and retail trade (NAICS 44–45) had the lowest response rates; for these sectors, the response rate was about 86 percent for each year. Thus, about 14 percent of the firms in these industries identified in Component 1 data or through other sources were missing from Component 2 data. The two industries with the largest number of eligible firms had response rates that met or exceeded 90 percent: 91.2 and 92.0 percent of manufacturing firms (NAICS 31–33) and 90.2 and 90.9 percent of health care and social assistance firms (NAICS 62) responded in 2017 and 2018, respectively.

By State

Figure 4-2 shows firm-level response rates across states. Response rates by state varied little, all exceeding the court-mandated response rate (other than firms for which state data were missing). Firms in Nebraska had the highest response rates, exceeding 98 percent for each year. Five states (Arkansas, Wisconsin, South Dakota, Kansas, and North Dakota) had firm response rates meeting or exceeding 93 percent for each year. Only three states (New Jersey, Arizona, and Nevada) had firm response rates lower

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Image
FIGURE 4-1 Component 2 firm response rates by NAICS code, 2017 and 2018.
SOURCE: Panel generated from Component 2 employer files for 2017 and 2018.
NOTE: No exclusion rules were applied.
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Image
FIGURE 4-2 Component 2 response rates by state, 2017 and 2018.
SOURCE: Panel generated from Component 2 employer files for 2017 and 2018.
NOTE: No exclusion rules were applied.
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

than 85 percent for each year (the lowest for Nevada at 82.3% for 2017). The state with the largest number of firms, California, had 87.9 and 89.0 percent of firms respond in 2017 and 2018, respectively. Texas, the state with the second largest number of firms, had response rates around 88 percent for each year. Thus, across all states, no more than 15 percent of firms identified through Component 1 data or through other sources are missing from the set of Component 2 filers.

NONRESPONSE AND EFFECT OF ALTERING THE COLLECTION PERIOD

As discussed in Chapter 2, because of court orders, the end date for Component 2 data collection changed multiple times over the fielding period. EEOC asked the panel to evaluate the effects of extending the data-collection period.

The panel first evaluated how the extension of the data-collection period affected the overall response rate and number of responding firms. Participation did not occur linearly over the data-collection period. Most firms participated before the first data-collection stop date of September 30, 2019 (Figure 4-3), between weeks 11–12 of the data-collection period. As of the 12th week of data collection (ending October 2, 2019), responses were obtained from 78.4 percent of eligible 2017 firms (n = 55,944; 87% of the final number of responding firms) and 79.9 percent of the eligible 2018 firms (n = 58,886; 88% of the final number of responding firms). These response rates were just short of the 82 percent response rate target.

On September 27, 2019, the deadline for participation was extended by six weeks to November 11, 2019 (Figure 4-3). As of the week ending November 13, 2019, 83.6 percent (n = 59,616; 93% of the final number of responding firms) of the eligible 2017 firms and 85.3 percent (n = 62,873; 94% of the final number of responding firms) of the eligible 2018 firms had participated. Had data collection closed at this date, the target response rate would have been met for both years, but the number of filers would have been reduced by 4,000 firms in each year.

The deadline was extended one final time, with data collection ending on February 19, 2020, yielding the final response rates of 89.8 percent for 2017 reports and 90.7 percent for 2018 reports. These final response rates exceeded the court-mandated response rate. In sum, with 19 additional weeks of data collection, an additional 8,156 firms participated in the 2017 Component 2 data collection, increasing the response rate by 11.4 percentage points; an additional 7,974 firms participated in the 2018 data collection, increasing the response rate by 10.8 percentage points.

The panel then examined whether stopping data collection early would have disproportionately impacted specific types of establishments.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Image
FIGURE 4-3 Cumulative response rate by date of collection, Component 2, 2017 and 2018.
SOURCE: Panel generated from Component 2 employer files for 2017 and 2018.
NOTE: No exclusion rules were applied.

Establishments were the focus, because establishments are the unit used most often for EEOC’s enforcement purposes. To facilitate comparisons with later chapters, examined establishments have at least one employee (Component 2 reports for establishments in 2017 and 82 establishments in 2018 were excluded due to reporting no employees). The total number of employees reported in Component 2 sometimes appeared unreliable (see Chapter 5), in which case the employee totals from their Component 1 filings were used instead (if available) for classifying establishments by size. Establishments that reported only via Type 6 reports (less than 9,500 establishments in each year) were also excluded. Type 6 reports can be filed by establishments with fewer than 50 employees, and these reports do not contain the necessary sex, race/ethnicity, occupation, pay-band, and hours-worked information collected by the Component 2 data collection (see Chapters 1 and 2). Additionally, establishments filing Type 6 reports are instructed to file using NAICS code 999999 (Type 6) for industry classification further limiting the utility of these data for analysis. (EEOC, 2020e).

Overall, by the end of the original collection date (September 30, 2019), information had been reported for 727,474 establishments with at least

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

one employee for 2017 (81.0% of the establishments included in the final set of reports; Table 4-3) and for 764,367 establishments with at least one employee for 2018 (81.7% of the establishments included in the final set of reports; Table 4-5). The initial six-week deadline extension to November 11, 2019 substantially increased the number of establishments reported. Six weeks after the initial deadline, information had been reported on an additional 111,760 establishments with at least one employee for 2017 and an additional 118,676 for 2018, yielding 93.5 and 94.4 percent of the final set of establishments for 2017 and 2018, respectively. At the end of data collection (February 2020), information had been reported on 897,683 establishments with at least one employee for 2017 and on 935,528 establishments for 2018, increasing the total number of establishments by 170,209 (19.0% of the final total) for 2017 and by 171,161 (18.3% of the final total) for 2018 over the original data-collection stopping date.

In general, there was little variation in the characteristics of establishments reported over the data-collection period. That is, the presence of establishments of various sizes, federal contractor status, and single establishments versus those reported by firms filing consolidated reports increased at similar rates over the data-collection period.

Reducing the data-collection period would have most substantially affected collection of data from self-filers not on the Component 2 master list. Among this group, only 24.8 percent (2017) and 21.1 percent (2018) of the final set of 4,581 (2017) and 5,099 (2018) establishments reported by the initial deadline. Other establishments not on the initial Component 2 master list but reported by PEOs were reported early: 88.1 percent (2017) and 87.5 percent (2018) of this group were reported by the initial deadline. In contrast, 81.1 percent (2017) and 81.9 percent (2018) of establishments on the initial master list reported by the initial deadline.

Extending the data-collection period was similarly effective across establishments of varying sizes. As of the initial deadline, among the establishments reported for 2017, 84.0 percent had 1–100 employees, 10.8 percent had 100–249 employees, 3.3 percent had 250–499 employees, 1.2 percent had 500–999 employees, and 0.7 percent had 1,000 or more employees (Table 4-4). The size distribution was similar for those 170,209 establishments that were reported for 2017 after the initial deadline, and for which size information is available. 86.4 percent had 1–100 employees, 9.3 percent had 100–249 employees, 2.5 percent had 250–499 employees, 1.1 percent had 500–999 employees, and 0.7 percent had 1,000 or more employees. Thus, extending the deadline was similarly beneficial for including establishments of all sizes. If establishments are examined by time of reporting, as shown in Tables 4-3 through 4-6, at the end of the initial deadline 611,027 establishments with 1–100 employees were reported in 2017 and 645,698 establishments of this size in 2018. Six weeks later, an

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

additional 96,253 of the smallest establishments had been reported for 2017 and an additional 99,435 reported for 2018. The set of the smallest establishments reported by the initial deadline was 80.6 percent (2017) and 81.6 percent (2018) of the final set of establishments of this size, and the six-week extension brought the total to 93.3 percent (2017) and 94.2 percent (2018) of the final set of smallest establishments. Although there are fewer large establishments (1,000 or more employees), a similar pattern holds. As of the initial deadline, 80.7 percent (2017) and 80.0 percent (2018) of the final set of large establishments had been reported, increasing to 94.8 percent (2017) and 94.8 percent (2018) after the six-week extension. Rates are similar for establishments of intermediate sizes.

Almost all establishments were in multi-establishment firms, and response timing varied slightly for single establishments compared to those in multi-establishment firms. Single establishments were slightly more likely to report by the initial deadline—84.0 percent of single establishments in both 2017 and 2018 reported by the initial deadline compared to 81.0 percent (2017) and 81.6 percent (2018) of establishments in multi-establishment firms. Six weeks later, establishments in multi-establishment firms outpaced single-establishment reports slightly, with 91.9 percent (2017) and 92.6 percent (2018) of single establishments and 93.5 percent (2017) and 94.4 percent (2018) of all establishments in firms that filed consolidated reports having been reported.

Response timing was similar for federal contractors and employers that were not federal contractors. As of the initial deadline 345,907 establishments categorized as federal contractors for 2017 (360,606 for 2018) had been reported, or 81.0 percent (2017) and 82.4 percent (2018) of the final set of reported federal-contractor establishments. Federal contractors made up 47.6 percent (2017) and 47.2 percent (2018) of the establishments that reported by the initial deadline. By the end of data collection, 427,291 federal-contractor establishments were reported for 2017 and 437,505 for 2018.

In sum, establishments that reported early and those that reported later had many similar characteristics. There was some fluctuation in the characteristics of reporting establishments in the intermediate weeks of data collection, but there was little change in the cumulative distributions of establishment characteristics over the data-collection period for most of the characteristics in question. For instance, at the initial deadline, 47.2 percent of establishments reported in 2018 were federal contractors compared to 46.8 percent at the end of the data-collection period. The largest variation in filer characteristics across the data-collection period was for self-filers that were not part of the Component 1 data collection and thus self-identified as being eligible. Clearly, shortening the data-collection period would have omitted many eligible establishments.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-3 2017 Characteristics of Component 2 Establishments, by Completion Date (Count)

Cumulative Within Listed Time Period
Establishment Characteristics Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Total 727,474 839,234 845,429 851,318 897,683 727,474 111,760 6,195 5,889 46,365
Size
Less than 100 611,027 707,280 712,418 717,386 758,103 611,027 96,253 5,138 4,968 40,717
100–249 78,810 89,134 89,818 90,486 94,596 78,810 10,324 684 668 4,110
250–499 23,925 26,904 27,129 27,306 28,209 23,925 2,979 225 177 903
500–999 8,461 9,752 9,838 9,888 10,272 8,461 1,291 86 50 384
1,000 or More 5,251 6,164 6,226 6,252 6,503 5,251 913 62 26 251
Federal Contractor
Yes 345,907 406,409 408,567 411,161 427,291 345,907 60,502 2,158 2,594 16,130
No 381,468 432,661 436,698 439,993 470,226 381,468 51,193 4,037 3,295 30,233
Submission Status
Single Establishment 20,035 21,914 22,305 22,598 23,856 20,035 1,879 391 293 1,258
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Cumulative Within Listed Time Period
Establishment Characteristics Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Consolidated Report 707,439 817,320 823,124 828,720 873,827 707,439 109,881 5,804 5,596 45,107
On Component 2 Master List
Yes 706,293 814,304 820,239 825,735 870,357 706,293 108,011 5,935 5,496 44,622
No 21,181 24,930 25,190 25,583 27,326 21,181 3,749 260 393 1,743
Self-Filer 1,135 2,734 2,913 3,125 4,581 1,135 1,599 179 212 1,456
Filed Through PEO 20,046 22,196 22,277 22,458 22,745 20,046 2,150 81 181 287
Submitting Unit
Self-Filed Firm 692,635 800,807 806,873 812,520 855,239 692,635 108,172 6,066 5,647 42,719
Submission by PEO or Alternate Filer 34,839 38,427 38,556 38,798 42,444 34,839 3,588 129 242 3,646

SOURCE: Panel generated from Component 2 employer and establishment files for 2017.

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-4 2017 Characteristics of Component 2 Establishments, by Completion Date (Percent)

Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Total 100 100 100 100 100 100 100 100 100 100
Size
Less than 100 84.0 84.3 84.3 84.3 84.5 84.0 86.1 82.9 84.4 87.8
100–249 10.8 10.6 10.6 10.6 10.5 10.8 9.2 11.0 11.3 8.9
250–499 3.3 3.2 3.2 3.2 3.1 3.3 2.7 3.6 3.0 1.9
500–999 1.2 1.2 1.2 1.2 1.1 1.2 1.2 1.4 0.8 0.8
1,000 or More 0.7 0.7 0.7 0.7 0.7 0.7 0.8 1.0 0.4 0.5
Federal Contractor
Yes 47.6 48.4 48.3 48.3 47.6 47.6 54.2 34.8 44.0 34.8
No 52.4 51.6 51.7 51.7 52.4 52.4 45.8 65.2 56.0 65.2
Submission Status
Single Establishment 2.8 2.6 2.6 2.7 2.7 2.8 1.7 6.3 5.0 2.7
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Consolidated Report 97.2 97.4 97.4 97.3 97.3 97.2 98.3 93.7 95.0 97.3
On Component 2 Master List
Yes 97.1 97.0 97.0 97.0 97.0 97.1 96.6 95.8 93.3 96.2
No 2.9 3.0 3.0 3.0 3.0 2.9 3.4 4.2 6.7 3.8
Self-Filer 5.4 11.0 11.6 12.2 16.8 5.4 42.7 68.8 53.9 83.5
Filed Through PEO 94.6 89.0 88.4 87.8 83.2 94.6 57.3 31.2 46.1 16.5
Submitting Unit
Self-Filed Firm 95.2 95.4 95.4 95.4 95.3 95.2 96.8 97.9 95.9 92.1
Submission by PEO or Alternate Filer 4.8 4.6 4.6 4.6 4.7 4.8 3.2 2.1 4.1 7.9

SOURCE: Panel generated from Component 2 employer and establishment files for 2017.

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-5 2018 Characteristics of Component 2 Establishments, by Completion Date (Count)

Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Total 764,367 883,043 889,007 897,407 935,528 764,367 118,676 5,964 8,400 38,121
Size
Less than 100 645,698 745,133 750,027 757,602 791,020 645,698 99,435 4,894 7,575 33,418
100–249 79,733 93,338 94,007 94,517 97,620 79,733 13,605 669 510 3,103
250–499 24,784 28,161 28,403 28,584 29,549 24,784 3,377 242 181 965
500–999 8,778 10,047 10,154 10,239 10,623 8,778 1,269 107 85 384
1,000 or More 5,374 6,364 6,416 6,465 6,716 5,374 990 52 49 251
Federal Contractor
Yes 360,606 418,780 420,549 422,442 437,505 360,606 58,174 1,769 1,893 15,063
No 403,568 464,003 468,195 474,690 497,732 403,568 60,435 4,192 6,495 23,042
Submission Status
Single Establishment 20,755 22,891 23,216 23,451 24,708 20,755 2,136 325 235 1,257
Consolidated Report 743,612 860,152 865,791 873,956 910,820 743,612 116,540 5,639 8,165 36,864
On Component 2 Master List
Yes 740,357 854,544 860,220 868,299 904,218 740,357 114,187 5,676 8,079 35,919
No 24,010 28,499 28,787 29,108 31,310 24,010 4,489 288 321 2,202
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Self-Filer 1,076 2,607 2,850 3,166 5,099 1,076 1,531 243 316 1,933
Filed Through PEO 22,934 25,892 25,937 25,942 26,211 22,934 2,958 45 5 269
Submitting Unit
Self-Filed Firm 725,365 838,245 844,119 852,504 890,111 725,365 112,880 5,874 8,385 37,607
Submission by PEO or Alternate Filer 39,002 44,798 44,888 44,903 45,417 39,002 5,796 90 15 514

SOURCE: Panel generated from Component 2 employer and establishment files for 2018.

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-6 2018 Characteristics of Component 2 Establishments, by Completion Date (Percent)

Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Total 100 100 100 100 100 100 100 100 100 100
Size
Less than 100 84.5 84.4 84.4 84.4 84.6 84.5 83.8 82.1 90.2 87.7
100–249 10.4 10.6 10.6 10.5 10.4 10.4 11.5 11.2 6.1 8.1
250–499 3.2 3.2 3.2 3.2 3.2 3.2 2.8 4.1 2.2 2.5
500–999 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.8 1.0 1.0
1,000 or More 0.7 0.7 0.7 0.7 0.7 0.7 0.8 0.9 0.6 0.7
Federal Contractor
Yes 47.2 47.4 47.3 47.1 46.8 47.2 49.0 29.7 22.6 39.5
No 52.8 52.6 52.7 52.9 53.2 52.8 51.0 70.3 77.4 60.5
Submission Status
Single Establishment 2.7 2.6 2.6 2.6 2.6 2.7 1.8 5.4 2.8 3.3
Consolidated Report 97.3 97.4 97.4 97.4 97.4 97.3 98.2 94.6 97.2 96.7
On Component 2 Master List
Yes 96.9 96.8 96.8 96.8 96.7 96.9 96.2 95.2 96.2 94.2
No 3.1 3.2 3.2 3.2 3.3 3.1 3.8 4.8 3.8 5.8
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Establishment Characteristics Cumulative Within Listed Time Period
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Published Deadline
(9/30/19)
6 Weeks Later
(10/1/19–11/11/19)
8 Weeks Later
(11/12/19–11/25/19)
12 Weeks Later
(11/26/19–12/23/19)
16 Weeks Later
(12/24/19–2/15/20)
Self-Filer 4.5 9.1 9.9 10.9 16.3 4.5 34.1 84.4 98.4 87.8
Filed Through PEO 95.5 90.9 90.1 89.1 83.7 95.5 65.9 15.6 1.6 12.2
Submitting Unit
Self-Filed Firm 94.9 94.9 95.0 95.0 95.1 94.9 95.1 98.5 99.8 98.7
Submission by PEO or Alternate Filer 5.1 5.1 5.0 5.0 4.9 5.1 4.9 1.5 0.2 1.3

SOURCE: Panel generated from Component 2 employer and establishment files for 2018.

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

COMPONENT 2 MEMBERSHIP CHANGES OVER TIME

Establishments and firms change from year to year. Births and deaths of firms and establishments are expected. Thus, the panel asked whether EEOC data resemble benchmark data in terms of the number of births and deaths of firms over time. If not, data that might appear to have roughly correct totals could actually hide considerable error, failing to capture the tremendous amount of change occurring in the economy. Alternatively, and as will be shown to be the case, the data could show too much variability over time, demonstrating that established firms and establishments are instead appearing as births or deaths, complicating the task of investigators seeking to examine an employer’s historical patterns.

Across all private U.S. firms in 2018, there was an establishment birth rate of 9.2 percent and an establishment death rate of 8.7 percent (Census Bureau, 2020).4 From Census Bureau data, establishment birth and death rates vary by firm size. The smallest firms (less than 20 employees and thus not eligible for the Component 2 data collection) have the highest establishment birth and death rates (11–12% for each in 2018); firms eligible for Component 2 (those with 100 or more employees) have lower establishment birth and death rates.5 In data reported by the Census Bureau, establishment birth and death rates also varied by establishment size, with the highest birth and death rates for the smallest-sized establishments.6

EEOC asked the panel to evaluate the same reporting unit’s response rates for 2018 Component 2 data and 2017 Component 2 data. Since EEOC indicated that enforcement most often occurs at the establishment level, the panel focused on how many establishments that reported in 2017 appeared only in reports from 2017 filers, and how many establishments that reported in 2018 appeared only in reports from 2018 filers. This internal comparison of reporting across 2017 and 2018 allowed insights into possible nonresponse at the establishment level over the two years. Some establishments that appeared in the 2017 Component 2 data did not appear in the 2018 reporting year because the establishments closed their doors (i.e., establishment death). Similarly, we expect that some establishments that appeared in the 2018 reporting year will be newly reported because they opened during 2018 (i.e., establishment birth). However, churn beyond the level seen in the population may arise because a firm failed to report on an establishment in one reporting year of the Component 2 data collection;

___________________

4 The Bureau of Labor Statistics reports quarterly establishment birth and death rates of about three percent (BLS, 2021c).

5https://data.census.gov/cedsci/table?q=BDSTIMESERIES.BDSFAGEFSIZE&tid=BDSTIMESERIES.BDSFAGEFSIZE&hidePreview=true

6https://data.census.gov/cedsci/table?q=BDSTIMESERIES.BDSEAGEESIZE&tid=BDSTIMESERIES.BDSEAGEESIZE&hidePreview=true

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

because of a merger, acquisition, or spinoff of establishments that led to difficulty linking establishments over time; because of shifts in the size of a firm over the two years that changed eligibility; or because of difficulties merging the EEOC data for analysis.

Merging Establishments Over Time

The 2017 and 2018 Component 2 establishment-level data files were merged using the following two steps:7

  1. Merge based on unique matches of the firm’s Employer Identification Number (EIN) and the establishments’ address fields (Address and Address2). Establishments were considered merged if a unique one-to-one match based on the combination of these variables was identified.
  2. For establishments that could not be matched in Step 1, unique matches were searched based on the combination of the firm’s EIN, the establishment’s zip code, and the establishment’s six-digit NAICS code across the two years. Establishments with unique matches based on these criteria were considered merged.

Failure to match across the two files may be due to establishment births or deaths, linking errors, or non-reporting of establishments by firms. Most failure to match across years occurred at the establishment level (more than 94% of establishments were reported by firms whose EINs appeared in both 2017 and 2018 Component 2 data; see Liao, 2021b). When defining eligible establishments, the panel excluded two general categories of reports: Type 2 reports (consolidated reports designed primarily for validation purposes, with data typically duplicated elsewhere in greater detail) and Type 6 reports (reports for establishments with fewer than 50 employees for which the filer chose not to provide detailed employee information). There were 94 firms in 2017 and 153 firms in 2018 that filed only Type 2 reports and were thus excluded from the analysis. Additionally, there were 198 firms in 2017 and 224 firms in 2018 that filed only Type 6 reports and were thus excluded. In addition, there were firms and establishments with extremely large outlying numbers of employees. The panel adopted an outlier rule at the firm level, excluding any firm and its associated establishments if the total number of employees exceeded a threshold of 1.4 million (the size of the largest U.S. employer). This rule resulted in the exclusion of an additional 33 firms in 2017 and 29 firms in 2018. This section examines

___________________

7 EEOC data include an establishment identification number, but the numbering is not consistent over time and thus inappropriate for merges.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

a total of 897,770 establishments in 2017 and 935,610 establishments in 2018, including a small number of firms that reported no employees in the Component 2 data collection.

Additionally, some establishments reported differing characteristics across the two years. Establishments were sometimes reported as federal contractors in one year but not the other. Furthermore, establishment size can change over time, and thus the size of establishments was averaged across the two filing years. Further research could examine this decision in more detail.

By Sample Sources

For establishments with at least one employee, there were 715,224 establishments that appeared in both 2017 and 2018 Component 2 data, while 182,546 establishments appeared only in 2017 data and 220,386 appeared only in 2018 data (Table 4-7). Thus, 80 percent of 2017 establishments could be matched with 2018 establishments, and 76 percent of 2018 establishments could be matched with 2017 establishments. Among the remainder, it is unknown how many existed in only one year (i.e., there was a birth or death in 2018) and how many could not be matched because of poor data.

The failures to match are two to three times that of establishment birth and death rates reported by the Census Bureau across all establishments. There are several potential explanations for the large number of nonmatches: some firms may have chosen to participate in the Component 2 data collection in only one year; some may have reported establishments on Type 6 reports in one but not both years; and errors linking establishments across years may have led to non-matches, even for establishments that existed in both years.

Appearance in only one of the data-collection years varies notably by whether establishments were single establishments or part of multi-establishment firms. Single establishments were much more likely to appear in both 2017 and 2018 data than were those that were part of multi-establishment firms—90–95 percent of single establishments appeared in both years, compared to 75–80 percent of establishments from multi-establishment firms. It is possible that address and NAICS codes were more consistently identified for single establishments over the two years, facilitating greater matching of units across time. It is also possible that establishments within firms were created or closed, or that firms inadvertently omitted establishments from their reports. Other characteristics of establishments differed less between those that appeared in one or both years of Component 2 data. For instance, establishments in firms that were not part of the Component

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-7 Characteristics of Component 2 Establishments, by Year

Total Establishments

2017
Total Establishments

2018
In 2017 Only In 2018 Only
N % of 2017 N % of 2018
Total 897,770 935,610 182,546 20.3 220,386 23.6
Submission Status
Single establishment 23,643 24,708 1,394 5.9 2,459 10.0
Consolidated report 874,127 910,902 181,152 20.7 217,927 23.9
Filer Is Federal Contractor
Yes 423,744 437,661 92,094 21.7 106,011 24.2
No 473,986 497,887 90,414 19.1 114,315 23.0
Missing 40 62 38 95.0 60 96.8
On Component 2 Master List
Yes 870,436 904,293 178,467 20.5 212,324 23.5
No 27,334 31,317 4,079 14.9 8,062 25.7
Submitting Unit
Self-filed firm 855,538 890,193 173,250 20.3 207,905 23.4
Submission by PEO or alternate filer 42,232 45,417 9,296 22.0 12,481 27.5

SOURCE: Panel generated from Component 2 employer and establishment files for 2017 and 2018.

NOTE: Excludes firms reporting more than 1.4 million employees and establishments covered in Type 6 reports. If the data reported for 2017 and 2018 conflict with regard to how establishments are categorized, then the 2018 data are used.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

2 master list were more likely to appear only in 2018 data (25.7% of these establishments were new in 2018) than to appear only in 2017 data (14.9% of these establishments reported for 2017 but not 2018). Federal-contractor establishments had similar rates to those that were not federal contractors—for instance, 22 percent of federal contractors in 2017 data appeared only in 2017 data, compared to 19 percent of reporting establishments that were not federal contractors in 2017.

By Establishment Size

Examining changes over time by establishment size poses challenges because establishments could have changed size over the two years for which data were collected. As such, establishment size was averaged over the two reporting years. Single-establishment firms with fewer than 100 employees are not eligible for filing the Component 2 instrument and thus are not required to appear in the data file; multi-establishment firms with a total size greater than 100 employees but with establishments with fewer than 100 employees are required to report on their establishments. Thus, it is not surprising to see reduced consistency in reporting for establishments with fewer than 100 employees compared to larger firms. This pattern is partially seen in Table 4-8.

Establishment size has a curvilinear relationship with the appearance of an establishment in both reporting years. Of the smallest establishments (less than 100 employees), 21.4 percent appeared only in 2017 data but not in 2018 data, and 24.7 percent appeared only in 2018 data but not in 2017 data. Midsize establishments—those with 100–499 employees—were slightly less likely to appear in only one year of reporting; for instance, only about 14 percent of midsized establishments appeared only in 2017 data but not in 2018 data, and around 17 percent appeared only in 2018 data but not in 2017 data. Larger establishments were slightly more likely to appear in a single reporting year—more than 19 percent of establishments with 500 or more employees appeared in only one reporting year. Table 4-8 also includes economy-wide establishment birth and death rates from the Census Bureau’s Business Dynamics Statistics (BDS). For establishments with at least 100 employees (the most directly comparable to those required to file EEOC reports), birth and death rates fail to exceed two percent in either year.

By Industry Sector

There is dramatic variation in the presence of establishments across both years of the Component 2 data collection across NAICS sectors (Table 4-9). Other than the unclassified sector, the largest proportion of exiting or entering establishments is in the management of companies and

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-8 Years of Appearance of Component 2 Establishments, by Establishment Size

Mean Establishment Size in 2017–2018 Total Establishments 2017 Total Establishments 2018 In 2017 Only In 2018 Only 2017 Census Bureau Business Dynamics Statistics (BDS) Establishment 2018 Census Bureau BDS Establishment
N % of 2017 N % of 2018 Entry Rate Exit Rate Entry Rate Exit Rate
Total 897,705 935,546 182,487 20.3 220,328 23.6
1–4 15.896 15.041 15.435 14.548
5–9 3.634 3.342 3.363 3.216
10–19 2.843 2.597 2.436 2.402
20–99 1.98 2.09 1.792 1.774
Less than 100 756,725 789,838 162,009 21.4 195,122 24.7
100–249 94,680 97,841 12,994 13.7 16,155 16.5 1.100 1.142 1.128 0.909
250–499 28,410 29,598 4,027 14.2 5,215 17.6
500–999 10,538 10,761 1,890 17.9 2,113 19.6 1.054 0.732 0.801 0.498
1,000 or More 7,352 7,508 1,567 21.3 1,723 22.9 0.587 0.600 0.383 0.396

SOURCE: Panel generated from Component 2 establishment files for 2017 and 2018; https://www2.census.gov/programs-surveys/bds/tables/timeseries/bds2019_ez.csv. Excludes firms reporting more than 1.4 million employees and establishments covered in Type 6 reports.

NOTE: Establishments not reporting the number of employees in both 2017 and 2018 are omitted.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-9 Years of Appearance of Component 2 Establishments, by Industry

NAICS Code Label Total In 2017 Only In 2018 Only
2017 2018 N % of 2017 N % of 2018
11 Agriculture, Forestry, Fishing, and Hunting 3,172 3,119 864 27.2 811 26.0
21 Mining, Quarrying, and Oil and Gas Extraction 5,981 6,329 1,612 27.0 1,960 31.0
22 Utilities 6,499 6,637 789 12.1 927 14.0
23 Construction 16,144 17,325 4,290 26.6 5,471 31.6
31–33 Manufacturing 67,028 70,211 12,991 19.4 16,174 23.0
42 Wholesale Trade 46,992 48,844 7,083 15.1 8,935 18.3
44–45 Retail Trade 200,603 201,073 29,717 14.8 30,187 15.0
48–49 Transportation and Warehousing 30,038 31,792 4,796 16.0 6,550 20.6
51 Information 32,564 33,525 5,840 17.9 6,801 20.3
52 Finance and Insurance 92,660 96,169 18,375 19.8 21,884 22.8
53 Real Estate and Rental and Leasing 37,271 38,952 8,692 23.3 10,373 26.6
54 Professional, Scientific, and Technical Services 57,219 61,494 16,206 28.3 20,481 33.3
55 Management of Companies and Enterprises 8,323 8,469 3,112 37.4 3,258 38.5
56 Administrative and Support and Waste Management and Remediation Services 40,976 44,469 12,279 30.0 15,772 35.5
61 Educational Services 5,537 5,686 1,378 24.9 1,527 26.9
62 Health Care and Social Assistance 119,160 127,032 26,800 22.5 34,672 27.3
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
NAICS Code Label Total In 2017 Only In 2018 Only
2017 2018 N % of 2017 N % of 2018
71 Arts, Entertainment, and Recreation 7,773 8,443 1,239 15.9 1,909 22.6
72 Accommodation and Food Services 84,262 89,054 17,130 20.3 21,922 24.6
81 Other Services (except Public Administration) 26,902 28,179 4,766 17.7 6,043 21.4
92 Public Administration 770 991 182 23.6 403 40.7
99 Unclassified 7,896 7,817 4,405 55.8 4,326 55.3

SOURCE: Panel generated from Component 2 employer and establishment files for 2017 and 2018. Excludes firms reporting more than 1.4 million employees and establishments covered in Type 6 reports.

NOTE: Establishments not reporting the number of employees in both 2017 and 2018 are omitted.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

enterprises sector (NAICS 55), in which just under 40 percent of establishments observed in 2017 data appear only in 2017, with a similar percentage in the reverse direction. Similarly, about one-third of the establishments in the administrative and support and waste management and remediation services (NAICS 56) observed in each year appear only in that year. The highest consistency rates are in the retail trade (NAICS 44–45), wholesale trade (NAICS 42), and utilities (NAICS 22) sectors; in each of these sectors, around 15 percent of each year’s filers appear in only one reporting year.

There are multiple reasons why establishments may appear or disappear from one year to another. In addition to births and deaths, there can be mergers or consolidations that reduce the number of establishments; or splits or divestitures that increase the number of establishments. It is also possible that a firm’s motivation and capacity to respond may change from year to year. For example, if key reporting staff turnover, this may reduce a firm’s capacity to respond (Tomaskovic-Devey et al., 1994). Nevertheless, the challenges that the panel faced in identifying whether the same establishment was reported across the two years of the Component 2 data collection provide direction for EEOC in future data collections. First, although there are unique identifiers attached to firms for reporting purposes, the establishments within a firm do not necessarily use unique identifiers consistently in reporting. Assigning a unique identifier to establishments as well as firms will facilitate examination of data for the same firm over time, allowing EEOC to evaluate whether pay gaps are systemic. Second, individual establishment identifiers will permit EEOC to evaluate whether there is differential nonresponse on establishments within firms, and to more efficiently target nonresponse follow-up efforts.

CUMULATIVE EFFECT OF INCOMPLETE FIRM LISTS AND NONRESPONSE ON REPRESENTATION

There are two general approaches to understanding coverage compared to external benchmark data. The first approach—and the one taken for this report—is to compare characteristics of units from the data collection being evaluated (2017 and 2018 Component 2 data collections) to an external set of high-quality data from a well-regarded source such as the Census Bureau or the Bureau of Labor Statistics (BLS) that is considered to have better coverage properties. Ideally, these two sources would have identical measurement properties on key constructs of interest, be measured over the same timeframe, and apply (or be made to apply) to the same population. These ideal conditions are often not met, but comparisons can still be useful. The second approach is to link individual units from one data source (with lower coverage) to a second data source (with better or different coverage). This micro-level approach to evaluating coverage is beyond the

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

scope of this report. However, EEOC has been in talks with BLS to link the EEO-1 frame with the BLS Business Register as part of a data modernization plan (EEOC, 2021i).

To compare Component 2 data to external data from the Census Bureau, firms that reported zero employees across all establishments were excluded. This resulted in the total exclusion of 338 firms in 2017 and 411 firms in 2018, yielding an analytic dataset of 63,762 firms and 897,683 establishments for 2017 and 66,449 firms and 935,528 establishments for 2018.

The aggregate coverage evaluation to external benchmark data involved two sources. First, the panel examined how well Component 2 respondents aligned with Component 1 respondents for each year. Although the Component 1 data collection formed a frame for the Component 2 data collection, these collections were separated in time and collected by different data collectors. Furthermore, many firms and establishments filed Component 2 data but did not file Component 1 data. As such, the coverage of Component 2 data was expected to increase relative to Component 1 data, except that (1) Component 1 data contained a greater number of firms and establishments with 0 employees (deleted for these analyses), and (2) federal contractors with at least 50 employees but fewer than 100 employees were required to file Component 1 instruments but not Component 2 instruments. That is, Component 1 and 2 instruments have largely similar eligibility requirements, except that federal contractors with at least 50 employees were required to participate in the Component 1 data collection but not in the Component 2 data collection. It is unknown whether this increase will vary across characteristics of firms and establishments (except for an expected decrease in federal contractors).

Next, Component 2 filers were compared to distributions obtained from outside business registers as external benchmarks. The first external benchmark comes from the Census Bureau’s Business Register, from which tabulations are made for the BDS program (Census Bureau, 2021d). Because the Census Bureau permits tabulations by size of firm, information from the Census Bureau was used as the benchmark on firm-level data. Establishments with no employees were excluded from BDS tabulations. The second external benchmark comes from BLS’s Quarterly Census of Employment and Wages (QCEW). Because BLS permits tabulations by establishment size, BLS information was added to the benchmarks for the establishment level. Establishments with no employees were included in the QCEW tabulations if they had a history of submitting unemployment insurance.

While data from these external benchmarks are useful for examining Component 2 data coverage, each source of data is based on different assumptions:

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
  • EEOC data differ from BDS and QCEW in covering only firms with 100 or more employees. Firms with fewer than 100 employees can voluntarily submit data, but those data represent only a small proportion of all small firms and are excluded from EEOC’s target population. For this reason, valid comparisons require imposing similar size restrictions on EEOC, BDS, and QCEW data. Both BDS and QCEW provide statistics based on size but with important differences to be described later.
  • While EEOC bases eligibility on firm size, the data do not directly produce a firm-level database. There is a filer-level database (but a filer may be responding for multiple firms) and an establishment-level database (but a firm may consist of multiple establishments). It is possible to simulate a firm-level database by aggregating across establishments, but firms cannot always be uniquely identified (e.g., a firm might be recorded with the ID of a submitting PEO instead of that of the specific firm).
  • QCEW and BDS provide data on the size of establishments but not of establishment size by firm size. Thus, one can select establishments with 100 or more employees, but cannot select firms that collectively have 100 employees or more and then examine the size of establishments within these firms. Since QCEW and BDS do not impose a size limit (except that BDS omits establishments with no employees), both QCEW and BDS will show a larger representation of small firms than will EEOC. On the other hand, if a limit of 100 or more employees per establishment is applied, then more restrictive conditions are imposed on QCEW than are imposed by EEOC, excluding establishments that should be covered by EEOC. Limiting EEOC data to establishments with 100 or more employees would increase comparability with benchmark data for these firms but would exclude over 80 percent of establishments reported in Component 2 data.
  • Size is measured differently across the three databases. EEOC and QCEW measure size separately for each establishment for each year (i.e., QCEW uses March employment [BLS, 2021d], while BDS uses the average firm size across two years [Census Bureau, 2021e]).
  • BDS differs in its method of counting firms and establishments in industry-level tabulations. If a firm has multiple establishments, industry-level tabulations count the firm in every industry within which the firm has an establishment. Thus, the overall total number of firms may not match the sum of the number of firms in the detailed industry-based statistics because the firm is counted more than once (i.e., once for each industry). Establishments form a
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
  • better measure of industry-level coverage because industry is assigned at the establishment level. Thus, establishments are counted only once.8
  • BLS systematically collects lists of establishments from most multilocation employers through the Multiple Worksite Report, based on Unemployment Insurance accounts.9 This may allow BLS to better identify establishments than does the Census Bureau’s BDS, so it reports on more small establishments and fewer large establishments.
  • There are also differences in how industry codes are assigned. EEOC filers submit their own codes, while BDS10 and QCEW (BLS, 2021e) assign codes to establishments.

To evaluate how well Component 2 data reflect the comparison groups in both 2017 and 2018, we use a coverage ratio:

CoverageRatio _ U n i t s t = 100 * # U n i t s C o m p o n e n t 2 t # U n i t s C o m p a r i s o n t

Values above 100 indicate that Component 2 data contain more units—either firms or establishments—than the comparison group. Values below 100 indicate that Component 2 data contain fewer units than the comparison group. The comparison group is either the units from Component 1 data, the number of firms from the Census Bureau’s BDS, or the number of establishments from the Census Bureau’s BDS or BLS’s QCEW, and is calculated separately for 2017 and 2018. The calculations are labeled as coverage ratios, but the set of responding employers and establishments reflects the joint process of sample frame coverage and nonparticipation.

Size: Differential Coverage of Firms and Establishments by the Component 1 and Component 2 Instruments

Starting with the firm-level comparison between Component 1 coverage and Component 2 coverage, more firms filed Component 2 instruments

___________________

8 “… in the BDS, geography and industry are defined at the establishment level, and not the firm level. So firms with multiple establishments operating in different geographies or industries will show up in the total firm data for each geography or industry in which they operate at least one establishment” (Census Bureau, 2021e).

9https://www.bls.gov/respondents/mwr/

10 BDS methodology largely follows county business patterns (Census Bureau, 2021g), which come from the economic census (https://bhs.econ.census.gov/ombpdfs/). NAICS codes are assigned by the Census Bureau from NAICS classification surveys, including reports about business primary activities.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

than Component 1 instruments in almost all size categories for both years (see Table 4-10). The overall coverage ratio was 91.7 in 2017 and 90.9 in 2018 for all firms, but this includes firms that were not in the target population for the Component 2 data collection. In particular, firms with fewer than 100 employees were not required to participate in the Component 2 data collection (see Chapter 2), but some of these firms (federal contractors with 50 or fewer employees) were part of the target population for the Component 1 data collection. As such, the panel focused only on firms with 100 or more employees. The coverage ratio for these firms overall was 101.8 for 2017 and 100.7 for 2018. The most substantial increase was in firms with 1,000 or more employees, with coverage ratios of 103.1 and 103.8 in 2017 and 2018, respectively. The panel suspects that some of the overcoverage of large firms resulted from PEO reporters. As noted by NORC in its report to EEOC, “some PEOs do not differentiate the independent employers for which they provide services while others … do” (EEOC, 2020e, p. 74). The coverage ratio relative to Component 1 data is about 100 in both 2017 and 2018 for firms with 500–999 employees. Smaller firms—those with 100–499 employees—varied over the two years, with a coverage ratio of 101.8 in 2017 and of 100.2 in 2018.

The coverage ratios for establishments varied across the size groups (see Table 4-11). The smallest establishments were eligible for inclusion in the Component 2 data collection if they were part of multi-establishment firms. Despite exclusion of the smallest firms from the Component 2 data collection, the coverage ratio for establishments with fewer than 100 employees was 116.6 in 2017 and 116.8 in 2018. Coverage ratios for establishments with 100 or more employees were less than 100 for both years overall, with about five to six percent fewer establishments reported in Component 2 data than in Component 1 data, but exceeded 100 for the largest establishments. The number of establishments with 1,000 or more employees was higher in Component 2 data than in Component 1 data by 1.8 percent in 2017 and 1.5 percent in 2018.

Overall, Component 2 data-collection response compliance was at least as high as Component 1 compliance. Thus, there does not seem to have been notably higher refusal to participate in EEOC’s initial pay-data collection effort than in the standard Component 1 data collection. The over-reporting of very large firms likely arose due to two sources of measurement error: errors in reporting total employment that were not recognized or corrected at the data-processing stage, and the bundling of multiple firm reports by PEOs.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-10 Number of Firms in EEO-1 Data and Census Bureau BDS: 2017 and 2018

Firm Size Category Component Coverage Ratio (100* C2/C1) Census Bureau BDS Number of Firms Coverage Ratio (100* C2/Census Bureau BDS)
C1 C2
2017
Total 69,561 63,762 91.7 5,264,715 1.2
1–99 7,034 139 2.0 5,151,159 0.0
100 or more (subtotal) 62,527 63,623 101.8 113,556 56.0
100–499 46,759 47,583 101.8 90,537 52.6
500–999 7,385 7,401 100.2 10,845 68.2
1,000 or more 8,383 8,639 103.1 12,174 71.0
2018
Total 73,064 66,449 90.9 5,292,681 1.3
1–99 7,154 73 1.0 5,178,570 0.0
100 or more (subtotal) 65,910 66,376 100.7 114,111 58.2
100–499 49,468 49,558 100.2 91,623 54.1
500–999 7,727 7,774 100.6 10,915 71.2
1,000 or more 8,715 9,044 103.8 11,573 78.1

SOURCE: Component 2 employer files for 2017 and 2018; https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_fz.csv

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees.

Industry: Differential Coverage of Firms and Establishments by the Component 1 and Component 2 Instruments

Focusing only on firms with at least 100 employees, firms are slightly undercovered in almost every industry compared to Component 1 data, with coverage ratios between 90 and 100 for almost every industry (Table 4-12). NAICS codes for firms overcovered relative to Component 1 data in both years are NAICS 54 (professional, scientific, and technical services), 55 (management of companies and enterprises), 56 (administrative and support and waste management and remediation services), and 92 (public administration), with coverage ratios between 100.7 and 180.2. NAICS codes 54, 55, and 56 are the NAICS sectors for PEO reporters; as such, this overrepresentation is likely occurring because of the extensive use of PEO

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-11 Number of Establishments, by Data Source: 2017 and 2018

Size of Establishments Component Coverage Ratio BLS QCEW QCEW Coverage Ratio Census Bureau BDS BDS Coverage Ratio
C1 C2 (100* C2/C1) (100* C2/QCEW) (100*C2/BDS)
2017
Total 797,618 897,683 112.5 9,420,257 9.5 7,071,934 12.7
Fewer than 100 650,441 758,098 116.6 9,239,146 8.2 6,887,265 11.0
Total 100 or More 147,177 139,585 94.8 181,111 77.1 182,484 76.5
100–499 130,374 123,040 94.4 163,758 75.1 164,723 74.7
500–999 10,604 10,232 96.5 11,352 90.1 12,452 82.2
1,000 or More 6,199 6,313 101.8 6,001 105.2 7,494 84.2
2018
Total 829,302 935,528 112.8 9,596,646 9.7 7,098,769 13.2
Fewer than 100 677,355 791,259 116.8 9,411,793 8.4 6,910,182 11.5
Total 100 or More 151,947 144,269 94.9 184,853 78.0 188,587 76.5
100–499 134,528 127,242 94.6 167,138 76.1 167,879 75.8
500–999 11,056 10,570 95.6 11,524 91.7 12,880 82.1
1,000 or More 6,363 6,457 101.5 6,191 104.3 7,828 82.5

SOURCE: Panel generated from Component 2 employer and establishment files for 2017 and 2018: https://data.bls.gov/cew/apps/table_maker/v4/table_maker.htm#type=12&year=2017&size=0,1,2,3,4,5,6,7,8,9&agg=21&supp=0 and https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_ez.csv (for 2017); and https://data.bls.gov/cew/apps/data_views/data_views.htm#tab=Tables and https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_ez.csv (for 2018).

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, and firms and establishments with 0 employees. For 2017 Component 1, no firms, no establishments, and no employees were excluded due to outliers; for 2017 Component 2, 33 firms, 332 establishments, and 892,955,294 employees were excluded due to outliers. For 2018 Component 1, no firms, no establishments, and no employees were excluded due to outliers; for 2018 Component 2, 29 firms, 293 establishments, and 354,413,111 employees were excluded due to outliers.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

reporters and of PEOs reporting their own NAICS as belonging to the firms for which they were filing.

Compared to Component 1 data, coverage of establishments varied notably across industries (Table 4-13). First, among establishments with at least one employee, there were more establishments reported in Component 2 data than in Component 1 data (coverage ratio greater than 100) for most industries. In particular, there were more establishments reported in Component 2 data than in Component 1 data in both years for 14 NAICS codes: 22 (utilities), 31–33 (manufacturing), 42 (wholesale trade), 44–45 (retail trade), 51 (information), 52 (finance and insurance), 53 (real estate and rental and leasing), 54 (professional, scientific, and technical services), 55 (management of companies and enterprises), 56 (administrative and support and waste management and remediation services), 62 (health care and social assistance), 71 (arts, entertainment, and recreation), 81 (other services except public administration), and 92 (public administration). Coverage was higher for Component 2 data in 2017 in NAICS codes 23 (construction) and 72 (accommodation and food services), and was higher in 2018 for NAICS codes 21 (mining, quarrying, and oil and gas extraction) and 48–49 (transportation and warehousing). Only two sectors were undercovered in both years, with the largest undercoverage in NAICS code 61 (educational services), followed by 11 (agriculture, forestry, fishing, and hunting). These patterns are likely linked to confusion in the definitions of firms and establishments in both Components 1 and 2 data associated with PEO filers.

Size: Component 2 Data Versus External Benchmarks

Looking next at coverage of firms relative to Census Bureau benchmark numbers, firms of all sizes were systematically underrepresented (Table 4-10). About two in five firms with 100 or more employees were missing from Component 2 data. Firms with fewer than 100 employees were excluded from the Component 2 target population.

For firms with at least 100 employees, the panel estimated that about 40 percent were missing from the Component 2 data, depending on firm size. Overall, the coverage ratio for firms with 100 or more employees was 56.0 in 2017 and 58.2 in 2018. Among the target population of firms with 100 or more employees, undercoverage was more severe for smaller firms (those with 100–499 employees) than for larger firms. The coverage ratio for these smaller firms was 52.6 for 2017 and 54.1 for 2018. The coverage ratio for firms with 500–999 employees was 68.2 in 2017 and 71.2 in 2018, and for firms with 1,000 or more employees, it was 71.0 in 2017 and 78.1

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-12 Number of Firms, by Industry and Data Source: 2017 and 2018

2017 2018
NAICS Code Industry Component Coverage Ratio (100*C2/C1) Census Bureau BDS Coverage Ratio (100*C2/Census Bureau) Component Coverage Ratio (100*C2/C1) Census Bureau BDS Coverage Ratio (100*C2/Census Bureau)
C1 C2 C1 C2
11 Agriculture, Forestry, Fishing, and Hunting 617 587 95.1 297 197.6 660 599 90.8 293 204.4
21 Mining, Quarrying, and Oil and Gas Extraction 472 473 100.2 1,002 47.2 500 494 98.8 1,053 46.9
22 Utilities 390 381 97.7 558 68.3 396 395 99.7 542 72.9
23 Construction 3,845 3,619 94.1 7,503 48.2 4,126 3,776 91.5 7,875 47.9
31–33 Manufacturing 10,965 10,373 94.6 16,476 63.0 11,324 10,796 95.3 16,399 65.8
42 Wholesale Trade 2,976 2,939 98.8 11,401 25.8 3,207 3,058 95.4 11,183 27.3
44–45 Retail Trade 4,018 3,721 92.6 10,635 35.0 4,177 3,799 91.0 10,578 35.9
48–49 Transportation and Warehousing 1,812 1,687 93.1 6,347 26.6 1,922 1,749 91.0 6,299 27.8
51 Information 1,509 1,435 95.1 3,696 38.8 1,579 1,516 96.0 3,658 41.4
52 Finance and Insurance 3,862 3,686 95.4 6,016 61.3 3,904 3,775 96.7 5,985 63.1
53 Real Estate and Rental and Leasing 1,269 1,205 95.0 3,827 31.5 1,343 1,259 93.7 3,812 33.0
54 Professional, Scientific, and Technical Services 6,051 6,389 105.6 11,355 56.3 6,372 6,843 107.4 11,506 59.5
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
2017 2018
NAICS Code Industry Component Coverage Ratio (100*C2/C1) Census Bureau BDS Coverage Ratio (100*C2/Census Bureau) Component Coverage Ratio (100*C2/C1) Census Bureau BDS Coverage Ratio (100*C2/Census Bureau)
C1 C2 C1 C2
55 Management of Companies and Enterprises 2,584 2,683 103.8 16,888 15.9 2,701 2,719 100.7 16,558 16.4
56 Administrative and Support and Waste Management and Remediation Services 3,212 5,319 165.6 12,345 43.1 3,621 5,770 159.3 12,449 46.3
61 Educational Services 1,420 1,318 92.8 5,035 26.2 1,541 1,350 87.6 5,072 26.6
62 Health Care and Social Assistance 11,320 10,711 94.6 22,677 47.2 11,755 10,972 93.3 22,975 47.8
71 Arts, Entertainment, and Recreation 1,217 1,175 96.5 3,605 32.6 1,284 1,216 94.7 3,658 33.2
72 Accommodation and Food Services 2,689 2,787 103.6 12,599 22.1 2,982 2,866 96.1 12,560 22.8
81 Other Services (except Public Administration) 2,168 2,122 97.9 5,950 35.7 2,320 2,172 93.6 5,929 36.6
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
92 Public Administration 131 236 180.2 164 245 149.4
99 Unclassified 0 777 32 1,007 3146.9

SOURCE: Panel generated from Component 2 employer files for 2017 and 2018. https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_sec_fz.csv

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, establishments with 0 employees, and firms with fewer than 100 employees. In Census Bureau tabulations, establishments with fewer than 100 employees were excluded. For 2017 Component 1, no firms, no establishments, and no employees were excluded due to extreme outliers (i.e., reporting more employees than the largest U.S. employer). For 2017 Component 2, 33 firms, 332 establishments, and 892,955,294 employees were excluded due to extreme outliers. For 2018 Component 1, no firms, no establishments, and no employees were excluded due to extreme outliers. For 2018 Component 2, 29 firms, 293 establishments, and 354,413,111 employees were excluded due to extreme outliers. Firms with multiple establishments in multiple industries may be counted more than once in the Census Bureau statistics. These statistics are not fully comparable to the Census Bureau statistics in Table 4-10 or the EEOC data presented here.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

TABLE 4-13 Number of Establishments, by Industry and Data Source: 2017 and 2018

NAICS Code Industry 2017 2018
Component Coverage Ratio

(100* C2/C1)
Census Bureau BDS Coverage Ratio

(100* C2/BDS)
Component Coverage Ratio

(100* C2/C1)
Census Bureau BDS Coverage Ratio

(100* C2/BDS)
C1 C2 C1 C2
11 Agriculture, Forestry, Fishing and Hunting 3,289 3,157 96.0 930 339.5 3,363 3,110 92.5 873 356.2
21 Mining, Quarrying, and Oil and Gas Extraction 6,090 5,984 98.3 6,729 88.9 5,671 6,326 111.5 6,732 94.0
22 Utilities 6,018 6,321 105.0 12,251 51.6 6,040 6,637 109.9 12,597 52.7
23 Construction 16,117 16,301 101.1 19,410 84.0 17,511 17,291 98.7 20,126 85.9
31–33 Manufacturing 65,469 66,852 102.1 53,525 124.9 67,558 70,153 103.8 54,993 127.6
42 Wholesale Trade 31,747 46,907 147.8 100,445 46.7 33,037 48,805 147.7 100,508 48.6
44–45 Retail Trade 164,455 202,176 122.9 375,030 53.9 165,505 201,007 121.5 371,420 54.1
48–49 Transportation and Warehousing 29,385 28,389 96.6 52,339 54.2 29,843 31,767 106.4 52,708 60.3
51 Information 28,631 32,567 113.7 70,813 46.0 27,303 33,519 122.8 73,100 45.9
52 Finance and Insurance 80,627 92,323 114.5 215,028 42.9 81,331 96,084 118.1 216,525 44.4
53 Real Estate and Rental and Leasing 27,690 37,145 134.1 86,261 43.1 27,141 38,943 143.5 87,739 44.4
54 Professional, Scientific, and Technical Services 52,831 56,996 107.9 94,405 60.4 56,191 61,427 109.3 93,150 65.9
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
55 Management of Companies and Enterprises 8,284 8,379 101.1 44,950 18.6 8,221 8,414 102.3 43,960 19.1
56 Administrative and Support and Waste Management and Remediation Services 37,620 41,010 109.0 75,673 54.2 39,883 44,419 111.4 73,247 60.6
61 Educational Services 6,218 5,533 89.0 13,747 40.2 6,401 5,676 88.7 13,736 41.3
62 Health Care and Social Assistance 109,848 118,864 108.2 220,112 54.0 118,803 126,908 106.8 226,832 55.9
71 Arts, Entertainment, and Recreation 7,272 7,727 106.3 13,973 55.3 8,146 8,431 103.5 14,238 59.2
72 Accommodation and Food Services 79,562 84,127 105.7 172,810 48.7 89,030 88,982 99.9 174,944 50.9
81 Other Services (except Public Administration) 21,522 26,933 125.1 52,736 51.1 23,141 28,170 121.7 52,606 53.5
92 Public Administration 619 747 120.7 683 990 144.9
99 Unclassified 0 8,856 0 8,272

SOURCE: Component 2 employer and establishment files for 2017 and 2018. https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_sec_fz.csv

NOTE: Excludes firms reporting more than 1.4 million employees, establishments covered in Type 6 reports, establishments with 0 employees, and firms with fewer than 100 employees. In Census Bureau statistics, establishments with fewer than 100 employees were excluded. For 2017 Component 1, no firms, no establishments, and no employees were excluded due to extreme outliers. For 2017 Component 2, 33 firms, 332 establishments, and 892,955,294 employees were excluded due to extreme outliers. For 2018 Component 1, no firms, no establishments, and no employees were excluded due to extreme outliers. For 2018 Component 2, 29 firms, 293 establishments, and 354,413,111 employees were excluded due to extreme outliers. Firms with multiple establishments in multiple industries may have been counted more than once in Census Bureau statistics. These statistics are not fully comparable to the Census Bureau statistics in Table 4-11 or EEOC data presented here.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

in 2018. Thus, there appears to be a notable undercoverage of firms and, therefore, a notable undercoverage of establishments and employees in the Component 2 data collection.

Evaluating the coverage of establishments is complicated by two factors: the Component 2 data collection only included establishments belonging to firms with at least 100 employees (making comparisons with external databases difficult), and filers may have submitted either Type 6 or Type 8 reports for establishments with fewer than 50 employees. (Type 6 reports contain only the total number of employees, without specifying sex, race/ethnicity, or pay bands.)

For establishments with 100 or more employees, Table 4-11 shows there were 139,585 establishments in the 2017 Component 2 data compared with 182,484 according to BDS, resulting in a coverage rate of 76.5 percent. Enlarging the focus to include all eligible establishments, separate BDS statistics show there were 1,681,167 establishments in 2017 associated with firms having at least 100 employees (not in tables).11 Comparable Component 2 data counts included 897,683 establishments with complete data and 436,197 establishments in Type 6 reports, for a total of 1,333,880 establishments, or 79 percent of the BDS count. These two coverage rates are fairly similar, but they hide the fact that much of the Component 2 dataset is not usable for pay-equity analysis because filers exercised their option to submit only total number of employees (i.e., Type 6 reports) without also specifying sex, race/ethnicity, or pay bands (i.e., Type 8 reports). Focusing only on establishments with pay data, the Component 2 data count of 897,683 establishments was 53 percent of the BDS count. Thus, in terms of providing usable data, the Component 2 data collection shows undercoverage of more than 40 percent for both firms and establishments.

Benchmark data by size of establishment for eligible firms (those with at least 100 employees) cannot be obtained from publicly available data. This is particularly relevant when trying to evaluate whether larger versus smaller establishments are more likely to be included in firm reports. Generally, smaller establishments are eligible for inclusion in the Component 2 data collection if they are part of multi-establishment firms with at least 100 employees across all establishments. However, QCEW and BDS do not permit joint identification of firms with at least 100 employees and the size of the establishment. Thus, reliable benchmark data by size of establishment among firms with at least 100 employees cannot be calculated. Many of

___________________

11 See https://www2.census.gov/programs-surveys/bds/tables/time-series/bds2019_fz.csv, which allows subsetting to firms with >100 employees for 2017 and 2018. The comparable statistics for 2018 are 1,690,034 establishments in BDS; 935,610 establishments with full data in Component 2 data; and 442,629 Type 6 reports, resulting in coverage rates of 82 percent if all establishments are counted and 55 percent if only establishments with usable data are counted.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

the smallest establishments (<100 employees) are within firms with fewer than 100 employees; unfortunately, in the benchmark data, it is impossible to differentiate these smallest establishments located within ineligible firms from small establishments located within eligible firms. Thus, although this group of firms is presented in the tables, it is not discussed relative to the benchmark data, except to note that Component 2 data provide poor coverage of such establishments, as might be anticipated.

By definition, any establishment with 100 or more employees will be in a firm with more than 100 employees and thus meet the size eligibility criteria for the Component 2 data collection. In comparisons to either BLS’s QCEW or the Census Bureau’s BDS, about one-fourth of establishments with more than 100 employees were missing from Component 2 data (Table 4-11). Combining across all establishments with more than 100 employees yielded a coverage ratio of 77.1 in 2017 and 78.0 in 2018. The coverage of establishments varied widely by establishment size. The largest establishments (with 1,000 or more employees) were overcovered relative to BLS data, with coverage ratios of 105.2 and 104.3 in 2017 and 2018, respectively, but undercovered relative to BDS data (coverage ratios of 84.2 in 2017 and 82.5 in 2018). This higher coverage (overcoverage relative to BLS benchmarks) of large establishments but undercoverage of large firms could indicate measurement error in the size of establishments (see Chapter 5) or that firms or filers did not report individual establishments separately in their filings. Establishments with 500–999 employees had coverage ratios above 80–90, depending on the data source, indicating that 10–20 percent of establishments of this size are missing in each year. Coverage ratios fall to around 75 for establishments with 100–499 employees.

Industry: Component 2 Data Versus External Benchmarks

Industry classifications of firms in the Component 2 data collection were compared with industry classifications in the Census Bureau’s BDS. It is well established that separate administrative data systems sometimes assign different NAICS codes to the same firm and/or establishment (Foster et al., 2021). EEOC filers and respondents to the Census Bureau’s surveys that create the Business Register may select different industries. EEOC also asks the respondents to self-assign industry classifications, which is different from using trained coders. Additionally, farms and agricultural businesses are not part of the target population for the economic census (Census Bureau, 2021f) and are thus largely excluded from the data collected by the Census Bureau but are included in Component 2 data (less than 600 firms with this NAICS code are included in the Component 2 data). EEOC collects Component 2 data from all private employers that meet certain eligibility criteria, including agricultural entities.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

At the firm level, the Business Register—and thus the estimates from BDS—count U.S. firms by industry in any applicable sector; thus, such firms can be double- or triple-counted (or more) in firm-level tabulations by industry. In contrast, firms and establishments in EEOC data are assigned to only one industry, as selected by the filer. For example, according to BDS, in 2017 the total number of firms with 100 or more employees was 113,556 (Table 4-10), but the total number of firms across industry-level tabulations was 158,212 (Table 4-12, summing over the column), a 39 percent increase. It is unknown how the multiple-industry classification varies across industries. Thus, estimates of undercoverage are expected to be overstated in firm-level industry comparisons, and caution is advised when interpreting coverage of firms at the industry level, compared to national benchmarks.

In general, coverage of firms varied across industries (Table 4-12). There was a higher coverage of firms with NAICS code 11 (agricultural firms), but all other industries were undercovered compared to BDS in both years. The highest coverage ratio was for NAICS code 22 (utilities), which was around 70 in both years. Industries with the lowest coverage ratios included those with NAICS codes 61 (educational services—although it is not clear if the Census Bureau’s eligibility criteria here match EEOC’s criteria), 55 (management of companies and enterprises), 42 (wholesale trade), 48–49 (transportation and warehousing), and 72 (accommodation and food services), all of which had coverage ratios below 30.

Coverage of establishments also varied across industries (Table 4-13). Similar to firms, establishments with NAICS code 11 (agricultural establishments) had a higher coverage ratio in Component 2 data than in Census Bureau data because the Census Bureau does not cover agricultural operations. Unlike at the firm level, establishments with NAICS codes 31–33 (manufacturing) were overcovered in Component 2 data compared to the Census Bureau benchmark data, with a coverage ratio of 124.9 in 2017 and 127.6 in 2018. Establishments with NAICS codes 21(mining, quarrying, and oil and gas extraction) and 23 (construction) had the next highest coverage ratios, at or above 84 in both years. The lowest coverage was for establishments with NAICS code 55 (management of companies and enterprises), with a coverage ratio less than 20 in both years.

SUMMARY

This chapter examined the pay data available in the Component 2 collections for reporting years 2017 and 2018. We found firm-level response rates were high, exceeding the court-mandated response rate. Response rates were high among all subgroups that the panel examined.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

When data collection resumed by court order, the response rate was slightly lower than the 82 percent target. A single extension brought the response rate above the target, to roughly 84 percent, and a second extension raised the rate to about 90 percent. The extensions increased the number of self-identified filers that were not on the master list to be surveyed.

However, data quality issues appeared when examining coverage (data quality is discussed further in Chapter 5). The lack of consistent, unique IDs interfered with the examination of birth and death rates among establishments. Furthermore, the number of employees associated with each establishment was often overstated, leading to small- and midsized establishments appearing to be large establishments. The bundling of multiple firms in a single report by PEOs also appeared to result in the false association of some employees with the reporting PEOs rather than the original firms. PEOs sometimes reported their own industry codes for those employers for which they filed, leading to errors in correct identification of the industries in which the establishments were active.

Ultimately, we found that the data covered less than 60 percent of U.S. private firms and establishments, partly due to nonresponse but primarily due to inadequacies in EEOC’s frame. Without weighting for nonresponse and poor coverage in the survey frame, the data are not appropriate for generating population statistics.

CONCLUSIONS AND RECOMMENDATIONS

Key conclusions and recommendations from the chapter’s findings appear below.

CONCLUSION 4-1: As collected, 2017–2018 Component 2 data are limited by significant data coverage issues related primarily to EEOC’s master list of potential respondents. In addition to respondent coverage, the panel identified issues with both nonresponse and measurement, which should be recognized when using these data.

CONCLUSION 4-2: The 2017–2018 Component 2 data have inconsistent and non-unique firm and establishment identifiers, which impede the maintenance of the master list, trend analysis over time, and data quality checking possible when merging by identifiers.

RECOMMENDATION 4-1: EEOC should improve the coverage of its master list, perhaps using an interagency agreement between EEOC and the Bureau of Labor Statistics to appropriately maintain business registers.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×

RECOMMENDATION 4-2: When preparing national and subnational statistics for the public, EEOC should adopt statistical weighting to adjust for possible undercoverage and nonresponse biases.

RECOMMENDATION 4-3: EEOC should use consistent and unique firm and establishment identifiers, facilitating data merges and data checking.

Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 107
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 108
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 109
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 110
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 111
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 112
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 113
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 114
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 115
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 116
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 117
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 118
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 119
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 120
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 121
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 122
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 123
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 124
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 125
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 126
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 127
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 128
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 129
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 130
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 131
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 132
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 133
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 134
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 135
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 136
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 137
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 138
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 139
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 140
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 141
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 142
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 143
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 144
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 145
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 146
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 147
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 148
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 149
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 150
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 151
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 152
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 153
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 154
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 155
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 156
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 157
Suggested Citation:"4 Do All Eligible Employers Receive and Respond to the Component 2 Instrument?." National Academies of Sciences, Engineering, and Medicine. 2023. Evaluation of Compensation Data Collected Through the EEO-1 Form. Washington, DC: The National Academies Press. doi: 10.17226/26581.
×
Page 158
Next: 5 Measurement Quality »
Evaluation of Compensation Data Collected Through the EEO-1 Form Get This Book
×
Buy Paperback | $50.00 Buy Ebook | $40.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The U.S. Equal Employment Opportunity Commission (EEOC) expanded EEO-1 data collection for reporting years 2017 to 2018 in an effort to improve its ability to investigate and address pay disparities between women and men and between different racial and ethnic groups. These pay disparities are well documented in national statistics. For example, the U.S. Census Bureau (2021) found that Black and Hispanic women earned only 63 percent and 55 percent as much, respectively, of what non-Hispanic White men earned.

Evaluation of Compensation Data Collected Through the EEO-1 Form examines the quality of pay data collected using the EEO-1 form and provides recommendations for future data collection efforts. The report finds that there is value in the expanded EEO-1 data, which are unique among federal surveys by providing employee pay, occupation, and demographic data at the employer level. Nonetheless, both short-term and longer-term improvements are recommended to address significant concerns in employer coverage, conceptual definitions, data measurement, and collection protocols. If implemented, these recommendations could improve the breadth and strength of EEOC data for addressing pay equity, potentially reduce employer burden, and better support employer self-assessment.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!