Affirmative Action and Nondiscrimination Obligations of Contractors and Subcontractors; Equal Opportunity Survey, 53032-53042 [E6-14922]

Agencies

• DEPARTMENT OF LABOR
• Office of Federal Contract Compliance Programs
• Federal Contract Compliance Programs Office

[Federal Register Volume 71, Number 174 (Friday, September 8, 2006)]
[Rules and Regulations]
[Pages 53032-53042]
From the Federal Register Online via the Government Printing Office [www.gpo.gov]
[FR Doc No: E6-14922]


=======================================================================
-----------------------------------------------------------------------

DEPARTMENT OF LABOR

Office of Federal Contract Compliance Programs

41 CFR Part 60-2

RIN 1215-AB53


Affirmative Action and Nondiscrimination Obligations of 
Contractors and Subcontractors; Equal Opportunity Survey

AGENCY: Office of Federal Contract Compliance Programs, Labor.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: The Office of Federal Contract Compliance Programs (OFCCP) is 
publishing a final rule rescinding the Equal Opportunity Survey (EO 
Survey) requirement in order to more effectively focus enforcement 
resources and eliminate a regulatory requirement that fails to provide 
value to either OFCCP enforcement or contractor compliance. This rule 
allows OFCCP to better direct its resources for the benefit of victims 
of discrimination, the government, contractors, and taxpayers.

DATES: Effective Date: September 8, 2006.

FOR FURTHER INFORMATION CONTACT: Director, Division of Policy, 
Planning, and Program Development, Office of Federal Contract 
Compliance Programs, 200 Constitution Avenue, NW., Room N3422, 
Washington, DC 20210. Telephone: (202) 693-0102 (voice) or (202) 693-
1337 (TTY).

SUPPLEMENTARY INFORMATION:

I. Background

    On January 20, 2006, OFCCP published a Notice of Proposed 
Rulemaking (NPRM), proposing to rescind a rule requiring designated 
nonconstruction contractors to prepare and file an EO Survey with 
OFCCP. 71 FR 3374. Created in 2000, the EO Survey was intended to 
further the goals of Executive Order 11246, as amended. The Executive 
Order requires that Federal Government contractors and subcontractors 
``take affirmative action to ensure that applicants are employed, and 
that employees are treated during employment, without regard to their 
race, color, religion, sex, or national origin.'' Section 202(1). 
Affirmative action under the Executive Order means more than passive 
nondiscrimination; it requires that contractors take affirmative steps 
to identify and eliminate impediments to equal employment opportunity. 
The affirmative steps include numerous recordkeeping obligations 
designed to assist the contractor, in the first instance, and also 
OFCCP in monitoring the contractor's employment practices.
    The EO Survey contains information about personnel activities, 
compensation and tenure data, and certain information about the 
contractor's affirmative action program. OFCCP recordkeeping rules 
require contractors to maintain information necessary to complete the 
EO Survey, although not in the format called for by the survey 
instrument. See 65 FR 26100

[[Page 53033]]

(May 4, 2000). The specific objectives of the EO Survey were:
    (1) To improve the deployment of scarce federal government 
resources toward contractors most likely to be out of compliance;
    (2) To increase agency efficiency by building on the tiered-review 
process already accomplished by OFCCP's regulatory reform efforts, 
thereby allowing better resource allocation; and
    (3) To increase compliance with equal opportunity requirements by 
improving contractor self-awareness and encourage self-evaluations.
    See 65 FR 68039 (Nov. 13, 2000); see also 65 FR 26101 (May 4, 
2000).
    OFCCP has carefully analyzed the extent to which the EO Survey has 
accomplished its stated objectives. This analysis included two studies 
that focused on the predictive ability of the EO Survey. The first 
study, the Bendick & Eagan Report,\1\ analyzed whether the pilot EO 
Survey results could be used to predict whether a contractor would have 
findings of non-compliance. The Bendick & Eagan Report did not 
demonstrate that the EO Survey is a good predictor of noncompliance \2\ 
because as the Report acknowledged, data problems and other 
methodological issues prevented Bendick & Eagan from conducting a full-
scale analysis of the pilot EO Survey's predictive power. Although the 
report stated that the EO Survey results might in the future be a way 
of finding contractors that are not in compliance, the report 
identified four ``handicaps'' that allowed it to present ``only a 
preliminary examination'' of the data's ability to differentiate 
between non-compliant and compliant establishments.\3\
---------------------------------------------------------------------------

    \1\ The Bendick & Eagan Report was produced by Bendick & Eagan 
Economic Consultants, Inc., and was entitled The Equal Opportunity 
Survey: Analysis of a First Wave of Survey Responses (September 
2000) (It was referred to in the NPRM as the Bendick Report, but is 
referred to here as the Bendick & Eagan Report to distinguish it 
from the comment submitted by Dr. Marc Bendick on March 2, 2006).
    \2\ The Executive Summary to the Bendick & Eagan Report 
concluded that the EO Survey ``can enhance the efficiency and 
effectiveness in OFCCP's monitoring of contractors' compliance with 
Executive Order 11246,'' but later acknowledges that its report 
provides ``only an exploratory, rather than a full-scale analysis of 
the Survey's predictive power.'' Bendick & Eagan Report at i-ii. The 
Bendick & Eagan Report did find ``preliminary positive indications 
of predictive power,'' which suggest that ``predictors based on the 
EO Survey are likely eventually to demonstrate substantial power.'' 
(Bendick & Eagan Report at 25) (emphasis added). The exploratory 
nature of its analysis, however, prevented a definitive finding on 
any correlation between predictive variables, generated from the EO 
Survey, and determinations of noncompliance.
    \3\ Bendick & Eagan Report at 18-27.
---------------------------------------------------------------------------

    The second study, the Abt Report,\4\ analyzed whether EO Survey 
data could be used to develop a model to more effectively target those 
contractors engaging in systemic discrimination. The following summary 
of the key findings of the Abt Report was presented in the NPRM (71 FR 
3374):
---------------------------------------------------------------------------

    \4\ The Abt Report was produced by Abt Associates of Cambridge, 
Massachusetts and was entitled An Evaluation of OFCCP's Equal 
Opportunity Survey.

    Abt found the model's predictive power to be only slightly 
better than chance. Screening on the basis of the model produced 
large numbers of false positives, that is, the model predicted 
numerous instances of systemic discrimination in the sample where 
OFCCP identified none. Specifically, using a cutoff for the 
probability that an establishment discriminates near the overall 
rate, the model suggests that 637 out of the 1,888 establishments in 
the study discriminate, yet only 42 (6.5%) of these are true 
positives. Thus, of 637 establishments that would be classified by 
the EO Survey results as suspected of having systemic 
discrimination, 93% would be false positives. Abt Report at 33. Even 
at a higher cutoff rate, where only 143 establishments are 
inspected, 127 were found to have no systemic discrimination, so the 
false positive rate remains high at 89% (i.e., 127/143).
    Furthermore, the EO Survey model wrongly classifies a 
significant portion of true discriminators as non-discriminators, 
and thus would not target them for compliance evaluations. If the 
637 establishments were chosen for review on the basis of the EO 
Survey model, 1,251 establishments would not have been reviewed. 
This group of 1,251 predicted by the EO Survey to lack 
discriminators would, in fact, have contained 21 of the 63 cases 
(33%) of systemic discrimination. Under the higher cutoff rate, 
about 75% of the establishments (47 contractors) that were found to 
have systemic discrimination would not have been reviewed under the 
EO Survey model.\5\
---------------------------------------------------------------------------

    \5\ Abt Report at 33-35. See also NPRM at 71 FR 3375-76.

    Based on the results of the studies, and the evaluation of new 
initiatives implemented by OFCCP to accomplish the same objectives of 
the EO Survey but in different ways,\6\ OFCCP concluded that the EO 
Survey failed to meet its objectives, and proposed removing the EO 
Survey requirement from covered contractors' obligations under the 
Executive Order. The preamble to the proposed rule discusses in depth 
the results of the studies and the reasons for OFCCP's proposal to 
rescind the EO Survey. 71 FR 3374-78.
---------------------------------------------------------------------------

    \6\ For an explanation of these initiatives, see the discussion 
in Section C below.
---------------------------------------------------------------------------

    OFCCP received a total of 2,736 comments on the NPRM. Of those, 
1,707 comments (62%) supported the proposal to discontinue the EO 
Survey and 1,029 comments (38%) opposed the proposed rule. Most of the 
comments focused on (1) The Abt Report; (2) the alleged intrinsic value 
of the EO Survey; and/or (3) the implications of rescinding the EO 
Survey.
    After considered review of the comments, OFCCP concludes that the 
objectives of the Executive Order 11246 program can be better 
accomplished through means other than the EO Survey, and publishes this 
final rule to rescind the EO Survey filing requirement. There are no 
differences between the proposed rule and the final rule.

II. Discussion of the Comments

A. Comments on the Abt Report

    Many of the commenters who support the proposal to rescind the EO 
Survey cited the Abt Report and the conclusions that OFCCP drew from 
it. For example, the Silicon Valley Industry Liaison Group stated:

    [I]t is clear to our member companies that the EO Survey has no 
internal value to the company * * *. Abt Associates, indicated that 
the EO Survey does not accomplish what it was constructed to do: 
find systemic discrimination. * * * In the Jan. 20, 2006 Proposed 
Rule, OFCCP states ``that the EO Survey misdirects valuable 
enforcement resources and does not meet any of its three objectives 
set out in the November 13, 2000 preamble.'' Since the EO Survey 
lacks efficacy and has no internal value to the contractor, we 
applaud the Agency for its recommendation to withdraw its use.\7\
---------------------------------------------------------------------------

    \7\ Silicon Valley Industry Liaison Group (SVILG) March 17, 2006 
letter. The SVILG comprises one of the largest liaison groups in the 
country with 272 members, including many leading high-tech, bio-tech 
and other major employers in Northern California.

    Likewise, the U.S. Chamber of Commerce stated, ``The Abt study--an 
impartial, comprehensive and statistically sound assessment of the 
value of the Survey--provides a sound regulatory basis for OFCCP to 
eliminate the Survey and search for new ways to select establishments 
for audit.'' \8\ Noting the Abt Report's findings concerning the false 
positive and false negative rate generated by the EO Survey data, the 
National Association of Manufacturers commented:
---------------------------------------------------------------------------

    \8\ Crowell & Moring LLP March 28, 2006 letter at 4-5 
(representing the U.S. Chamber of Commerce).

    Simply stated, any system that targets compliant contractors for 
audit, thus punishing those employers striving to comply with their 
affirmative action and non-discrimination obligations, while 
allowing non-compliant contractors to avoid detection, utterly fails 
to serve any legitimate regulatory or enforcement purpose and should 
be eliminated. Indeed, continuing a system that consciously targets 
a significant

[[Page 53034]]

number of compliant contractors violates fundamental principles of 
due process.\9\
---------------------------------------------------------------------------

    \9\ Fortney & Scott, LLC March 27, 2006 letter at 4 
(representing the National Association of Manufacturers).

    Conversely, many commenters who support retention of the EO Survey 
suggest that the Abt study is flawed, and thus no valid inferences 
regarding the EO Survey's predictive power can be drawn from the Abt 
study.\10\ For example, the Florida Federation of Business and 
Professional Women's Clubs, Inc. stated:
---------------------------------------------------------------------------

    \10\ Leadership Conference on Civil Rights March 20, 2006 letter 
at 2.

    The proposal to eliminate the EO Survey cites the findings from 
a research consultant. However, the consultant's analysis was based 
upon a skewed sample because contractors who did not respond or 
provided questionable information were not included.\11\
---------------------------------------------------------------------------

    \11\ Florida Federation of Business and Professional Women's 
Clubs, Inc. March 21, 2006 letter.

    The National Women's Law Center noted:
    OFCCP attempts to justify its proposal with findings from the 
study it commissioned by Abt Associates. Essentially, OFCCP 
concludes that the Survey's predictive power is little better than 
chance, and produces so many false positives and false negatives as 
to be virtually useless in targeting those contractors that have 
engaged in systemic discrimination. However, neither these nor any 
other conclusions about the EO Survey's predictive power can be 
validly drawn from the Abt study, because the study sample given to 
Abt by OFCCP, and on which these conclusions are based, was 
hopelessly skewed and unrepresentative of the contractor 
community.\12\
---------------------------------------------------------------------------

    \12\ National Women's Law Center March 28, 2006 letter at 3-4.

    Given the significance of the Abt study, the commenters' major 
critiques of the study are addressed below. For presentation purposes, 
these critiques have been grouped into three areas:
    1. The Abt study should have been based upon a larger group of 
federal contractors.
    2. The sample used by Abt was skewed.
    3. The Abt study inappropriately focused on systemic 
discrimination, rather than all violations.
1. The Abt Study Should Have Been Based Upon a Larger Group of Federal 
Contractors
    Some of the comments in opposition to the proposal maintain that 
the Abt study is flawed because it was not based upon a larger group of 
federal contractors. Other commenters focused on the decline in the 
number of EO Surveys OFCCP distributed each year. For example, the 
Leadership Conference on Civil Rights stated:

    The EO Survey's distribution was dramatically reduced--from 
50,000 contractors to 10,000--thus undermining the reach of the 
instrument and raising questions about OFCCP's commitment to carry 
out the intent of the law. Further, to our knowledge, the data 
collected through the EO Survey has never been used by OFCCP for 
targeting of compliance reviews.\13\
---------------------------------------------------------------------------

    \13\ Leadership Conference on Civil Rights March 20, 2006 letter 
at 2.

    In contrast, Maly Consulting LLC suggested that OFCCP should not 
have sent out any EO Surveys before OFCCP did ``a complete job to 
determine its viability.'' \14\
---------------------------------------------------------------------------

    \14\ Maly Consulting LLC March 27, 2006 letter at 2.
---------------------------------------------------------------------------

    OFCCP acknowledges that the number of EO Surveys sent out declined. 
In fact, the NPRM specifically notes that ``OFCCP mailed 53,000 EO 
Surveys between December 2000 and March 2001, 10,000 in December 2002, 
10,000 in December 2003, and 10,000 in December 2004.'' (71 FR 3375) 
The reason for this decline was noted in the January 2003, OFCCP notice 
in the Federal Register seeking a two-year extension of the Paperwork 
Reduction Act clearance (68 FR 4797) and the NPRM to this final rule. 
That is:

    Time constraints and a number of data problems affected an 
earlier pilot study of the EO Survey data [the Bendick & Eagan 
Report] in such a way so as not to be able to assess the Survey's 
predictive power. To perform a study that is not limited by these 
obstacles, OFCCP has engaged an outside contractor to study the 
Survey data. The contractor will assess data from the EO Survey 
submissions as part of its study. * * * OFCCP requests a two-year 
extension of PRA authorization for the EO Survey, involving 10,000 
EO Surveys per year. The two-year extension will permit OFCCP to 
complete the ongoing study of the EO Survey. Ten-thousand Surveys is 
the number the outside contractor needs to assess the Survey's 
reliability for finding employers that discriminate against their 
employees.\15\
---------------------------------------------------------------------------

    \15\ 68 FR 4797, 4798 (2003). See also NPRM at 71 FR 3375.

    Without a complete validation study of the utility of the EO 
Survey, it would not have been useful to send EO Surveys to the broader 
contractor community. Indeed, it was logical and consistent with the 
Paperwork Reduction Act to send only a sufficient number of EO Surveys 
to develop the predictive model and to fully test and validate the EO 
Survey.
    Regarding the Abt study, the limiting factor was not the number of 
EO Surveys sent out but rather the number of compliance evaluations 
that could be completed. As the Bendick & Eagan Report noted, one of 
the Bendick & Eagan Report's methodological shortcomings was its 
inability to compare compliance evaluations with EO Survey results.\16\ 
Undertaking such a comparison was one of the essential goals of the Abt 
study. Regardless of the number of EO Surveys, OFCCP expected to be 
able to conduct only 2,250 compliance reviews for the study. Thus, it 
was expected that only about 2,250 EO Surveys could be linked to 
completed compliance evaluations. This linkage is crucial to the study 
because without it there is no possibility of modeling the data on the 
EO Survey to a systemic discrimination outcome.
---------------------------------------------------------------------------

    \16\ Bendick & Eagan Report at 20-21.
---------------------------------------------------------------------------

    Based on the 2,250 estimate, Abt determined that about 10,000 EO 
Surveys would have to be sent out. (This is the number that was sent 
out in December 2002, 2003 and 2004.) As detailed in Chapter 2 of the 
Abt Report, the selection of the establishments was done in the 
following manner:

    The target population consisted of a subset of the 95,961 
establishments with EEO-1 contractor records for FY2000. The subset 
excluded the following categories:
     Establishments that were sent EO Surveys the previous 
year.\17\
---------------------------------------------------------------------------

    \17\ In addition to minimizing the burden on a single 
contractor, this avoided the problem cited in the Bendick & Eagan 
Report of contaminating the EO Survey data by conducting compliance 
evaluations prior to collection of EO Survey responses.
---------------------------------------------------------------------------

     Establishments that the OFCCP reviewed within the last 
two years (FY2001 and FY2002).
     Establishments associated with a parent company for 
which the OFCCP has approved a Functional Affirmative Action 
Program.
     Any establishment that had the same parent company as 
an establishment that had asserted that the OFCCP lacked 
jurisdiction (for reasons that comprised five categories).
     A small number of establishments that had very 
questionable records.
     Establishments that were among the 6,863 to which EO 
Surveys were sent in April 2000, in connection with the pilot study.
     All establishments of two large companies that have 
traditionally contested jurisdiction and were not sent EO Surveys on 
the previous round.
    The resulting subset contained 26,451 establishments. A sample 
of approximately 10,000 establishments was drawn from this sampling 
frame, according to an allocation among a detailed set of 
strata.\18\
---------------------------------------------------------------------------

    \18\ Abt Report at 3.

    The strata were based upon three factors: region, industry and 
establishment size. The details of the strata are presented on page 4 
of the Abt Report. Page 5 of the Abt Report presents the number of 
---------------------------------------------------------------------------
establishments in each stratum:

    Because of the random rounding in the allocation procedure, the 
actual total sample

[[Page 53035]]

size was 10,018 establishments. The actual sample was obtained by 
selecting a simple random sample of establishments from each of the 
276 final strata. * * * The subsample [for review] was selected in 
three parts, an initial sample of 3,300 and two supplementary 
samples (of 1,000 and 2,100, respectively), as experience with the 
reviews led to revisions in the initial assumptions. Thus, the total 
size of the subsample was 6,400.

    The 6,400 random review subsample was reported in footnote 2 on 
page 3375 of the NPRM. As was also reported in that note:

    Of these 6,400, only 3,723 establishments responded to the EO 
Survey. Of these 3,723, only 2,651 had data that allowed OFCCP to 
complete a compliance evaluation. Thus, OFCCP completed about 2,651 
compliance evaluations. However, of the 2,651, a significant number 
(763) had missing or incoherent data on the EO Survey, and were not 
used in the study. Ultimately the study focused on 1,888 cases that 
had completed compliance reviews and had reliable EO Survey data.

    The number of completed evaluations on contractors that returned 
the EO Survey (2,651) actually exceeded OFCCP's original goal of 
completing 2,250 evaluations for the study by almost 18%. Moreover, the 
3,618 establishments that were not ``used'' by Abt \19\ could not have 
an impact on the results of their analysis because the original 10,018 
establishments (both the 6,400 review subsample and the 3,618 non-
review subsample) were drawn in a random fashion.
---------------------------------------------------------------------------

    \19\ Abt Report, Appendix E, Table A.
---------------------------------------------------------------------------

    If EO Surveys had been sent out to all establishments with EEO-1 
contractor records for FY2000, OFCCP still would have only been able to 
complete about 2,651 compliance evaluations. Thus, it is unlikely that 
sending the EO Survey to more contractors would have altered the 
results of the study. On the contrary, the approach of sending out the 
minimum number of EO Surveys necessary to conduct a statistically valid 
study not only reduced the burden on federal contractors but also 
minimized the burden on OFCCP and its resources. The selection strategy 
utilized by Abt produced a representative sample of federal contractors 
while avoiding the contamination issues mentioned in the Bendick & 
Eagan Report. In sum, an adequate number of establishments were sent 
the EO Survey.
2. The Sample Used by Abt Was Skewed
    The second major criticism of the Abt Report concerned whether the 
sample it used was representative. Despite the efforts by Abt to 
produce a representative sample of Federal contractors for the study, 
several commenters opposing the proposal maintain that the Abt study 
was flawed because it did not use the data from all of the contractors 
who were sent the EO Survey. For example, the National Women's Law 
Center stated:

    The integrity of OFCCP's sample was compromised from the 
beginning. Any contractor that refused to respond to the EO Survey 
(10%), asserted that OFCCP lacked jurisdiction (27%), or went out of 
business (5%) was simply dropped from the sample. * * * Another 15% 
of the contractors were dropped from the sample because they had 
submitted responses to the Survey that contained internal 
inconsistencies too extreme to address with ``suitable cleaning.'' 
As a result, more than half of the original sample of 10,000 
contractors was dropped before the study even began and before Abt 
built its model of predictive power. Ultimately, the study sample 
was whittled down to 1,888 contractors for whom Abt had both a 
Survey containing adequate data and the results from a CR conducted 
by OFCCP.\20\
---------------------------------------------------------------------------

    \20\ National Women's Law Center March 20, 2006 letter at 4.

---------------------------------------------------------------------------
    Similarly, the Bendick Comment stated:

    [T]his OFCCP conclusion is not justified by the Abt Report 
because the sample of employers OFCCP provided to Abt was not 
appropriate for the study of the Survey's predictive power. The 
sample consisted of 2,226 firms for which both a compliance audit 
and a Survey response was available. If the employer refused to 
answer the EO Survey or provided only apparently-incorrect data, 
then that firm was simply dropped from the sample. Firms which were 
not included in the sample totaled 3,352 of 6,400 firms which could 
have been included in the study. That is, 52.4%--more than half--of 
firms were omitted from the data before Abt began its analysis.\21\
---------------------------------------------------------------------------

    \21\ Dr. Marc Bendick March 2, 2006 comment at 3 (footnote 
omitted) (emphasis in original) (hereinafter ``Bendick Comment''). 
The Bendick Comment also asserted that the 2000 Bendick & Eagan 
Report ``found exactly the reverse of what the [NPRM] says it 
found,'' pointing specifically to the NPRM's statement that the 
Bendick & Eagan Report ``failed to find a correlation between the 
predictive variables, generated by the EO Survey, and determinations 
of noncompliance.'' Id. at 2 (citing 71 FR 3374). Despite the 2006 
Bendick Comment, the 2000 Bendick & Eagan Report specifically 
stated: ``The EO Survey data collected in the April 2000 wave does 
not offer circumstances in which the full predictive power of the 
survey can be revealed. Four handicaps are important to note. * * * 
Considering these four circumstances, this report presents only a 
preliminary examination of the ability of selected variables drawn 
from the EO survey to differentiate establishments likely to have 
non-compliance findings from those not likely to have such 
outcomes.'' Bendick & Eagan Report at 20-23. In the EO Survey NPRM, 
OFCCP acknowledged that ``data problems prevented Bendick from 
conducting a full-scale analysis of the pilot EO Survey's predictive 
power. The report stated that the EO Survey results might in the 
future be a way of finding contractors that discriminate, but the 
pilot EO Survey did not.'' 71 FR 3374-75 (citing Bendick & Eagan 
Report at 18-27).

    Numerous commenters, including American Federation of Government 
---------------------------------------------------------------------------
Employees, Local 12, AFL-CIO, echoed the following sentiment:

    The proposal to eliminate the EO Survey cites the findings from 
a research consultant. However, the consultant's analysis was based 
on a skewed sample because contractors who did not respond or 
provided questionable information were not included. Earlier 
research by a different consultant concluded that the very 
contractors who did not comply with the EO Survey in the first place 
were more likely to be in violation of the law.\22\
---------------------------------------------------------------------------

    \22\ American Federation of Government Employees, Local 12, AFL-
CIO March 17, 2006 letter at 1.

    To address these concerns about the Abt sample, it is necessary, as 
a preliminary matter, to examine the composition of the 6,400 
establishments that were sent the EO Survey and in the review subsample 
used by Abt. Table B presented in Appendix E of the Abt Report provides 
a breakdown of the 6,400 establishments in the review subsample 
selected by Abt.
    Of the 6,400 contractors sent EO Surveys and in the subsample used 
by Abt, 2,004 were either out of business or asserted that they did not 
have to respond (e.g., they were not federal contractors with at least 
50 employees). These establishments were excluded from the analysis 
because it would have been difficult and an inefficient use of 
resources to include them in the model. It would have been nonsensical, 
if not impossible, for OFCCP to complete compliance evaluations on the 
330 establishments who were out of business. Further, including the 
small number of establishments that claimed they didn't have to respond 
to the EO Survey, but should have in the Abt study, could not have 
significantly skewed the results of the analysis given they were also 
randomly selected.
    Of the remaining 4,396 contractors, 3,723 (about 85%) responded to 
the EO Survey with data that either passed the initial OFCCP check with 
an ``OK'' status or submitted data that generated an ``edit condition 
report.'' However, OFCCP had not completed compliance evaluations on 
all of these contractors. As stated in the NPRM, OFCCP completed 
compliance evaluations on only 2,651 of the contractors that responded 
to the EO Survey with data (about 71% of 3,723). This represented the 
pool of available matches of EO Survey data and systemic discrimination 
determinations.
    As was discussed in the NRPM, after further evaluating the data, 
Abt focused on the set of 1,888 cases that had completed compliance 
reviews and

[[Page 53036]]

what Abt considered reliable EO Survey data. The results of Abt's 
analysis of these cases were presented in the NPRM. See 71 FR 3375 n. 2 
and Abt Report, Appendix E, Table B.
    Before the report was finalized, OFCCP asked Abt to analyze the 
data with ``relaxed'' edits due to this very concern that the cases 
being omitted from the analysis would bias the results. Appendix E 
presents Abt's findings with the relaxed edits and, ``The result, in 
brief, was that [Abt] emerged with the same four predictor variables. 
The coefficients were somewhat different, but not greatly so. The 
qualitative interpretation is pretty much the same.'' \23\
---------------------------------------------------------------------------

    \23\ Abt Report, Appendix E, at 1-2.
---------------------------------------------------------------------------

    Based upon this analysis, OFCCP concluded that Abt's data quality 
standards did not have a significant impact on the results of the 
study. In short, OFCCP concluded that excluding those establishments 
from the sample which Abt ultimately analyzed would not have changed 
Abt's conclusion regarding the predictive power of the EO Survey.
    There remains a group of 673 non-respondents out of the subsample 
of 6,400, or 10.5%. The supposition by many commenters is that this 
omitted group contains a high portion of noncompliant contractors. Such 
speculation cannot be verified. In fact, there could be any number of 
reasonable explanations for the number of non-respondents. For example, 
contractors may have been unable to properly complete the EO Survey or 
simply may not have returned it to OFCCP.\24\ Moreover, one could just 
as easily speculate that the non-respondents are not under the 
jurisdiction of OFCCP and chose to ignore the EO Survey. Whatever the 
reason, because the review subsample was randomly drawn, the relatively 
low non-response rate is unlikely to have a statistically significant 
impact on the results of the Abt Report.
---------------------------------------------------------------------------

    \24\ In a related comment, the U.S. Chamber of Commerce 
observed: ``Many Survey responses had to be disregarded due to 
clearly erroneous data, demonstrating the difficulties that 
employers had in providing accurate information.'' Crowell & Moring 
LLP March 28, 2006 letter at 4 (representing the U.S. Chamber of 
Commerce).
---------------------------------------------------------------------------

    Finally, some commenters who argue for retaining the EO Survey cite 
the difference in the results of the Bendick & Eagan and Abt reports as 
evidence that the Abt Report is flawed. For example, the Bendick 
Comment stated:

    In the sample studied by Abt, only 3.0% of firms were found out 
of compliance (engaged in systemic discrimination). In the sample 
analyzed in the Bendick Report, 38.4% of the firms surveyed were 
found out of compliance. Thus, the data set Abt analyzed was clearly 
not representative of all federal contractors.\25\
---------------------------------------------------------------------------

    \25\ Bendick Comment at 3 (footnote omitted).

    The reason for this difference is not because the Abt Report is 
flawed or skewed, but because the Abt Report appropriately focused on 
systemic discrimination, which is the focus of OFCCP's enforcement 
strategy, while the Bendick & Eagan Report studied non-compliance in 
its broadest sense, of which systemic discrimination is only one part. 
Directly comparing the results of the two studies is not really 
appropriate and can be misleading.\26\ Since systemic discrimination 
violations are a subset of the types of non-compliance that OFCCP finds 
in its reviews, and the most harmful to workers, it is not at all 
surprising that the rate of systemic discrimination in the sample used 
by Abt is lower than the rate of non-compliance in the sample used by 
Bendick & Eagan, which included both a wide variety of paperwork 
violations and systemic discrimination violations.
---------------------------------------------------------------------------

    \26\ The National Women's Law Center acknowledges that a 
comparison of the findings of the Bendick & Eagan Report and Abt 
Report may not be appropriate, but submits that it should have led 
OFCCP to question the Abt sample: ``This comparison of noncompliance 
rates may not be an apples-to-apples comparison because of the 
narrow scope of violations OFCCP used in framing its study and in 
conducting [compliance reviews] * * *. Still, the dramatic 
difference in rates of noncompliance found through OFCCP's 
[compliance reviews] should have led OFCCP, at a minimum, to 
question the representativeness of the sample it was using.'' 
National Women's Law Center March 28, 2006 Letter at 5, n. 22. It 
should be noted that OFCCP did review the sample and methodology 
used by Abt and determined it to be statistically valid.
---------------------------------------------------------------------------

    In short, the sample Abt used was appropriate, statistically valid, 
and did not skew the results.
3. The Abt Report Inappropriately Focused on Systemic Discrimination, 
Rather Than All Violations
    The third major criticism of the Abt Report was its focus on 
systemic discrimination. Several commenters who support retaining the 
EO Survey assert that the Abt Report inappropriately focused on 
systemic discrimination, rather than all violations. They believe that 
by focusing only on systemic discrimination, the study underestimated 
the true benefit of the EO Survey. A typical example of this comment is 
that from Schaeffer and Schaeffer LLC:

    OFCCP expressed its intent during the formal rulemaking in 2000 
when the agency said that the data in all three parts of the EO 
Survey were intended ``to provide indicators of potential compliance 
problems for which further inquiry may be appropriate.'' OFCCP also 
stated ``The survey responses do not prove that a problem exists, 
but rather are used as an indicator to guide OFCCP compliance 
evaluations.'' * * * While OFCCP's emphasis on systemic compensation 
discrimination is a very positive development in many respects for 
which the agency should be commended, the question remains whether 
it is the proper standard for the EO Survey to meet.\27\
---------------------------------------------------------------------------

    \27\ Schaeffer and Schaeffer LLC March 28, 2006 letter at 4-5 
(emphasis in original).

    The National Women's Law Center emphasized, ``Systemic 
discrimination may be OFCCP's enforcement focus, but it is not the sum 
total of OFCCP's legal mandate nor the EO Survey's only purpose. This 
cordoning off of the Survey's scope itself may bias the Abt study's 
findings.'' \28\
---------------------------------------------------------------------------

    \28\ National Women's Law Center March 28, 2006 letter at 4, n. 
14.
---------------------------------------------------------------------------

    Systemic discrimination is indeed the proper standard for the EO 
Survey to measure. OFCCP's mission is based on the underlying principle 
that employment opportunities generated by Federal dollars should be 
available to all Americans on an equitable and fair basis. To fulfill 
this mission, it is OFCCP's stated policy to focus on increasing 
outreach efforts and targeting systemic discrimination in order to make 
better use of its resources. This policy has proven to be very 
effective. For example, in September 2004, OFCCP secured $5.5 million 
in salary adjustments and other financial remedies for 2,021 current 
and former female employees of a major financial institution who had 
been subjected to illegal compensation discrimination. This was OFCCP's 
fourth largest case in terms of monetary recovery, and was the first 
systemic compensation discrimination case to be filed in a quarter 
century. In FY 2005, OFCCP recovered a record $45.2 million for 14,761 
American workers who had been subjected to unlawful employment 
discrimination--a 56 percent increase over recoveries in FY 2001.
    Central to this policy is scheduling and focusing OFCCP's 
compliance evaluations on those cases most likely to result in findings 
of systemic discrimination and the recovery of make whole relief for 
victims of discrimination. It has long been widely recognized that 
compliance evaluations consume significant resources, that OFCCP can 
only conduct evaluations on a portion of all federal contractors, and 
that a large portion of the evaluations conducted do not result in 
findings of systemic discrimination.\29\ Therefore, it

[[Page 53037]]

is crucial to OFCCP's policy that the evaluations that are conducted be 
better targeted. Since OFCCP is focusing its compliance evaluations on 
systemic discrimination and, as noted by Schaeffer and Schaeffer, the 
stated purpose of the EO Survey was to provide an indication when 
further inquiry may be appropriate,\30\ it was appropriate for the Abt 
Report to focus on cases of systemic discrimination rather than 
generally on all types of non-compliance (including, largely, 
affirmative action program paperwork requirements).
---------------------------------------------------------------------------

    \29\ See, e.g., Bendick Comment at 5.
    \30\ Schaeffer and Schaeffer LLC March 28, 2006 letter at 4-5.
---------------------------------------------------------------------------

    Some commenters also cite the Bendick & Eagan Report to show that 
the EO Survey has value. For example, National Employment Lawyers 
Association stated:

    The Bendick study found a correlation between the predictive 
variables generated by the EO Survey and determinations of non-
compliance. That report examined 31 predictive variables and found 
28 of them (90.4%) to have some predictive power, including 11 
(35.5%) in which the predictive power was ``statistically 
significant.'' \31\
---------------------------------------------------------------------------

    \31\ National Employment Lawyers Association March 20, 2006 
letter at 2.

    Aside from the data issues discussed on pages 20 to 23 of the 
Bendick & Eagan Report, OFCCP has determined that the report's use of 
the broad term ``non-compliance'' instead of systemic discrimination 
inflates the predictive power of the variables. Since it was never 
OFCCP's intention to issue violations solely based upon the EO Survey, 
OFCCP is required to follow-up the EO Survey results with a compliance 
evaluation to actually make a finding of ``non-compliance.'' The 
correlation of the broad definition of non-compliance used in the 
Bendick & Eagan Report with the predictor values in the EO Survey would 
do little to advance OFCCP's goal of targeting systemic discrimination 
and recovering make whole relief for those who suffered from 
discrimination. On the contrary, by including other violations in the 
definition of non-compliance, this approach would divert resources from 
investigating the potential cases of systemic discrimination toward 
cases involving just paperwork violations. The Bendick Comment 
acknowledges that ``OFCCP resources permit only a very small proportion 
of federal contractors to be reviewed each year--at the time the 
Bendick Report was completed, less than 4 percent of contractors each 
year.'' \32\ Thus, it is critical to OFCCP's enforcement strategy that 
these resources be used efficiently to protect workers actually harmed 
by discrimination, remedy that discrimination, and bring violators into 
compliance.
---------------------------------------------------------------------------

    \32\ Bendick Comment at 5.
---------------------------------------------------------------------------

4. Conclusion
    After careful consideration of these comments, OFCCP continues to 
believe that the Abt Report is statistically sound and supports its 
conclusion that the EO Survey data does not, in any meaningful way, 
improve OFCCP's ability to target for review those contractors engaging 
in systemic discrimination.

B. Comments on the Alleged Intrinsic Value of the EO Survey

    The second major area discussed by commenters is the alleged 
intrinsic value of the EO Survey. This view, as articulated by the 
Leadership Conference on Civil Rights is that ``Even if the data 
collected [on the EO Survey] does not automatically prove 
discrimination, it provides a picture of a contractor's workforce that 
otherwise would not be available. It is the potential for this 
increased level of scrutiny that provides the incentive for contractor 
self-examination.'' \33\ By contrast, the National Association of 
Manufacturers ``heartily endorses elimination of the EO Survey as an 
overly burdensome, expensive, and wholly ineffective regulatory 
requirement that unnecessarily duplicates other equal employment 
opportunity (``EEO'') and affirmative action reporting obligations.'' 
\34\
---------------------------------------------------------------------------

    \33\ Leadership Conference on Civil Rights March 20, 2006 letter 
at 3.
    \34\ Fortney & Scott March 27, 2006 letter at 2 (representing 
the National Association of Manufacturers).
---------------------------------------------------------------------------

    The main points raised by supporters of the EO Survey about its 
alleged intrinsic value are:
    1. The EO Survey is the only reliable method to collect 
compensation data.
    2. The EO Survey enhances the tiered review process.
    3. The EO Survey facilitates effective self-evaluations by federal 
contractors.
1. The EO Survey Is the Only Reliable Method to Collect Compensation 
Data
    The concern that the EO Survey is the only reliable method to 
collect compensation data was expressed by numerous commenters, 
including the National Employment Lawyers Association, which stated:

    The Notice indicates that if the EO Survey is discontinued, 
OFCCP will use the EEO-1 data to predict the likelihood of whether a 
contractor will be found out of compliance. Although EEO-1 counts 
are useful, the data from the EO Survey are even more useful. * * * 
The EO Survey also contains compensation data that EEO-1 counts do 
not provide. Eliminating the EO Survey would jettison an extremely 
useful tool for identifying discrimination.\35\
---------------------------------------------------------------------------

    \35\ National Employment Lawyers Association March 20, 2006 
letter at 3.

    The Unitarian Universalist Association of Congregations suggested 
that the compensation data on the EO Survey is useful to OFCCP for 
---------------------------------------------------------------------------
targeting purposes:

    The EO Survey is a particularly important tool because it, for 
the first time, would provide OFCCP with pay data from all federal 
contractors every two years. That information could be used by OFCCP 
to help identify unequal pay practices, and better target its 
limited enforcement resources.\36\
---------------------------------------------------------------------------

    \36\ Unitarian Universalist Association of Congregations March 
20, 2006 letter.

    While the EO Survey collects data on compensation by EEO-1 
category, the Abt Report indicates that the data have no relation to 
the determination of systemic discrimination and contrary to these 
assertions is not a useful tool for enforcement purposes. The 
proponents of the EO Survey apparently believe that the mere collection 
of this data will have some beneficial effect. However, there is no 
evidence that the specific compensation data collected by the EO Survey 
can be used to predict compensation discrimination. Rather, the data is 
collected in such a raw and aggregate form that it cannot be used to 
compare similarly situated employees, and thus has negligible value in 
predicting compensation discrimination. The U.S. Chamber of Commerce 
agreed with OFCCP's assessment of the predictive value of the 
---------------------------------------------------------------------------
compensation data collected by the EO Survey:

    [T]he compensation data required by the Survey, submitted on an 
EEO-1 category basis, fails to provide any information useful to 
OFCCP in identifying contractors appropriate for audit. Because the 
data is reported on a broad EEO-1 category basis, the OFCCP cannot 
use the data to assess the compensation of similarly-situated 
employees. The data likewise cannot be subjected to a valid 
statistical analysis, and the Survey ignores the myriad non-
discriminatory factors that may impact compensation. Indeed, any 
methodology that could be employed with respect to compensation data 
generated by the Survey would be wholly at odds with the draft 
guidance issued by OFCCP in November 2004 regarding systemic 
analyses of compensation.\37\
---------------------------------------------------------------------------

    \37\ Crowell & Moring LLP March 28, 2006 letter at 3 
(representing the U.S. Chamber of Commerce).

    Even if there were some small marginal utility to EO Survey 
compensation data, the minimal benefit of the data would be outweighed 
by the

[[Page 53038]]

burden on the contractor to complete the EO Survey, and on OFCCP to 
process and use the EO Survey. Moreover, the obligation to expend 
resources to complete the EO Survey could discourage contractors from 
conducting a more thorough and useful evaluation of their personnel 
data. The necessity to collect and process EO Survey data could divert 
scarce OFCCP resources from more vigorously enforcing equal employment 
laws in a more effective manner.
    OFCCP believes that remedying compensation discrimination is 
important to its mission. But the EO Survey fails as a means of 
targeting it. As previously discussed, the Abt Report demonstrated that 
using the EO Survey for targeting would direct compliance officers away 
from contractors who are discriminating. In addition, the EO Survey 
would direct them--93% of the time--to contractors who are not 
discriminating.
    Further, the EO Survey is not the only source of compensation data 
available to OFCCP. First, OFCCP collects compensation data pursuant to 
Item 11 of the Scheduling Letter sent out to contractors selected for a 
compliance evaluation. The compensation data collected at initial desk 
audit stage is vastly superior to EO Survey compensation data. The data 
collected at the desk audit is more refined than the EO Survey data and 
is also specifically tailored to the contractor's job groups. In 
contrast, the EO Survey data is collected by EEO-1 category, which are 
likely too aggregate and result in the grouping of dissimilar jobs. As 
demonstrated by the Abt Report, studying the differences in pay 
averages for aggregate-level employee groups, which is the only type of 
compensation analysis the EO Survey data permits, is not even 
predictive of compensation discrimination. Finally, the desk audit data 
is likely to be more current and accurate, due to the interaction 
between the compliance officer and the contractor. In contrast to the 
computer program-based EO Survey, during a desk audit, a compliance 
officer reviews the compensation data, and can inquire about issues 
with the data, thus providing the contractor with the opportunity to 
correct any erroneous data submissions.
    In addition to the compensation data produced at the desk audit, 
other tools are available for pay assessments. Each Federal contractor 
is required by regulation to conduct a compensation self-analysis as 
part of its mandated affirmative action plan. See 41 CFR 60-2.17(b)(3). 
Certain covered contractors are required, pursuant to 41 CFR 60-2.1 to 
create and annually update an Affirmative Action Program evaluating the 
impact of all of their employment practices, including compensation, on 
women and minorities and to correct any problems identified.
    In sum, the EO Survey is not reliable and it is not the only means 
available for collecting such data. OFCCP collects compensation data as 
part of the desk audit process, and contractors are required to collect 
such data as part of its affirmative action obligations. The Paperwork 
Reduction Act specifically requires that the data collected have 
utility. 44 U.S.C. 3506(c)(3)(A). It does not appear that the EO Survey 
meets this threshold. It is unnecessary to maintain the EO Survey to 
collect compensation data, as other tools accomplish the same purpose, 
with better results for the agency.
2. The EO Survey Enhances the Tiered Review Process
    Some commenters assert that the EO Survey enhances OFCCP's tiered 
review process. For example, the AFL-CIO stated:

    The EO Survey enhances the effectiveness of the tiered-review 
system by enabling OFCCP to more accurately determine which level or 
type of compliance review is appropriate for a particular 
contractor. * * * [T]he tiered-review program is designed to ensure 
that the agency bases its level of review of a contractor on the 
likelihood of uncovering substantive violations, as determined at 
the early stages of review. Thus, is it [sic] essential that those 
early-stage targeting determinations are as accurate as possible, 
and the initial data collected by the EO Survey helps ensure that 
accuracy by providing essential information about each contractor in 
a format intended for such targeting. Based on that information, the 
agency can then more accurately decide what level of review would be 
a most effective expenditure of its resources, be it an off-site 
review of contractor records, targeted on-site reviews at a 
contractor's facility that focus on specific issues, or full-scale 
on-site reviews that concentrate on multiple issues. Without the EO 
Survey, the agency is less able to decide what level of review is 
most appropriate, and risks expending resources on a level of review 
inappropriate for that contractor.
* * * * *
    OFCCP contends that it can better build upon the tiered-review 
process through use of new procedures such as Active Case Management 
(used in connection with desk audit reviews) and proposed standards 
for identifying systemic compensation discrimination * * * 
[H]owever, these procedures would seem to factor into the tiered-
review process only after the initial selection stages. The EO 
Survey would accordingly surpass these procedures in terms of its 
capacity to build upon the tiered-review process by identifying 
contractors with systemic pay discrimination issues before deciding 
what level of review to conduct. * * * Thus, not only is the EO 
Survey an effective tool for research management, but the 
alternatives proposed by [OFCCP] are wholly inadequate.\38\
---------------------------------------------------------------------------

    \38\ AFL-CIO March 28, 2006 letter at 7-9 (emphasis in 
original).

    As discussed above, the EO Survey data is not useful in the 
selection process. And it is precisely at those early-stage targeting 
determinations that the AFL-CIO deemed ``essential'' that the EO Survey 
fails. Nor is its data useful in the tiered review process.\39\
---------------------------------------------------------------------------

    \39\ Assuming even minimal utility, such utility is outweighed 
by the cost to OFCCP to send out, process, input, and use the EO 
Survey data.
---------------------------------------------------------------------------

    The desk audit data is collected at the initial stages of the 
compliance review process and can be used to determine the appropriate 
level or type of review, as it is presented in a more timely, accurate, 
detailed, and less-aggregated form than the EO Survey data. Under its 
Active Case Management (ACM) procedures, OFCCP opens a larger number of 
reviews than in the past, uses automated statistical methods, and ranks 
and prioritizes establishments for a full review based on the 
probability that discrimination would be uncovered during a more in-
depth review. OFCCP closes cases during the desk audit if no 
statistical indicators are found that imply the presence of 
discrimination and thereby warrant further attention. More resources 
are then focused on full scale compliance evaluations of establishments 
where statistical indicators of systemic discrimination are found. In 
other words, using the ACM procedures and desk audit data is far 
superior in the tiered review process than using the EO Survey data.
    Furthermore, as discussed in the EO Survey NPRM, the findings of 
the Abt Report support OFCCP's conclusion that the EO Survey does not 
enhance the tiered-review process: ``[B]ecause the EO Survey has 
limited utility in predicting which contractors are engaged in systemic 
discrimination, it follows that EO Survey data would have limited 
utility in predicting whether and how the selected contractors are 
discriminating.'' 71 FR 3377. In sum, the aggregate nature of the data 
collected in the EO Survey, along with OFCCP's review of the Abt 
Report, demonstrate that the EO Survey does not enhance the tiered 
review process.
3. The EO Survey Facilitates Effective Self-Evaluations by Federal 
Contractors
    Some of the commenters opposed to the proposed rule assert that the 
very process of responding to the EO Survey can cause federal 
contractors to perform

[[Page 53039]]

self-evaluations, which will reduce discrimination without the need of 
a direct action by OFCCP. For example, the Leadership Conference on 
Civil Rights stated:

    By requiring contractors to report information they already are 
obligated to maintain, the EO Survey aims to give contractors 
greater incentive to undertake regular self-analysis--or self-
audits--without placing a heavy resource burden on OFCCP. 
Encouraging such proactive self-audits helps promote contractor 
compliance with existing legal obligations without adding on new 
responsibilities. * * * \40\
---------------------------------------------------------------------------

    \40\ Leadership Conference on Civil Rights March 20, 2006 letter 
at 3.

---------------------------------------------------------------------------
    Similarly, the American-Arab Anti-Discrimination Committee stated:

    Particularly with respect to pay inequities based on race or 
gender, the EO Survey created documentation of pay data that allowed 
employees complaining of pay inequities to precisely pinpoint such 
inequities, while also allowing employers to point to their EO 
Survey responses to counter allegations of pay inequities. Without 
the EO Survey, the task of identifying problem employers becomes 
more difficult, and discrimination problems can only be addressed 
retroactively, after the harm has been done and via an often 
prohibitively expensive and time-consuming process.\41\
---------------------------------------------------------------------------

    \41\ American-Arab Anti-Discrimination Committee March 20, 2006 
letter at 1-2.

    The effectiveness of the EO Survey in promoting self-evaluations, 
however, is undermined by EO Survey data itself, which is presented in 
such an aggregate form that it cannot be used to identify 
discrimination. As previously explained, the data gathered by the EO 
Survey include information, in summary form, about personnel 
activities, compensation and tenure data, and information about the 
contractor's affirmative action program. None of this information alone 
is sufficient to indicate discrimination or the lack thereof in any 
contractor establishment. The data is aggregated, which makes it 
virtually impossible to determine whether similarly situated employees 
or applicants are treated equally.
    Commenters noted the lack of utility of EO Survey data in 
performing self-evaluations. Morgan, Lewis & Bockius LLP stated:

    Because the EO Survey does not group similarly situated 
employees and includes no data regarding employees' qualifications 
or the qualifications of any position, no analysis of EO Survey data 
will satisfy the referenced legal standards for assessing unlawful 
discrimination. With respect to grouping of employees, the EO Survey 
aggregates positions into general EEO-1 occupational categories such 
as Officials and Managers and Professionals. The EEO-1 occupational 
categories do not only contain employees who are similarly situated 
in terms of hiring, promotions, compensation, and termination 
decisions, but countless other non-similarly situated categories * * 
*. In addition to comparing dissimilar employees, the EO Survey does 
not capture any data on applicants' or employees' qualifications. 
Because the EO Survey data does not group similarly situated 
employees and fails to address qualifications, it does not serve as 
a useful basis for conducting a self-evaluation of personnel 
practices to ensure nondiscrimination. * * * \42\
---------------------------------------------------------------------------

    \42\ Morgan, Lewis & Bockius LLP March 27, 2006 letter at 4-5. 
Morgan, Lewis further claims that remedying perceived disparities 
resulting from an analysis of the EO Survey data may cause 
contractors to inadvertently violate Title VII. Id. at 5-6.

    Specifically referencing compensation self-analyses, the U.S. 
Chamber of Commerce, as described previously, noted that the data is 
reported on a broad EEO-1 category basis, which OFCCP cannot use to 
assess the compensation of similarly-situated employees and that the 
data cannot be subjected to a valid statistical analysis. The U.S. 
Chamber of Commerce also stated that the EO Survey ignores the myriad 
non-discriminatory factors which may affect compensation.\43\ Indeed, 
the EO Survey compensation data cannot be used to comply with OFCCP's 
new voluntary guidelines for performing compensation self-evaluations. 
See Voluntary Guidelines for Self-Evaluation of Compensation Practices 
for Compliance With Nondiscrimination Requirements of Executive Order 
11246 With Respect to Systemic Compensation Discrimination, 71 FR 35114 
(June 16, 2006) (``Voluntary Guidelines''). Specifically, EO Survey 
compensation data is reported in EEO-1 category groupings, whereas the 
Voluntary Guidelines require contractors to group employees who are 
similarly situated, which means they perform similar work and occupy 
positions which are similar in responsibility level, and similar in the 
skills and qualifications involved in the positions. 71 FR 35120. The 
compensation data, as reported on the EO Survey, cannot satisfy the 
standards of the Voluntary Guidelines.\44\
---------------------------------------------------------------------------

    \43\ Crowell & Moring LLP March 28, 2006 at 3 (representing the 
U.S. Chamber of Commerce).
    \44\ The broad EEO-1 category groupings under the EO Survey will 
also not be useful for OFCCP when it investigates compensation 
discrimination, as the groupings are too aggregate to satisfy the 
``similarly situated'' standard.
---------------------------------------------------------------------------

    The ``similarly situated'' standard is also used in the recently 
published Interpreting Nondiscrimination Requirements of Executive 
Order 11246 With Respect to Systemic Compensation Discrimination, 71 FR 
35124 (June 16, 2006) (``Systemic Standards''). The Systemic Standards 
are standards OFCCP uses in investigating potential systemic 
compensation discrimination. These Systemic Standards will make OFCCP 
more effective at rooting out systemic pay discrimination.
    Some commenters who support the proposed rulemaking stated that the 
EO Survey is not an effective self-evaluation tool or that there are 
more effective means to induce contractors to perform self-evaluations. 
For example, the Equal Employment Advisory Council (EEAC) asserts that 
based on its own survey: ``[T]he EO Survey simply does not `provide 
contractors with a useful tool for self-evaluation,' evidenced by the 
fact that 96% of all establishments responding to a survey conducted by 
EEAC reported that `completing the Survey was not useful in monitoring 
company EEO and affirmative action compliance.' '' \45\ Morgan, Lewis & 
Bockius LLP states that Title VII, and its potential to result in 
punitive damages liability, is a more effective incentive for self-
evaluation than the EO Survey.\46\
---------------------------------------------------------------------------

    \45\ EEAC March 21, 2006 letter at 7 (emphasis in original).
    \46\ Morgan, Lewis & Bockius, LLP March 27, 2006 letter at 6.
---------------------------------------------------------------------------

    Other commenters point to OFCCP's recent initiatives as more 
effective inducements for self-evaluation. For example, the American 
Bakers Association stated:

    ABA supports the premise of the EO Survey as it requires baking 
companies who have federal contracts to take affirmative steps to 
identify and eliminate impediments to equal employment opportunity. 
However, the Survey imposes a significant administrative burden on 
ABA members who are required to complete the EO Survey. * * * Any 
beneficial role that the EO Survey was intended to provide through 
reinforcement of contractor obligations has, in recent years, been 
accomplished through other agency initiatives. For example, outreach 
seminars and workshops, recommendations as to self-evaluation 
methods, and enhanced reference (and instructional) material on the 
OFCCP Web site all have contributed greatly to the awareness of 
contractors and their ability to access the important information 
relevant to their programs.\47\
---------------------------------------------------------------------------

    \47\ American Bakers Association March 13, 2006 letter at 1-2. 
It also stated that numbers of false positives and false negatives 
generated by the EO Survey demonstrate that the EO Survey has 
minimal benefit in improving contractor self-awareness and 
encouraging self-awareness. Id.

    Likewise, the National Association of Manufacturers stated, ``[We] 
support OFCCP's continuing efforts to provide accessible compliance 
resources, particularly through its website, which are far more 
effective in assisting federal contractors in mastering their

[[Page 53040]]

compliance obligations than expending time and resources on completing 
a non-useful EO Survey.'' \48\
---------------------------------------------------------------------------

    \48\ Fortney & Scott, LLC March 26, 2006 letter at 6 
(representing the National Association of Manufacturers).
---------------------------------------------------------------------------

    Indeed, as detailed in the NPRM, OFCCP has significantly increased 
its compliance assistance efforts in recent years to heighten 
contractors' awareness of their equal opportunity obligations and to 
encourage self-evaluations through methods other than the EO Survey. 
OFCCP's compliance assistance includes over 1,000 regular compliance 
assistance seminar