Equal Employment Opportunity: Information on Personnel Actions,  
Employee Concerns, and Oversight at Six DOE Laboratories	 
(18-FEB-05, GAO-05-190).					 
                                                                 
In April 2002, GAO identified the need to strengthen equal	 
employment opportunity (EEO) oversight at three Department of	 
Energy (DOE) national weapons laboratories and recommended that  
DOE and the Department of Labor's (DOL) Office of Federal	 
Contract Compliance Programs (OFCCP) collaborate to ensure the	 
laboratories complied with EEO requirements. GAO was subsequently
asked to examine six other DOE laboratories and determine (1)	 
whether differences exist for managerial and professional women  
and minorities compared with men and Whites in salaries, merit	 
pay increases, separation patterns, and promotion rates; (2) what
EEO concerns laboratory women and minorities have raised; and (3)
what DOE and OFCCP have done to implement GAO's earlier 	 
recommendation. 						 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-190 					        
    ACCNO:   A17976						        
  TITLE:     Equal Employment Opportunity: Information on Personnel   
Actions, Employee Concerns, and Oversight at Six DOE Laboratories
     DATE:   02/18/2005 
  SUBJECT:   Comparative analysis				 
	     Data collection					 
	     Employee promotions				 
	     Employment discrimination				 
	     Employment of minorities				 
	     Fair employment programs				 
	     Interagency relations				 
	     Labor statistics					 
	     Laboratories					 
	     Merit compensation 				 
	     Racial discrimination				 
	     Salary increases					 
	     Sex discrimination 				 
	     Statistical data					 
	     Women						 
	     Minorities 					 
	     Contractors					 
	     Civilian employees 				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-190

Report to the Chair, Subcommittee on Energy, Committee on Science, House
of Representatives

February 2005

EQUAL EMPLOYMENT OPPORTUNITY

Information on Personnel Actions, Employee Concerns, and Oversight at Six
DOE Laboratories

Contents

Tables

Figures

Abbreviations

February 18, 2005Letter

The Honorable Judy Biggert Chair, Subcommittee on Energy Committee on
Science House of Representatives

Dear Madam Chair:

The Department of Energy (DOE) is the largest civilian contracting agency
in the federal government, with about 90 percent of its annual budget
spent on contracts. As part of its contract costs, DOE can reimburse its
contractors for litigation costs associated with cases brought against
them.1 Since fiscal year 1998, DOE has approved nearly $57 million for
reimbursement to its contractors for equal employment opportunity (EEO)
litigation costs.2 About $10 million of the $57 million is attributable to
six of DOE's multiprogram laboratories, and pending lawsuits could
increase this amount substantially.45 These six laboratories-Argonne,
Brookhaven, ,3,Lawrence Berkeley, Oak Ridge, and Pacific Northwest
National Laboratories, and Idaho National Engineering and Environmental
Laboratory-are managed and operated for DOE by contractors that perform
basic scientific research and environmental remediation. These
laboratories employ about 21,000 staff, of which approximately 66 percent
are managers and professionals.6 Both former Secretaries Abraham and
Richardson issued policies calling for all DOE managers, including
contractors, to foster a culture that embraces diversity and to ensure
that all employees have equal opportunity with respect to hiring,
promotions, and professional development.

Under Title VII of the Civil Rights Act of 1964 (Title VII), as amended,
employers cannot discriminate against their employees or job applicants on
the basis of race, color, religion, sex, or national origin. The Equal
Employment Opportunity Commission (EEOC) has primary responsibility for
enforcing compliance with the act for the U.S. workforce.

Executive Order 11246, as amended, which applies to federal contractors,
such as those that manage and operate many of DOE's laboratories,
prohibits the same type of discrimination as prohibited by Title VII, and
also requires that employers take affirmative action to ensure that
employees and job applicants are treated fairly without regard to race,
color, religion, sex, or national origin. The Department of Labor enforces
the executive order and has assigned this responsibility to its Office of
Federal Contract Compliance Programs (OFCCP). OFCCP investigates
complaints of employment discrimination, conducts compliance evaluations,
and takes administrative and enforcement actions when necessary. Under an
agreement between EEOC and OFCCP, EEOC generally investigates individual
complaints of discrimination against federal contractors, while OFCCP
generally investigates discrimination

complaints filed against federal contractors involving groups of people or
patterns of discrimination.7

Under the executive order, DOE is responsible for ensuring that its
contracts contain the EEO provisions OFCCP requires, for cooperating with
OFCCP, and for providing information and assistance as needed. In
addition, the Federal Acquisition Regulation (FAR) requires DOE to ensure
that it carries out the requirements of the subpart of the FAR concerning
EEO and that it cooperates with and assists OFCCP in fulfilling that
office's responsibilities.8 The primary responsibility for complying with
EEO requirements rests with the laboratory managers. In addition to
implementing its responsibilities concerning EEO, DOE has a policy to
promote diversity in its contractors' workforce.9 The DOE Acquisition
Regulation (DEAR), which supplements the FAR, requires that management and
operating contracts contain a clause requiring contractors to prepare
diversity plans. DOE's policy in pursuit of diversity is also reflected in
its contracting officers' guidance.

In an April 2002 report,10 GAO identified the need to strengthen EEO
oversight at three of DOE's nine multiprogram laboratories-Los Alamos,
Sandia, and Lawrence Livermore National Laboratories----and recommended
that DOE and OFCCP work more collaboratively to ensure the laboratories'
compliance with EEO.11

As agreed with your office, this report examines the status of women and
minorities at DOE's other six multiprogram laboratories. Specifically, for
fiscal years 2001 through mid-2004, we examined (1) whether statistically
significant differences exist between women and minorities when compared
with men and Whites in their salaries, the amount of their merit pay
increases, and their separation patterns, and whether there were
substantial differences in the promotion rates of groupings of laboratory
staff by race/ethnicity and sex when compared with White men;12 (2) what
EEO concerns women and minorities raised at these laboratories; and (3)
what actions DOE and OFCCP have taken to implement GAO's recommendation to
work collaboratively. We did not draw conclusions about whether the
laboratories have or have not discriminated against any employee or group
of employees. Our analysis of differences in salaries, merit pay
increases, separations, and promotions focuses on managers and
professionals, while our review of employee concerns includes all
laboratory staff.

To determine whether there were differences in salaries, merit pay
increases, and separation patterns for women and minorities when compared
with men and Whites, and in promotion rates of groupings of laboratory
staff by race/ethnicity and sex when compared with White men, we used data
from each laboratory's personnel database for fiscal years 2001 through
mid-2004. For our analyses of statistically significant differences in
salaries, merit pay increases, and separations, we developed
laboratory-specific regression models. By statistically significant
differences, we mean that we are 95 percent confident that these
differences are too large to have been produced by chance or random
fluctuations and that they reflect real differences in the populations
being compared. Our analyses of statistically significant differences are
not designed to prove or disprove discrimination in a court of law like
analyses conducted by OFCCP or EEOC, nor do they establish whether the
differences are of practical significance that would require corrective
action by the laboratories. Rather, our analyses use a standard method
designed to provide information at an aggregate level about differences in
personnel actions, such as salaries, merit pay increases, and separation
patterns, for women and minorities at the laboratories that may need
further investigation. Because the laboratories have somewhat different
personnel systems and practices, our analyses of salaries, merit pay
increases, and separation patterns included variables specific to each
laboratory, as well as those common to all six. For example, we included
laboratory-specific controls for organizational structure and occupational
classification systems. Additionally, we included controls for age; tenure
at the laboratory or within grade; employment status (postdoctoral,
part-time and temporary status); management status; citizenship; and
education level where data were available. Consequently our analyses of
statistically significant differences are not exhaustive, nor do they
prove or disprove discrimination. Rather they are designed to identify
issues or problems that may need to be investigated further.

To determine whether promotion rates of groupings of laboratory staff by
race/ethnicity and sex were substantially different from White men, we
applied the 80 percent rule set out in the federal government's Uniform
Guidelines on Employee Selection Procedures.13 We determined that the
laboratories' personnel data were sufficiently reliable for our purposes.

To determine the EEO concerns of women and minorities at the laboratories,
we interviewed representatives from employee groups for women and
minorities at the laboratories and reviewed recent DOE and laboratory
surveys and studies. We also collected data from each laboratory on
complaints filed and investigated within the laboratory (internal
complaints). Finally, we examined complaints filed with organizations
outside of the laboratory, such as EEOC, OFCCP, or a state and local fair
employment practices agency (external complaints). We did not attempt to
prove or disprove the validity of these concerns, nor did we assess the
laboratories' efforts to address them. We determined that the data we
collected were sufficiently reliable for our purposes.

To determine the actions DOE and OFCCP have taken to implement our 2002
recommendation, we met with DOE and OFCCP officials responsible for
implementing our recommendation and for EEO oversight at DOE's
laboratories. We also examined the roles and responsibilities of DOE

Women and minority staff expressed a number of concerns about their fair
and equitable treatment at the laboratories. According to DOE and
laboratory studies and our interviews, concerns focused primarily around
underrepresentation, the lack of career development opportunities, and the
need to improve the laboratory work environment. For example, some women
and minorities explained that the laboratories could further career
development opportunities by offering mentoring programs. According to our
analysis of complaints filed and investigated within the laboratories, the
primary issues varied by laboratory and cannot be readily compared across
laboratories because each laboratory records and categorizes these
complaints differently. For external complaints filed under Title VII, the
most often cited issues we identified were pay-an area where we have
already identified statistically significant differences in the preceding
section-and terminations. The highest number of external cases filed dealt
with sex or race matters.

As a result of our April 2002 recommendation, OFCCP and DOE staff met to
discuss the possible creation of a more formal relationship through a
memorandum of understanding. While reviewing OFCCP's draft memorandum, the
Department of Labor raised questions about DOE's authority and
responsibility for EEO matters at the laboratories. As a result, OFCCP has
not sent the draft memorandum to DOE for coordination. OFCCP maintains
that Executive Order 11246, as amended, made OFCCP solely responsible for
enforcing federal contractors' compliance with EEO obligations. DOE
officials agree. However, DOE and OFCCP appear to disagree about the scope
of DOE's authority to implement DOE's diversity policies. DOE maintains
that its activities in pursuit of diversity under the DEAR provision and
its contracting guidance are independent of OFCCP's enforcement authority.
OFCCP officials, however, have raised concerns about whether DOE's
implementation of its diversity policies under the DEAR and its guidance
might encroach upon OFCCP's enforcement authority.

To understand the implications of the statistical differences we found and
evaluate their practical significance, we are recommending that the
Secretary of Labor direct OFCCP to work with the laboratories to determine
their causes and take the necessary corrective steps, if appropriate, to
address any EEO problems identified. We are also recommending that the
Secretaries of Labor and of Energy work together to define the scope of
DOE's contract administration oversight responsibility for diversity.

In commenting on a draft of this report both DOE and the Department of
Labor agreed to work together to resolve EEO oversight issues, although
neither specifically commented on our recommendations. However, DOE stated
that our analytical method differs from the method their laboratories
would use in their analyses. We disagree. The methodology that we use is
similar to the methodology that OFCCP has recently proposed that federal
contractors with more than 250 employees use to conduct annual
self-evaluations of compensation practices. Our analyses of salaries,
merit pay, and separation patterns at each of the laboratories included
factors that OFCCP's proposed guidance identifies as legitimately
affecting compensation, such as experience, education, and performance, as
well as other factors that the laboratories specifically identified.

In addition, both DOE and the Department of Labor commented that our
analytical methods differ from those OFCCP uses. We acknowledge these
differences, but note that our analyses and OFCCP's analyses have a
different purpose. As we state in our report, our analyses were not
designed to prove or disprove discrimination, but are a first step in
identifying whether differences exist that may require further
investigation. In contrast, OFCCP uses its analyses to determine whether
discrimination has occurred.

Background

The six nonweapons multiprogram laboratories we reviewed-Argonne,
Brookhaven, Lawrence Berkeley, Oak Ridge, and Pacific Northwest National
Laboratories, and Idaho National Engineering and Environmental
Laboratory-are managed by contractors from both universities and private
industry. The laboratories have a total workforce of about 21,000
employees and range in size from fewer than 2,500 at Lawrence Berkeley to
more than 5,000 at Idaho. Figure 1 shows the location of these six
multiprogram laboratories.

Figure 1: Location of the Six Multiprogram Laboratories Reviewed

Executive Order 11246, as amended, provides generally the same
prohibitions against discrimination for federal government contractors as
Title VII of the Civil Rights Act of 1964, as amended.14 The order states
that federal contractors will not discriminate against an employee or
applicant for employment on the basis of race, color, religion, sex, or
national origin. In addition to the requirements of Title VII, the order
further states that federal contractors will take affirmative action to
ensure that applicants and employees are treated without regard to their
race, color, religion, sex, or national origin in personnel actions,
including recruitment and hiring, pay, benefits, promotion, selection for
training, demotions and transfers, lay-offs, and termination. Under OFCCP
regulations, the contractors must develop affirmative actions plans that
spell out the steps they will take to ensure EEO.

OFCCP's regulations implementing the executive order require contractors,
including the laboratories, to submit data annually to EEOC on

specific job categories, by race/ethnicity and sex.15 Private-sector
employers provide annual employment statistics by sex for each of nine
major job categories and for each of five population groups: Whites,
Blacks, Hispanics, Asians or Pacific Islanders, and American Indians or
Alaskan Natives. We used data on race/ethnicity, sex, and the nine job
categories that the laboratories are required to provide to EEOC and
combined those data into three job category groups: managers and
professionals; technicians, clerks, and craft workers; and operatives,
laborers, and service workers.16

Figure 2 shows the six laboratories' staff by job category.

Figure 2: Composition of Staff by Job Category, 2003

According to their 2003 data, the six laboratories vary somewhat in the
composition of their staffs. While minorities account for less than 10
percent of the staff at Idaho, they account for more than 30 percent at
Lawrence Berkeley. Figure 3 shows the proportion of Whites to minorities
at the six laboratories.

Figure 3: Percentage of Laboratory Population by Race (Whites and
minorities), 2003

The ratios of men to women were similar across the laboratories, ranging
from 61 percent men and 39 percent women at Pacific Northwest to 75
percent men and 25 percent women at Idaho. Figure 4 shows the composition
of staff at all six laboratories by sex.

Figure 4: Percentage of Laboratory Staff by Sex, 2003

The laboratories take a number of actions to fulfill their EEO
responsibilities. These include, among other things,

o submitting information on the composition of their labor force to EEOC
and DOE;

o developing affirmative action programs that are designed not only to
improve the number of women and minorities for specific jobs in which they
are underrepresented but to ensure that the laboratory has fulfilled its
EEO responsibilities;

o preparing diversity plans, which detail the laboratories' efforts to
promote workforce diversity by training employees on the importance of
diversity at the laboratories and on the prevention of racial profiling;

o providing mechanisms through which staff can raise EEO concerns or
complaints; and

o developing annual self-assessments on their EEO performance for DOE's
review.

Appendix II provides more detailed data for each of the six laboratories.

Available Data Show Some Statistically Significant Differences in
Salaries, Merit Pay Increases, and Separation Patterns for Managerial and
Professional Women and Minorities and One Instance of Substantially Lower
Promotion Rates

For fiscal years 2001 through mid-2004, we found some statistically
significant differences in salaries, merit pay increases, and separation
patterns for managerial and professional women and minorities when
compared with men and Whites. We also found one instance in which
promotion rates for selected minorities were substantially lower than for
White men. Statistically significant differences do not prove or disprove
discrimination; rather, they provide information at an aggregate level
about differences in personnel actions and may indicate a need for further
investigation.

Women Tended to Receive Lower Salaries than Men, While Salaries for
Minorities Were Generally Equal to Those of Whites

We found statistically significant differences in salaries for women when
compared with men in managerial and professional job categories at five of
the six DOE laboratories, and for minorities in managerial and
professional job categories when compared with Whites at one laboratory.
These statistical differences remained after holding constant occupational
classification, organizational division, age, tenure at the laboratory or
within grade, employment status (including postdoctoral, part-time and

temporary status), management status, citizenship, and education level
where data were available.17 Table 1 presents the results of our analysis.

Table 1: Percent Differences in Salaries for Women and Minorities, Fiscal
Years 2001 through mid-2004

Laboratory         Women compared with men Minorities compared with Whites 
Argonne                              -3.6% 
Brookhaven                           -3.0% 
Idaho                                -2.0% 
Lawrence Berkeley                    -3.1%                           -1.5% 
Oak Ridge                            -3.6% 
Pacific Northwest                          

Source: GAO analysis of laboratory data.

Note: Negative numbers indicate that the group earned a lower salary than
its counterpart and blank spaces indicate that there were no significant
differences for that group compared with men and Whites.

Half of the Laboratories Showed Differences in Merit Pay Increases for
Either Women or Minorities or Both Compared with Men and Whites

While our analysis showed that the merit pay increases for full-time and
nontemporary managerial and professional women and minorities tended to be
statistically comparable to merit pay increases for their respective
counterparts at three of the laboratories, results were mixed at the other
three laboratories. These statistical differences remained after holding
constant occupational classification, organizational division, performance
score, age, salary, tenure at the laboratory or within grade, employment
status, management status, citizenship, and education level where data

were available.18 Merit pay increases were higher for women and minorities
at one, higher for women at another, and lower for minorities at the third
compared with men and Whites. Table 2 presents the results of our
analysis.

Table 2: Percent Differences in Merit Pay Increases for Women and
Minorities, Fiscal Years 2001 through mid-2004

Laboratory         Women compared with men Minorities compared with Whites 
Argonne                                                              -3.7% 
Brookhaven                            7.0%                            3.2% 
Idaho                                 5.3% 
Lawrence Berkeley                          
Oak Ridge                                  
Pacific Northwest                          

Source: GAO analysis of laboratory data.

Note: Positive numbers indicate that the group earned a higher merit pay
increase than its counterpart, negative numbers indicate that the group
earned a lower merit pay increase, and blank spaces indicate that there
were no statistically significant differences for that group compared with
men and Whites.

Separation Patterns for Women and Minorities Were Generally Comparable to
Men and Whites

Managerial and professional women tended to separate from their jobs
(leave) at a comparable rate as men at five of the six laboratories.
Similarly, managerial and professional minorities tended to leave their
jobs at a comparable rate as Whites at five of the laboratories. However,
at one laboratory, women were more likely to leave than men, and at
another laboratory, minorities were more likely to leave than Whites.
These statistical differences remained while holding constant performance
score, age, salary, tenure at the laboratory or within grade, employment
status, management status, citizenship, and education level where data
were available. Separations include both voluntary actions, such as
retirement, and involuntary actions, such as reductions in force and
terminations for cause. While a greater likelihood of separation is not
necessarily indicative of race or sex-based problems at the laboratories,
race and sex-based patterns of separation are important to evaluate in
order to ensure they are not indicative of broader concerns about women's
or minorities' treatment at the laboratories. Table 3 presents the results
of our analysis.

The 1.4 shown in table 3 indicates that women at Idaho were 40 percent
more likely to leave than men, while the 1.7 indicates that minorities at
Pacific Northwest were 70 percent more likely to leave than Whites. Had
our analysis shown that any group was less likely to leave than its
respective counterpart, we would have indicated this using a number less
than 1.

Table 3: Differences in the Likelihood of Separating from the Laboratories
for Women and Minorities, Fiscal Years 2001 through mid-2004

Laboratory         Women compared with men Minorities compared with Whites 
Argonne                                    
Brookhaven                                 
Idaho                                  1.4 
Lawrence Berkeley                          
Oak Ridge                                  
Pacific Northwest                                                      1.7 

Source: GAO analysis of laboratory data.

Note: Numbers greater than 1 indicate that the group has a greater
likelihood to leave than their respective counterpart and blank spaces
indicate that there were no significant differences for that group.

Promotions for Women and Minorities Were Generally Comparable to Those of
White Men

We found that management promotions for groupings of laboratory staff by
race/ethnicity and sex generally met or exceeded 80 percent of the
promotion rate of White men, with the exception of Idaho.19 At Idaho, an
additional Black man and woman, and four additional Asian men would have
been needed to meet 80 percent of the promotion rate of White men.
However, Idaho promoted nearly twice the number of Hispanics than would
have been needed to meet the 80 percent criterion for that group.

Table 4 shows, for each laboratory, the number of actual promotions in the
3-1/2-year period by race/ethnicity and sex and the additional number of
promotions needed to reach 80 percent of the White male promotion rate.
Blank spaces indicate that the 80 percent rule was met or exceeded (no
additional staff needed to meet 80 percent of the White male promotion
rate).

Table 4: Number of Promotions for Groupings by Race/Ethnicity and Sex and
Number of Additional Promotions Needed to Reach 80 Percent of the
Promotion Rate of White Men, Fiscal Years 2001 through mid-2004

                            Argonne             Brookhaven  
                             Promoted   Needed               Promoted  Needed 
White men                      419                             287 
Black men                        5                              10 
Hispanic men                     9                               8 
Asian men                       56                              27 
American Indian men              1                               0 
White women                    116                             119 
Black women                      8                              12 
Hispanic women                   4                               5 
Asian women                     15                               7 
American Indian women            0                               0 

Source: GAO analysis of laboratory data.

Notes: Numbers are rounded down.

Blank spaces indicate that the 80 percent rule was met or exceeded (no
additional staff needed to meet 80 percent of the White male promotion
rate).

Laboratory Staff Raised a Number of EEO Concerns

Women and minority staff expressed a number of concerns about their fair
and equitable treatment at the laboratories, focusing primarily on
underrepresentation, the lack of career development opportunities, and the
need to improve the laboratory work environment. We identified these
concerns through our interviews with representatives of the laboratories,
laboratory women's and minority employee groups, and our review of DOE and
laboratory surveys and studies of laboratory staff. Furthermore, according
to our analysis of Title VII complaints filed and investigated within the
laboratories (internal), the issues identified from data provided by the
laboratories varied and did not lend themselves to comparison across
laboratories because the laboratories do not record and categorize the
complaints in a consistent manner. Finally, our analysis of Title VII
complaints filed and investigated by an organization outside of the
laboratories (external) identified pay and terminations as the most often
cited areas of concern. The highest number of external cases filed dealt
with sex or race matters. As discussed in the preceding section, we found
statistically significant differences in pay. We did not attempt to prove
or disprove the validity of these concerns, nor did we assess the
laboratories' efforts to address them.

Employee Groups, Surveys, and Studies Indicated Three Primary Concerns

According to our interviews with representatives of employee groups and
our review of DOE and laboratory surveys and studies, women and minority
staffs' EEO concerns center primarily around three key areas: (1)
underrepresentation of women and minorities in science positions and
management positions, (2) lack of career development opportunities, and
(3) a laboratory work environment that needs improvement.

Underrepresentation of Women and Minorities

Women and minorities were primarily concerned with what they perceive as
underrepresentation at the laboratories, particularly in science and
management positions, according to our interviews and reviews of DOE and
laboratory surveys and studies. For example:

o The representatives stated a need for increased efforts on the part of
the laboratories to increase representation of women and minorities in
science positions. According to a representative we spoke with, this
shortage occurred in part because of limited recruiting at universities
with large numbers of minorities, such as historically black colleges and
universities. The representatives also cited a need for increased
transparency in the hiring and promotion process.

o More than 60 percent of the members of the Asian Pacific American
Committee at Oak Ridge believed that Asians were underrepresented in
management positions at the laboratory, according to a May 2002 survey.20
According to an official at Oak Ridge, the laboratory's management has
established programs to identify and overcome barriers to upward movement
within the laboratory. For example, the laboratory offered in-house
professional training for Asian staff and their managers, and the
laboratory added leadership training specifically for Asians to the
management training curriculum.

Lack of Career Development Opportunities

Some women and minority staff raised concerns about the limited career
development opportunities available at the laboratories. For example:

o Representatives of the employee groups suggested the laboratories would
benefit from instituting mentoring programs to help employees plan and
take control of their careers. The need for such programs is compounded by
the fact that women and minorities often do not have role models in
management positions, and therefore the laboratories need to address how
to help "grow staff" for such roles.

o According to an August 2003 study at Argonne,21 career development for
women ranked as one of the top three goals for the women's group to focus
its future efforts. It also cited women's lack of familiarity with career
paths for science and technology positions at the laboratory as a problem.
The report suggested that clearer identification of these paths would be
beneficial to all employees, particularly minorities, who lack a
significant number of role models at the laboratory.

Laboratory Work Environment

Although the representatives of the employee groups stated that management
has worked to create a better atmosphere for women and minorities at the
laboratories, they identified a number of issues that continue to concern
them. For example:

o According to staff at Idaho, diversity and EEO matters have lost
visibility as a management issue because the laboratory has changed its
Diversity/Affirmative Action Specialist within the human resources office
from a full-time, management position to a part-time nonmanagement
position. However, an Idaho official told us that the position focuses
full-time on EEO issues, but other human resources duties may be performed
if time permits.

o According to staff at Lawrence Berkeley, the laboratory established a
diversity committee to encourage understanding of cultural differences
among employees. However, the committee does not have upper management
representation and therefore possesses limited influence. According to a
Lawrence Berkeley official, the laboratory created the Best Practices
Diversity Council in January 2003 to address these concerns. The council
does have management representation and works to implement diversity best
practices and processes at both the division and laboratory level.

o According to a 2001 Update Survey conducted at Brookhaven,22 women
responded less favorably than laboratory staff overall about whether the
laboratory provided a working environment that was accepting of sex
differences. In addition, Black and Hispanic employees responded less
favorably than laboratory staff overall about whether laboratory
management supported diversity in the workplace.

Internal Complaints Covered a Variety of Issues

During the 3-1/2-year period we examined, five of the six laboratories
received 187 internal complaints dealing with race, color, religion, sex,
or national origin concerns.23 Because the process for collecting and
recording information, particularly specific data, on internal complaints
is different at each laboratory, we were not able to compare numbers and
types of complaints filed across the laboratories. Furthermore, the
laboratories have a number of avenues available to employees for filing
internal complaints, ranging from working through their direct supervisor
to working through the laboratory's human resources or EEO offices. As a
result, some of the laboratories do not have one centralized repository
for collecting specific or consistent information on each complaint.
Consequently, we were not able to ensure that the information we received
included the entire universe of internal complaints filed at the
laboratories. Nevertheless, the information does provide an overall
picture of the complaints dealing with race, color, religion, sex, or
national origin at each laboratory and general characteristics about the
complainants. In addition, the data represent complaints filed by staff
and investigated within the laboratories and do not necessarily indicate
any illegal activity on the part of the laboratories.

Table 5 presents the most often cited issues as classified by the
laboratories for the internal complaints at each laboratory that provided
these data.  24

Table 5: Most Often Cited Issues Identified from Internal Complaints on
File, Fiscal Year 2001 through June 25, 2004

Laboratory                  Issue                                          
Argonne                     Harassment                                     
Idaho                       Offensive comments, materials, and actions     
Lawrence Berkeley           Hostile work environment                       
Oak Ridge                   Failure to hire                                
Pacific Northwest           Sexual harassment                              

Source: GAO analysis of laboratory data.

Notes: Brookhaven did not submit data on internal complaints.

A single complaint may have more than one issue associated with it. As a
result, if a complaint has multiple issues, some of the issues may not be
for race, color, religion, sex, or national origin bases. We included any
complaints that had at least one such basis.

Our analysis of the type of staff filing these complaints at each
laboratory shows the following:

o At Argonne, Idaho, and Pacific Northwest, women filed the majority of
the complaints.

o At Oak Ridge, men filed the majority of the complaints.

o At Lawrence Berkeley, the complaints were evenly divided between men and
women.

o At Argonne, Idaho, and Pacific Northwest, White staff filed the majority
of the complaints.

o At Oak Ridge, Asian staff filed the highest number of complaints.

o At Lawrence Berkeley, Asian and Black staff filed the highest number of
complaints.

External Complaints Most Often Cited Pay and Terminations as Areas of
Concern

Laboratory employees filed 48 complaints dealing with Title VII concerns
between fiscal years 2001 through mid-2004. The highest number of external
cases filed dealt with sex or race matters.25 Lawrence Berkeley reported
the highest number of complaints (17), followed by Brookhaven (14). As
table 6 shows, pay and termination were the two issues cited most often.

Table 6: Issues Presented in External Complaints, Fiscal Year 2001 through
June 25, 2004

Issue                                                                Total 
Pay                                                                     10 
Termination and dismissal                                                6 
Harassment                                                               4 
Promotion                                                                4 
Layoff/reduction-in-force                                                3 
Americans with Disabilities Act                                          2 
Demotion                                                                 2 
Job Assignment                                                           2 
Performance evaluations or ratings                                       2 
Othera                                                                   7 
Total                                                                   42 

Source: GAO analysis of laboratory data.

Notes: The 42 issues cited in table 6 correspond to 34 external
complaints. Brookhaven did not submit information on issues associated
with its 14 external complaints.

Because a single complaint may have more than one issue associated with
it, the total number of issues does not match the total number of external
complaints filed. In addition, if a complaint has multiple issues, some of
the issues may not be for Title VII matters. We included any complaints
that had at least one Title VII basis.

a"Other" includes seven different issues that were each identified once.

As table 7 shows, minorities filed 34 of the 48 complaints, with Black
women filing the highest number of any minority group (11 of 48).

Table 7: External Complaints by Sex and Race/Ethnicity, Fiscal Year 2001
through June 25, 2004

                                             Sex       
Race/ethnicity                                Women                    Men 
White                                            11                      3 
Black                                            11                     10 
Hispanic                                          4                      2 
Asian                                             3                      4 
American Indian                                   0                      0 
Total                                            29                     19 

Source: GAO analysis of laboratory data.

OFCCP and DOE Have Distinct Roles and Responsibilities for EEO

In our April 2002 report, we recommended that DOE and OFCCP explore the
costs and benefits of a more formal, collaborative relationship toward
their common goal of EEO compliance at the laboratories. OFCCP and DOE
staff met to discuss the implementation of the recommendation and the
possible creation of a more formal relationship through a memorandum of
understanding. OFCCP provided the draft memorandum to the Department of
Labor's Office of the Solicitor for review to ensure that it complied with
applicable laws and regulations. As a result of its review, the Office of
the Solicitor raised questions about DOE's legal authority to monitor and
to enforce Executive Order 11246, as amended. As a result, OFCCP has not
sent the draft memorandum to DOE for coordination. In addition, in the
course of this review, GAO observed that the DEAR requires that DOE
management and operating contracts contain a diversity clause, requiring
the contractor to submit a diversity plan to DOE. Also, DOE's guidance for
contracting officers contains provisions calling on the contracting
officer to evaluate the contractor's diversity activities. While OFCCP has
not reached any conclusions, it has concerns that DOE's implementation of
these diversity provisions might encroach upon OFCCP's EEO enforcement
authority.

According to OFCCP, it is the only executive agency with the authority to
enforce a contractor's compliance with Executive Order 11246, as amended.
Before the executive order was amended in 1978, responsibility for
compliance with EEO obligations was dispersed among 11 different agencies,
including DOE. Because agency standards, procedures, and reporting
requirements were not uniform, contractors faced differing agency
requirements. The amended executive order concentrated full responsibility
for EEO contractor compliance, including the authority to issue
regulations and binding orders with the Secretary of Labor, who delegated
this responsibility to OFCCP. Under the executive order, the contracting
agencies are responsible for including the EEO clause in each nonexempt26
contract, cooperating with the Secretary of Labor by providing the
information and assistance the Secretary requires, and complying with the
terms of the executive order and implementing regulations and orders. The
obligations of contracting agencies, contractors, and subcontractors under
the executive order are set out in OFCCP regulations27 and provisions of
the Federal Acquisition Regulation (FAR).28

OFCCP maintains that DOE lacks authority as part of contract
administration to evaluate a laboratory contractor's compliance under the
contract with the EEO requirements of the executive order and its
implementing regulations. OFCCP states, for example, that DOE does not
have authority to reduce the contractor's fee for EEO violations cited by
OFCCP. In OFCCP's view, DOE's reduction of a contractor's fee would be
tantamount to imposing a monetary penalty on the contractor, which is not
authorized under the terms of the executive order and its implementing
regulations.29 In addition, according to OFCCP, to permit DOE to impose
standards or sanctions beyond those administered by OFCCP could
reintroduce inconsistencies and the conflicts between EEO objectives and
procurement objectives that led to the 1978 centralization of EEO
enforcement within the Department of Labor.

DOE officials agree that OFCCP is solely responsible for the
administration and enforcement of the EEO provisions of government
contracts. In this regard, DOE states that its EEO responsibilities are
limited to those conferred by the FAR, that is, to include the appropriate
provisions and clauses in solicitations and contracts and to refer
complaints to and cooperate with OFCCP, as prescribed by the FAR.

In addition to the FAR requirements for EEO, DOE has diversity policies
that are reflected in the DEAR requirement that each management and
operating contract contain a diversity clause, and in provisions of DOE's
reference guidance for contracting officers. DOE views its activities with
regard to the inclusion of the DEAR diversity clause and pursuit of
diversity as independent of OFCCP's jurisdiction. However, OFCCP has
pointed out several areas in the DEAR and DOE's guidance as possibly
infringing on OFCCP's EEO enforcement authority. For example:

o The DEAR requires each management and operating contract to contain a
clause requiring that the contractor submit a diversity plan to the DOE
contracting officer within 90 days of the effective date of the
contract.30  This plan, at a minimum, is to address the contractor's
approach for promoting diversity through its workforce, education and
community outreach, subcontracting, economic development, and the
prevention of racial profiling. DOE's policy guidance for contracting
officers31 states that this plan may discuss how the contractor has or
plans to establish and maintain result-oriented EEO and affirmative action
programs in accordance with the requirements of EEO and affirmative action
contract clauses and how the contractor's organization includes or plans
to include elements/dimensions of diversity that might enhance such
programs.32 OFCCP told us that these provisions raised concerns about
DOE's implementation of the diversity plan requirement.

o According to DOE's policy guidance for contracting officers, each DOE
contracting officer is to evaluate the contractor's performance in
implementing its diversity plan.33 In addition, contracting officers are
to evaluate the extent to which the plan demonstrates the contractor's
commitment, among other things, to diversity, cultural sensitivity, and
inclusion in all aspects of its business practices, the workplace, and
relations with the community at large.34

OFCCP believes that DOE's diversity plan requirements and policy may put
DOE in the position of evaluating contractors' compliance with the
requirements of the EEO laws that OFCCP is solely responsible for
enforcing and encroach upon OFCCP's enforcement authority. DOE states,
however, that its efforts to promote diversity under its contracts through
the diversity clause are independent of enforcement of the EEO clause and
OFCCP's authorities under the executive order. DOE views its efforts as
being in pursuit of an agency value of workforce diversity in carrying out
its mission.

Conclusions

DOE has issued policies that are intended to ensure that the department
maintains a respectful and productive work environment for both federal
and contractor employees. While our findings of statistically significant
differences in salaries, merit pay increases, and separation patterns for
managerial and professional women and minorities compared with men and
Whites and our finding of a difference at one laboratory in promotion
rates for certain minority groups compared with White men do not prove or
disprove discrimination, they may indicate a need for further
investigation.

OFCCP and DOE agree that OFCCP has sole responsibility under the executive
order for the administration and enforcement of the EEO provisions of
government contracts. However, the agencies do not appear to agree about
the extent to which DOE has authority to implement its diversity policies.
While DOE views its implementation of the DEAR diversity clause and
diversity guidance as independent of OFCCP's authorities, OFCCP has raised
concerns about whether DOE's implementation might encroach upon its
enforcement authority. Although we commend DOE for its desire to have a
diverse workforce and its initiative for including contract clauses to
achieve that goal, the departments of Labor and of Energy need to clarify
and reach agreement about DOE's role concerning its contractors' diversity
activities.

Recommendations for Executive Action

To understand the implications of the statistical differences we found and
evaluate their practical significance, we recommend that the Secretary of
Labor direct OFCCP to work with the laboratories to determine their causes
and take the necessary corrective steps, if appropriate, to address any
EEO problems identified.

We also recommend that the Secretaries of Labor and of Energy work
together to define the scope of DOE's contract administration oversight
responsibility for diversity.

Agency Comments and Our Evaluation

We provided DOE and the Department of Labor with draft copies of our
report for their review and comment. DOE's written comments on our report
did not address our recommendations, but DOE stated that it is ready to
continue its effort to work with OFCCP to define the working relationship
of the two departments. DOE also expressed concern about the accuracy of
our analyses stating that the criteria that we used differed from that
used by the laboratories and OFCCP. Actually, the methodology that we use
is similar to the methodology that OFCCP has recently proposed that
federal contractors with more than 250 employees use to conduct annual
self-evaluations of compensation practices.35 Our analyses of salaries,
merit pay, and separation patterns at each of the laboratories included
factors that OFCCP's proposed guidance identifies as legitimately
affecting compensation, such as experience, education, and performance, as
well as several other factors that some of the laboratories specifically
identified. In fact, we tested a number of modifications to our models in
response to questions posed by several of the laboratories. None of these
modifications resulted in significant changes to our findings.
Furthermore, as we state in our report, unlike the analyses conducted by
OFCCP, our analyses do not, and were not designed to, prove or disprove
discrimination. Rather, our analyses are intended to identify issues or
problems that may need to be investigated further.

DOE also commented that our analyses were questionable because we revised
the statistics for Idaho. DOE suggested that if similar vigor were applied
to all the report's statistics, other changes might result. We disagree.
The revisions to the Idaho statistics were not the result of a change to
our methodology, but resulted from including additional data provided by
the laboratory during the report's comment period. The other laboratories
had provided this data earlier. Instead of undermining the validity of our
analyses, our actions reflect our commitment to reporting complete and
accurate information.

Lastly, DOE restated its position that its role in fostering contractors'
workforce diversity does not conflict with OFCCP's role in enforcing
contractors' equal employment obligations.

In its written comments, the Department of Labor pointed out that OFCCP's
methods of statistical analysis differ from the analysis methods we used
in this report and that OFCCP's methods are designed to determine if a
contractor is engaged in unlawful employment discrimination. Our report
recognizes this role for OFCCP and clearly states that our analyses are
not designed to prove or disprove discrimination in a court of law like
OFCCP's analyses. Our report clearly states that our analyses are designed
to provide information at an aggregate level about differences in
personnel actions that may need further investigation. Because our
analyses identified statistically significant differences in salaries,
merit pay increases, and separation patterns, and substantial differences
in promotion rates for women and minorities at the six laboratories, we
have recommended that OFCCP work with the laboratories to determine the
causes of these differences and take the necessary corrective steps, if
appropriate, to address any EEO problems identified. In response to our
recommendations, the Department of Labor stated that OFCCP will continue
to work with DOE with regard to contract oversight.

DOE's and the Department of Labor's written comments are presented in
appendixes III and IV, respectively.

As agreed with your office, unless you publicly announce the contents of
this report earlier, we plan no further distribution until 30 days from
the report date. At that time, we will send copies of this report to
appropriate congressional committees; the Secretary of Energy; the
Secretary of Labor; the Chair, Equal Employment Opportunity Commission;
the Director, Office of Management and Budget; and other interested
parties. We will also make copies available to others upon request. In
addition, the report will be available at no charge on the GAO Web site at
http://www.gao.gov.

Key contributors to this report are listed in appendix V.

Sincerely yours,

Robin M. Nazzaro Director, Natural Resources and Environment

Scope and MethodologyAppendix I

This appendix details the methods we used to (1) determine whether there
were differences in salaries, merit pay increases, and separation patterns
for women and minorities when compared with men and Whites, and in the
promotion rates of groupings of laboratory staff by race/ethnicity and sex
compared with White men, for fiscal years 2001 through mid-2004; (2)
describe equal employment opportunity (EEO) concerns raised by laboratory
staff; and (3) determine the actions the Department of Energy (DOE) and
the Office of Federal Contract Compliance Programs (OFCCP) have taken to
implement our earlier recommendation to work more collaboratively to
ensure the laboratories' compliance with EEO.1 In addition, this appendix
provides information on the methods we used to develop the personnel data
provided in the background section of the letter and in appendix II.

Our review focused on personnel actions and EEO concerns at six DOE
multiprogram laboratories: Argonne National Laboratory in Argonne,
Illinois, and Idaho Falls, Idaho; Brookhaven National Laboratory in Upton,
New York; Idaho National Engineering and Environmental Laboratory in Idaho
Falls, Idaho; Lawrence Berkeley National Laboratory in Berkeley,
California; Oak Ridge National Laboratory in Oak Ridge, Tennessee; and
Pacific Northwest National Laboratory in Richland, Washington. We
interviewed and obtained data and documentation from relevant officials at
DOE's, OFCCP's, and EEOC's headquarters offices in Washington, D.C.; DOE
officials in the Chicago, Idaho, Oak Ridge, and Richland operations
offices and the Berkeley and Brookhaven site offices; and laboratory
officials at the six laboratories. We performed our statistical analysis
of salaries, merit pay increases, and separation patterns by comparing
women to men, and minorities to Whites. For our analysis of promotions, we
compared men and women in each racial/ethnic group to White men. Our
statistical analysis of laboratory staff includes all exempt management
and professional staff at the laboratories.2 Limited-term staff (such as
postdoctoral students who hold professional occupations on a temporary
basis) were included in the analysis of salaries, but not in the analyses
of merit pay increases, promotions, or separations. We reviewed only
exempt employees because they represent the majority of laboratory staff,
laboratories maintain personnel data for these employees, and their
salaries and benefits are not negotiated by a union.

To determine whether there were differences in salaries, merit pay
increases, and separation patterns for women and minorities when compared
with men and Whites, and in promotion rates for women and minorities in
each racial/ethnic group compared with White men, in fiscal years 2001
through mid-2004, we applied statistical tests to the laboratory-provided
employee data on personnel actions. We requested data for all exempt
management and professional staff for the period we reviewed. Because many
of the laboratories were undergoing changes in size, funding, and
structure, we chose to analyze the most recent 3-1/2 years of personnel
actions, rather than just 1 year. We believed the 3-1/2-year period would
provide a more accurate picture of laboratory compensation practices
because unique funding or organizational effects could produce marked
differences in separations, merit pay increases, and promotions in a
particular year. Because we were combining information across years, we
reviewed the data to ensure individuals' demographic information remained
constant (e.g., race, sex, education level). Where necessary, we either
excluded inconsistent cases or included variables in the models to adjust
for nonconstant data. We assessed the reliability of each laboratory's
data by (1) using advanced electronic testing, (2) reviewing documentation
on the data systems, and (3) interviewing knowledgeable staff about the
data and data system internal controls and quality reviews. Based on our
assessment, we determined that the personnel data from all laboratories
were sufficiently reliable for our purposes.

The federal government's Uniform Guidelines on Employee Selection
Procedures direct agencies to analyze personnel actions of groups
protected by Title VII of the Civil Rights Act. Under the guidelines,
agencies analyze personnel actions separately by race/ethnicity and by
sex. At congressional request, we performed our statistical analyses in
accordance with these guidelines. We analyzed the laboratory data at the
individual level using the complete population of exempt management and
professional laboratory staff. The laboratories have somewhat different
personnel systems and practices, and although we applied the same modeling
techniques for each laboratory, we performed the analyses separately for
each laboratory. We consulted with the laboratories regarding our
analytical approach to ensure that we were receiving the appropriate data
for the analyses. Additionally, we consulted with OFCCP about our
methodology. Our analyses are not designed to prove or disprove
discrimination in a court of law; rather, they are designed to provide
information at an aggregate level about race/ethnicity and sex differences
in personnel actions at the laboratories. Therefore, our results do not
indicate whether discrimination has occurred.

Determining Statistical Differences in Salaries, Merit Pay Increases, and
Separations

To determine whether there were statistically significant race/ethnicity
and sex differences in salaries, merit pay increases, and separations at
each laboratory, we used multivariate regression techniques. Salaries and
merit pay increases were modeled with Ordinary Least Squares regression,
and separations were modeled with Logistic regression. Race/ethnicity and
sex differences in salaries, merit pay increases, and separations were
considered statistically significant if the probability of the t-statistic
or chi-square value associated with the coefficient was 0.05 or lower. In
other words, we regard race/ethnicity and sex differences in salaries,
merit pay increases, and separations as significant if they have less than
a 5 percent probability of resulting from chance or random fluctuations.
Although a probability of .05 was used as a minimum indicator of
statistical significance, results for women's salaries at all five
laboratories had probabilities of .001 or less. That is, the observed sex
differences in salaries have, at most, a 1 in 1,000 chance of reflecting
only random fluctuations in salaries. We chose this analytic design
because it (1) is widely used in human capital literature to evaluate
differences in compensation and other employment-related subjects, (2)
allowed for the most straightforward and parallel analysis of the
laboratories' personnel data, and (3) is an appropriate statistical method
for answering our first objective.

The models for salaries, merit pay increases, and separations included
variables commonly used to explain variation in personnel actions, such as
education level, age, and laboratory tenure. In addition, we included
laboratory-specific variables, such as occupational category and
organizational division, to control for the laboratory's unique
classification system for comparing similarly situated individuals.
Including these variables allowed us to determine whether pay differences
between women and men and minorities and Whites existed despite their
equality in position and in other human capital characteristics, such as
tenure and education level. We controlled for these factors because they
are widely used in human capital models, are generally used by the
laboratories in their compensation analysis, and are available from all
six laboratories. We also included model-specific adjustment variables.
For example, we included a salary variable in the merit pay increase model
because merit pay increases are proportional to salary. Additionally,
because we were analyzing 3-1/2 years of data, we included variables to
adjust for nonconstant demographic data in the salary and merit pay
increase models. The variables included in the salary models account for
88, 88, 95, 91, 85, and 96 percent of the variance in salaries at Argonne,
Brookhaven, Idaho, Lawrence Berkeley, Oak Ridge, and Pacific Northwest,
respectively. The variables included in the merit pay increase models
account for 76, 74, 78, 69, 75, and 63 percent of the variance in merit
pay increases at Argonne, Brookhaven, Idaho, Lawrence Berkeley, Oak Ridge,
and Pacific Northwest, respectively.

Salary and merit pay increases represent an average for the 3-1/2-year
period, adjusting for the length of time employees were on board during
that period. In accordance with economic analysis literature, we used the
natural log of salary and merit pay increases in our models.3 Separations
from the laboratory include voluntary actions, such as retirements and
resignations, and nonvoluntary actions, such as terminations for cause. If
employees terminated their employment at a laboratory for any reason in
the 3-1/2-year period, they were coded as having separated. Temporary and
term employees were not included in the separation analyses.

We reviewed our final models with the laboratories and performed
additional analyses in response to some of their concerns. In particular,
for the salary analysis of one laboratory, the laboratory considered
performance score an important factor in determining pay and includes that
control in its own analyses of pay. Because we included both temporary
staff-who typically do not receive performance scores-and permanent staff
in our analyses, we did not include performance score. To address the
laboratory's concerns, we analyzed permanent staff only and included a
control for average performance score. No material changes in the estimate
for women's salaries resulted from the inclusion of average performance
score. Additionally, the laboratory considered tenure within grade to be
an important control, so we included this control in the salary analysis
for the same laboratory, which also resulted in no material change to the
estimate for women's salaries. Finally, one laboratory did not have tenure
within grade in their database, but considered it an important factor in
their pay equity studies. Consequently, we included a control indicating
whether someone received a promotion in the 3-1/2-year period in the
salary analysis as a proxy for tenure within grade. Our assumption was
that more recent changes in pay grade should show the greatest effect on
salary within grade, if tenure within grade is an important explanatory
factor in salary differences. There were no material changes in the
estimate of women's salaries when including this control for promotion.
Because there were no material changes in the results of these additional
analyses, we reported the results of our original analyses.

Promotions and the 80 Percent Rule

To determine whether promotions of management and professional women and
minorities reflect the diversity of the potential applicant pool (other
managers and professionals), we applied the 80 percent rule set out in the
federal government's Uniform Guidelines on Employee Selection Procedures.
We did not analyze promotions of nonmanagerial or nonprofessional staff
because the applicant pools were either external to the laboratories or
were from job categories within the laboratories that we did not examine.
The potential applicant pool consists of exempt, permanent laboratory
staff in managerial or professional positions at any time in the
3-1/2-year period.

Using the 80 percent rule, we first determined the proportion of
promotions for each race/ethnicity and sex group based on the number of
promotions received in the 3-1/2-year period and the total number of
laboratory staff in each group. We then determined whether the proportions
for minority men and women and White women represented at least 80 percent
of the proportion for White men. Since there are a limited number of
promotions every year, we examined promotions for the entire 3-1/2-year
period. We did not include postdoctoral and limited-term employees in our
promotion analysis because they do not receive promotions. If a personnel
action denoted either a competitive or career promotion, the
race/ethnicity and sex of the employee receiving the promotion was
recorded. If an individual received more than one promotion in the
3-1/2-year period, the individual was counted as having one promotion. The
number of individuals needed to reach 80 percent of the White male
promotion rate was rounded down. For example, where a minority group was
short of the 80 percent promotion rate by 2.8 people, that group would be
reported as being 2 people short. Unlike the analyses of salaries, merit
pay increases, and separations, we did not control for any factors that
might influence the likelihood of promotion. Additionally, we did not have
data on either who was eligible to compete for a promotion or who actually
applied for a promotion.

Women's and Minorities' Concerns at the Laboratories

To determine what concerns women and minorities raised at the
laboratories, we primarily analyzed the available information contained in
DOE and laboratory surveys and studies of laboratory staff since 2000 and
the results of structured interviews we conducted with representatives of
women's and minority groups at each laboratory. These women's and minority
groups have memberships spanning all position levels at the laboratories,
with the exception of the Argonne's women's group, which consisted solely
of scientific and technical positions. Idaho does not have women's and
minority employee groups, but we did speak to representatives from a
laboratory diversity committee associated with one of the laboratory's
divisions. In addition, although Lawrence Berkeley does not have a women's
group, it selected a small group of women for us to discuss EEO issues at
the laboratory. Finally, Pacific Northwest does not have minority employee
groups, although we did speak with a Black scientist selected for our
review by the laboratory. We included only those EEO staff concerns that
we considered most relevant. We did not attempt to describe all of the EEO
concerns raised or analyze the laboratories' efforts to address these
concerns. We also did not attempt to prove or disprove the validity of
these concerns.

Internal and External Complaints

For our analysis of complaints filed at the laboratories, we developed a
template to capture detailed complaint data on internal and external
complaints from fiscal year 2001 through June 25, 2004. Internal complaint
data include complaints filed and investigated within the laboratory. The
laboratories do not have a uniform policy for recording or managing
complaints. Instead, each laboratory maintains different processes and
thresholds for determining which complaints it will investigate and
record. (Brookhaven did not provide data for internal complaints filed
during the time period we examined.) Consequently, the numbers of internal
complaints are not comparable across the laboratories. Nevertheless,
internal complaint data do provide an indication of the types of Title VII
concerns raised at each laboratory and the general

characteristics of complainants.4 These data are reliable insofar as they
provide an indication of the types of Title VII concerns the laboratories
have experienced. However, because of the inconsistencies across and
within some laboratories in the processes for managing and recording
internal complaints, the data are neither comparable across the
laboratories nor exhaustive of all possible complaints made by laboratory
employees during the period we reviewed.

External complaint data include complaints filed with and investigated by
EEOC, OFCCP, or a state and local fair employment practices agency.5 We
determined that data on external complaints were reliable for our
purposes. However, due to the small number of cases, we did not discuss
specific details of complaints by laboratory in order to prevent the
identification of individual complainants. Because EEOC, OFCCP, and state
and local fair employment practices agencies notify a laboratory's legal
department, human resources, or EEO office when a complaint is filed
against the laboratory, we are assured that the data we received from the
laboratories contain the full universe of external complaints against them
and that the data are reliable for comparison across the laboratories.

DOE and OFCCP's Actions to Implement Our 2002 Recommendation

To determine the actions DOE and OFCCP have taken to implement our 2002
recommendation, we met with responsible DOE and OFCCP officials. We also
examined legislation, Executive Order 11246, as amended, OFCCP
regulations, the Federal Acquisition Regulation, and DOE directives on the
roles and responsibilities of DOE headquarters and field offices, OFCCP,
EEOC, and the contractors in ensuring that the laboratories comply with
EEO requirements.

Method for Developing Background Information in Appendix II

To provide descriptive information about laboratory staff, in terms of
race/ethnicity, sex, and job category, we obtained data from the
laboratories on the number of staff by race/ethnicity, sex, and job
category in 2003, as reported annually by the laboratories to EEOC on the
Employer Information Reports (EEO-1s). For ease of presentation, we
grouped the laboratory jobs into three categories: managers and
professionals, which constitute the majority of staff at each of the
laboratories; technicians, clerks, and craft workers; and operatives,
laborers, and service workers. These categories constitute eight of the
nine job categories required for the EEO-1s: officials and managers,
professionals, technicians, office and clerical, craft workers,
operatives, laborers, and service workers. The laboratories do not have
sales workers, which is the ninth category; therefore, sales workers were
not part of our analysis. We determined that the EEO-1 data were
sufficiently reliable for our purposes.

We conducted our review from February 2004 through December 2004 in
accordance with generally accepted government auditing standards.

Information on the Six Multiprogram LaboratoriesAppendix II

Contractors operate the six multiprogram laboratories we reviewed. These
laboratories have a total workforce of about 21,000 and vary in size from
fewer than 2,500 employees at Lawrence Berkeley to more than 5,000 at
Idaho. Each laboratory works in several research areas, which may include
basic science, medical research, information technology, environmental
restoration, national security, weapons nonproliferation, and nuclear
safety. In addition, they provide advanced scientific facilities that are
available for use by guest scientists from industry, academia, other
laboratories, and other countries. Figure 5 identifies the six
laboratories and provides some profile information on them.

Figure 5: Profile Information on the Six Laboratories Reviewed

aOn February 1, 2005, Battelle Energy Alliance, LLC, assumed management of
operations at Idaho.

This appendix also provides a brief description of each of the six
laboratories and describes the staff composition by (1) sex; (2)
race/ethnicity; (3) White men, White women, and minorities; (4) job
category; and (5) job category by White men, White women, and minorities.
We obtained this information from the EEO-1s the laboratories submitted in
2003.

Argonne National Laboratory

Argonne National Laboratory reports to DOE's Office of Science through the
Argonne Site Office. Argonne originated from a University of Chicago
laboratory that participated in the World War II effort to develop nuclear
weapons and that was the location of the first controlled nuclear chain
reaction. The laboratory was chartered as a national laboratory in 1946.
Managed and operated by the University of Chicago, it occupied two
sites-Argonne, Illinois, and a location 50 miles west of Idaho Falls,
Idaho until February 1, 2005, when the Idaho location-Argonne West-became
part of the newly established Idaho National Laboratory. Research at
Argonne falls into five broad categories: (1) basic science in the fields
of chemistry, biology, physics, mathematics, and computer science; (2)
scientific facilities, including the Advanced Photon Source, for
laboratory and visiting scientists; (3) development of energy sources for
the future; (4) environmental management; and (5) national security in
support of counterterrorism and the detection of weapons proliferation. In
2003, Argonne had 3,420 contractor employees. Figures 6 through 10 profile
Argonne staff.

Figure 6: Argonne Staff by Sex, 2003

Figure 7: Argonne Staff by Race/Ethnicity, 2003

Note: American Indians comprise less than 0.5 percent of Argonne's staff.

Figure 8: Argonne Staff by White Men, White Women, and Minorities, 2003

Figure 9: Argonne Staff by Job Category, 2003

Figure 10: Composition of Job Category Group at Argonne by White Men,
White Women, and Minorities, 2003

Brookhaven National Laboratory

Brookhaven National Laboratory, located in Upton, New York, reports to
DOE's Office of Science through DOE's Brookhaven Site Office. Brookhaven
was established in 1947 to promote basic research in the physical,
chemical, biological, and engineering aspects of the atomic sciences and
to construct and operate large scientific machines that individual
institutions could not afford to develop on their own. Brookhaven is
managed and operated by Brookhaven Science Associates. The laboratory's
major programs include (1) nuclear and high-energy physics, (2) physics
and chemistry of materials, (3) environmental and energy research, (4)
counterterrorism and weapons nonproliferation, (5) neurosciences and
medical imaging, and (6) structural biology. In 2003, Brookhaven had 2,839
contractor employees. Figures 11 through 15 profile Brookhaven staff.

Figure 11: Brookhaven Staff by Sex, 2003

Figure 12: Brookhaven Staff by Race/Ethnicity, 2003

Note: American Indians comprise less than 0.5 percent of Brookhaven's
staff.

Figure 13: Brookhaven Staff by White Men, White Women, and Minorities,
2003

Figure 14: Brookhaven Staff by Job Category, 2003

Figure 15: Composition of Job Category Group at Brookhaven by White Men,
White Women, and Minorities, 2003

Idaho National Engineering and Environmental Laboratory

Idaho National Engineering and Environmental Laboratory, which became part
of the newly established Idaho National Laboratory on February 1, 2005, is
an 890-square-mile section of southeast Idaho, with offices in Idaho
Falls. Idaho reports to DOE's Office of Nuclear Energy, Science, and
Technology through DOE's Idaho Operations Office. Established in 1949 as
the National Reactor Testing Station, the laboratory's initial mission was
to develop civilian and defense nuclear reactor technologies and to manage
spent fuel. Until February 1, 2005, when Battelle Energy Alliance, LLC
assumed management of operations at the laboratory, Idaho was managed and
operated by Bechtel BWXT Idaho, LLC, which consisted of Bechtel National,
Inc., BWX Technologies Company, and the Inland Northwest Research
Alliance, a consortium of eight regional universities. The laboratory's
primary missions include (1) environmental management-environmental
restoration of the site, waste management, disposition of spent nuclear
fuel, and high-level waste management; (2) energy programs-nuclear and
radiological research and nuclear reactor design and development, fossil
energy, energy efficiency, and renewable energy; (3) nonproliferation and
national security; and (4) scientific research. In 2003, Idaho had 5,075
contractor employees. Figures 16 through 20 profile the Idaho staff.

Figure 16: Idaho Staff by Sex, 2003

Figure 20: Composition of Job Category Group at Idaho by White Men, White
Women, and Minorities, 2003

Lawrence Berkeley National Laboratory

Lawrence Berkeley National Laboratory, located in Berkeley, California,
reports to DOE's Office of Science through DOE's Chicago Operations
Office. The laboratory was established at the university in 1931 by Ernest
Orlando Lawrence to advance scientific research through the development
and application of the cyclotron, an instrument that accelerates charged
atoms at high speed, to accomplish nuclear transmutations. It became a
federal facility in 1942. The laboratory is managed and operated by the
University of California and conducts research in the areas of (1)
biological, physical, and chemical sciences; (2) energy resources; (3)
computing sciences; (4) material sciences; and (5) environmental
remediation. In 2003, Lawrence Berkeley had 2,397 contractor employees.
Figures 21 through 25 profile Lawrence Berkeley staff.

Figure 21: Lawrence Berkeley Staff by Sex, 2003

Figure 22: Lawrence Berkeley Staff by Race/Ethnicity, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 23: Lawrence Berkeley Staff by White Men, White Women, and
Minorities, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 24: Lawrence Berkeley Staff by Job Category, 2003

Figure 25: Composition of Job Category Group at Lawrence Berkeley by White
Men, White Women, and Minorities, 2003

Oak Ridge National Laboratory

Oak Ridge National Laboratory, located in Oak Ridge, Tennessee, reports to
DOE's Office of Science through DOE's Oak Ridge Operations Office. The
laboratory was established in 1943 to develop a method for producing and
separating plutonium as part of the World War II effort to develop nuclear
weapons. The laboratory is managed and operated by a partnership of the
University of Tennessee and Battelle. Its primary missions include (1)
research and development of advanced materials; (2) biological and
environmental sciences and technology; (3) computational science and
advanced computing; (4) energy production and end-use technologies; (5)
instrumentation and measurement, such as biological and chemical detection
and measurement; and (6) neutron science and technology-using neutrons to
study the structure and dynamics of materials. In 2003, Oak Ridge National
Laboratory had 3,762 contractor employees. Figures 26 through 30 profile
Oak Ridge staff.

Figure 26: Oak Ridge Staff by Sex, 2003

Figure 27: Oak Ridge Staff by Race/Ethnicity, 2003

Notes: Sections do not total to 100 percent due to rounding. American
Indians comprise less than 0.5 percent of Oak Ridge's staff.

Figure 28: Oak Ridge Staff by White Men, White Women, and Minorities, 2003

Figure 29: Oak Ridge Staff by Job Category, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 30: Composition of Job Category Group at Oak Ridge by White Men,
White Women, and Minorities, 2003

Pacific Northwest National Laboratory

Pacific Northwest National Laboratory, located in Richland, Washington,
reports to DOE's Office of Science through the Pacific Northwest Site
Office. The laboratory was established in 1965 to perform research and
development for DOE's Hanford site, a World War II and Cold War plutonium
production facility. The laboratory's early missions included protecting
the environment, fabricating reactor fuel, and designing reactors.
Battelle manages and operates the laboratory. The laboratory's primary
missions include (1) ensuring efficient and productive uses of energy; (2)
environmental research, such as developing indicators of human and
ecosystem health; (3) physical, chemical, and biological science; (4)
protecting and improving workers' safety and health; (5) developing
software and hardware for scientific research and business systems; and
(6) supporting the nation's national security effort. In 2003 Pacific
Northwest had 3,380 contractor employees. Figures 31 through 35 profile
Pacific Northwest staff.

Figure 31: Pacific Northwest Staff by Sex, 2003

Figure 32: Pacific Northwest Staff by Race/Ethnicity, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 33: Pacific Northwest Staff by White Men, White Women, and
Minorities, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 34: Pacific Northwest Staff by Job Category, 2003

Note: Sections do not total to 100 percent due to rounding.

Figure 35: Composition of Job Category Group at Pacific Northwest by White
Men, White Women, and Minorities, 2003

Comments from the Department of EnergyAppendix III

Comments from the Department of LaborAppendix IV

GAO Contacts and Staff AcknowledgmentsAppendix V

GAO Contacts

Robin M. Nazzaro, (202) 512-3841 ([email protected]) Sherry McDonald, (202)
512-8302 ([email protected])

Staff Acknowledgments

In addition to the individuals named above, Doreen S. Feldman, Kerry
Hawranek, Dick Kasdan, Carol Kolarik, Grant Mallie, Rebecca Shea, Carol
Herrnstadt Shulman, Lisa Vojta, and Greg Wilmoth made key contributions to
this report.

(360426)
*** End of document. ***