Diversity In Law Firms

The U.S. Equal Employment Opportunity Commission

---

Executive Summary

Since 1975, the representation of women, African Americans, Hispanics and Asian Americans as professionals in larger Legal Service firms has increased substantially.

• Women increased from 14.4 percent in 1975 to 40.3 percent in 2002.
• African Americans from 2.3 percent to 4.4 percent.
• Hispanics from 0.7 percent to 2.9 percent.
• Asians from 0.5 percent to 5.3 percent.

There were parallel increases in J.D. degrees from 1982 to 2002.

• Women receiving law degrees increased from 33 percent in 1982 to 48.3 in 2002.
• African Americans from 4.2 percent to 7.2 percent.
• Hispanics from 2.3 percent to 5.7 percent.
• Asians from 1.3 percent to 6.5 percent.

Firm characteristics such as size, number of offices, locations, prestige and earnings rankings appear to have more effect on the proportion of minority legal professionals than the proportion of women legal professionals. However, both the proportion of women and the proportion of minorities are significantly higher in firms with more offices.

Minority legal professionals are likely to be associated with firms in the top ten legal markets (cities), and in firms ranked in the top 100 on the basis of prestige and/or earnings.

Large, nationally known law firms generally have a higher proportion of women and minorities than other types of law firms. There is also less variation in the proportion of women and minorities among these large, nationally known law firms.

In comparing associates and partners in a sample of large law firms, women, African Americans, Hispanics and Asians all have lower odds of being partners than White males.

Table of Contents

Introduction

Changes in the Employment of Women and Minorities

Law Firm Organizational Characteristics

Status Within the Firm: Partners and Associates

Additional Information

Methodological Appendix: Comparison of EEO-1 and NALP Surveys
Methodological Appendix

Selected References

Introduction

Purpose

The purpose of this report is to examine the employment status of women and minorities at law firms required to file EEO-1 reports. An employer is required to file an EEO-1 report if it employs 100 or more employees. Therefore, this study covers law firms which would be characterized as medium to large. Specifically, it examines employment status in a general sense to display the changes in the employment of minorities and women as attorneys since 1975. It also looks at the organizational characteristics of firms to explore the variations in the current employment of minorities and women. Finally, a major issue in law firms, the prospect of becoming a partner, is examined empirically to determine the relative likelihood of women and minorities being partners.

The Legal Profession

The importance of the legal profession in today's society is unquestionable. Lawyers are often powerful players in social, economic and political circles and as women and minorities become an increasing part of this profession, their ability to obtain public and private influence is increasing.¹

[L]awyers are very often key players in designing and activating the institutional mechanisms through which property is transferred, economic exchange is planned and enforced, injuries are compensated, crime is punished, marriages are dissolved and disputes are resolved. The ideologies and incentives of the lawyers engaged in these functions directly influence the lived experience of Americans, including whether they feel fairly treated by legal institutions (p. 346).²

However, perhaps more important than the influence of attorneys is the central role they play in maintaining social stability.

The persuasive power of law as a tool to change or eliminate certain or nonproductive behavior must, in part, be attributable to the respect and acquiescence afforded to the law and lawyers by those subject to it. . . . Hence, the development of law and its practice as a noble profession rather than as a trade or occupation (p. 1022).³

More specifically,

Patterns of stratification with the legal profession are important in their own right . . . but they are of particular concern to legal scholars and legal educators because principles of inequality among lawyers may suggest much about whether access to justice in our society is fairly distributed. If race, gender, and social class are determinants for entry into the profession and for the attainment of certain positions within the profession, it may imply that these same attributes affect the sorts of treatment individuals will receive by legal institutions, in part because they do not have access to lawyers who share a similar social background (Nelson, 1988, p. 368).⁴

Social scientists have researched many aspects of American law firms including size, geographic location, hiring and promotion patterns, legal specialties, profitability, and client characteristics. Several themes emerge from this literature.

Public Sector Employment

Many studies find that women and minorities are likely to hold jobs in the public sector. For example, Payne and Nelson (2003), in a study of the Chicago bar as of 1995, report that 20.7 percent of white women lawyers were employed by government or the judiciary, compared to 7.6 percent of white men. The percentages for African-American lawyers and Hispanic lawyers in government and the judiciary are even higher, 43.8 percent and 37.5 percent respectively. (See their Table 2-2).⁵

Private Sector Employment

Almost all studies find a substantial increase in the employment of women and minorities in private sector law firms. For example, in a study of ninety-seven elite law firms in Chicago, Los Angles, New York, and Washington, Elizabeth Chambliss (1997) states that " . . . the lawyers who work in elite law firms historically have been white Protestant men who graduated from prestigious law schools such as Harvard, Columbia, and Yale. As recently as 1970, women and people of color were almost completely excluded. Since 1970, the gender and race composition of elite law firms has changed considerably at the associate level. By 1980, 23.2% of the associates in the sample were women; by 1990, 36.2% of associates in the sample were women. Although the level of racial diversity is much lower, it too has increased. By 1980, 3.6% of associates in the sample were minorities; by 1990, 6.5% of associates were minorities" (pp.695-696).⁶

Information on Minorities

As a general rule, the available literature tends to focus more on women than minorities in the legal profession.⁷

Increasing Focus on Mechanisms

Although many of earlier studies concentrated on broad questions about the distribution of women and minorities across different sectors of the legal profession, recent studies are increasingly examining employment practices in large private law firms. Examples follow.

Attrition⁸

The 2003 NALP Foundation Study of entry-level hiring and attrition⁹ concluded that,

Compared to men as a whole, male minority associates were more likely to have departed their employers within 28 months (29.6 percent vs. 21.6% of men overall) and were far more likely to have departed within 55 months of their start date (68 percent) minority males departed vs. 52.3 percent of men overall . . . Female minority associates departed their law firm employers at somewhat greater rates than women as a whole, with the differential widening as the years in the job increased. Nearly two-thirds (64.4 percent) of female minority associates had departed their employers within 55 months compared to just over half (54.9 percent) of women overall (p. 23).¹⁰

Earnings

An examination of pay differences among University of Michigan Law School graduates by Noonan, Corcoran, and Courant (2003)

. . . compared male/female differences in earnings 15 years after graduation for two cohorts: (1) men and women who graduated from law school between 1972 and 1978, and (2) men and women who graduated from law school between 1979 and 1985. We find that the gender gap in earnings has remained relatively constant; 15 years after graduation, women in both cohorts earn approximately 60% of men's earnings. Penalties to part-time work and career interruptions¹¹ also remain steady. While within occupation sex segregation has declined over time, sex differences in hours worked have increased and assume a more prominent role in explaining the sex gap in lawyers' earnings (p. 1).¹²

Promotion

A study of eight large New York corporate law firms describes the traditional "up and out" system of promotions to partner as follows:

Women have fared poorly under the 'up and out' system. Using data supplied by the firms and the Martindale-Hubbell Law Directory, we tracked cohorts of first-year associates in the eight firms in periods beginning in 1973-74 and 1985-86 for a ten-year period to see how many associates had been elevated to partner. (The last cohort, those hired in 1985-86, were followed until 1994) . . . For each cohort except the first, where one-quarter of women associates (five of twenty) made partner, men associates gained partnership at a higher rate than women. For the entire period, 19% (362 of 1878) of men attainted partnership while only 8% (60 of 754) of women made partner" (p. 358).¹³

Research Methodologies

Most studies of legal employment have relied on public data sources or individual interviews with attorneys. With several notable exceptions (e.g., the continuing studies of the Chicago bar¹⁴ ), there have been relatively few systematic, large-scale sampling studies of American lawyers. Perhaps the most promising future development is the work currently being done by the After the JD (AJD) study. One of the strengths of the AJD study is the broad range of organizations supporting the project. In addition to the National Science Foundation, the AJD project obtained funding from a number of organizations interested in legal education and the profession, including Access Group, American Bar Foundation, Law School Admission Council, NALP and NALP Foundation, National Conference of Bar Examiners, and the Open Society Institute. The AJD project is based on a two-stage, scientific sampling design that first selects among geographic areas and then selects individual attorneys within those areas. The sample population consists of persons who first became members of a state bar in calendar year 2000 and who graduated from law school in the period July 1, 1998 through June 30, 2000. Approximately 9200 individuals received an initial questionnaire sent in March 2002, and a sample of these - about 20% -- are currently being interviewed face-to-face. Respondents will be re-contacted five and ten years after their admission to the bar. The forthcoming results of the AJD project will provide a rich and unparalleled source of data on attorney careers including first job after law school and all subsequent jobs as well as detailed descriptions of the current job such as partnership status, hours worked, and time devoted to different legal specialties.

The next section will examine trends in the legal profession over time. Various data sources are considered. The Current Population Survey covers attorneys, data from the American Bar Association covers prospective attorneys as reflected in law degrees conferred and the EEO-1 covers a range of legal professionals that are predominately attorneys but other job titles (such as non-lawyer accountants) as well. Despite the diversity of data sources, most of the trends suggest a uniform pattern of increased growth in the participation of women and minorities in the American legal profession..

Changes in the Employment of Women and Minorities

Three different data sources are used to examine how the employment of women and minorities in the legal professional has changed over time. These three data sets provide different perspectives on the employment of attorneys. The Equal Employment Opportunity Commission's (EEOC's) own EEO-1 report is used to reflect employment in large private law firms. Private employers with 100 or more employees are required to file annual EEO-1 reports with EEOC. They are also required to file separate reports for each of their establishments with 50 or more employees. By and large when companies in the Legal Services industry file such reports the professional job group provides a fairly representative index of diversity among associate attorneys.¹⁵ Due to the filing threshold of 100 employees, the EEO-1 data best captures the employment practices of large private firms. Of course, not all lawyers are employed by these types of organizations. Therefore, a second data set, the Current Population Survey¹⁶ was used to obtain a perspective on the more general labor market for attorneys. Finally, to obtain a sense of the availability of women and minority attorneys, data on law degrees (J.D. degrees) conferred is examined.¹⁷ Two different time periods are examined. For EEO-1 data it is possible to construct a relatively long time period from 1975 to 2002. Due to the limitations of the other data sets, the period from 1982 to 2002 is examined when using all three types of data.

FIGURE 1: EMPLOYMENT OF WOMEN EEO-1, 1975-2002

Image

Women

In 1975 women represented just 14.4 percent of all professionals in the legal services industry based on their filing of EEO-1 reports. By 2002, this figure increases dramatically to 40.3 percent. See Figure 1. It is interesting to compare these results to the percent of women receiving law degrees and the percent of women lawyers in the entire workforce as reflected in the Current Population Survey. See Figure 2.

Figure 2: Comparisons of
Degrees Conferred, EEO-1 Employment and
Current Population Survey Data
for Women

Image

In 1982 the percent of women reported as professionals in Legal Services on the EEO-1 is nearly identical to the percent of women receiving law degrees in that year. However, by 2002 the employment of women as professionals in these larger law firms is eight percentage points below degrees conferred. Employment of women lawyers reported in the Current Population Survey falls behind both the employment of women professionals in legal services as reported on the EEO-1 and law degrees obtained by women. Rates of change were computed for women over this time period in order to obtain a better sense of the relative differences over time. (Because the raw numbers in the three data sources differ in magnitude, the percentages are used to compute these rates.) The percent of women professionals in legal services on the EEO-1 increased by 23.6 percent during the period, while the rate of change for J.D. degrees conferred was 46.4 percent. CPS employment of women attorneys exhibit a rate of change of 112 percent over the period. This suggests the employment of women in the larger law firms required to file EEO-1 reports may not have kept pace with law degrees obtained by women or the employment of women attorneys in the general work force. Despite this, the employment of women in these firms remained higher than in the more general work force.

Minorities

African Americans represented 2.3 percent of these employees in 1975 and 4.4 percent in 2002. However, the percent of Asian professionals in Legal Services reported on the EEO-1 exceeds African American professionals by 2002. Starting at just 0.5 percent in 1975, Asians represent 5.3 percent in 2002. Hispanics increased from 0.7 percent to just less than 3 percent. Native American Alaskan Natives are poorly represented among these workers. See Figure 3.

Figure 3: EEO-1 Employment by Race/Ethnicity 1975-2002

Image

African Americans

As Figure 4 shows, law degrees earned by African Americans appear to consistently exceed the employment of African Americans as professionals in Legal Services in large private law firms (EEO-1 data) and as lawyers in the general work force (CPS data).

Figure 4: Comparisons of Degrees Conferred, EEO-1 Employment and
Current Population Survey Data for African Americans

Image

Further, unlike the employment patterns for women, the proportion of African Americans employed as lawyers in the general labor market and as professionals in law firms as captured by the EEO-1 data is fairly consistent. However, there is a slight difference in the manner in which these rates have changed over time. At the beginning of the period African Americans make up 2.9 percent of professional employment in the EEO-1 reports filed by Legal Service firms and climb to 4.4 percent in 2002. In the general work force figures captured by the CPS, African Americans start lower in 1982 at 2.7 percent and at the end of the period reaches 4.6 percent which slightly exceeds the EEO-1 figure. Rates of change based on these percentages reflect the same dynamics but produce much larger disparities in rates of change. From 1982 to 2002 the African American percentage of EEO-1 reported professionals in legal services increased 51.7 percent and employment of African American attorneys in the general work force increased at a rate of 70.4 percent. The increase in EEO-1 employment of African Americans as professionals in Legal Services did not keep pace with the change in law degrees earned by African Americans(a rate of change of 71.4 percent), but the CPS based rate of change and degrees conferred is similar. Thus, changes in the employment of African American professionals in private sector firms required to file EEO-1 reports lagged behind their increase as lawyers in the general work force and in their increased rate of receiving law degrees over the past twenty years.

Hispanics

In 1982 Hispanics were earning law degrees at a rate (2.3 percent) exceeding their representation as professionals in Legal Services and as attorneys in the general work force. By the end of the period this disparity continues. However, the more interesting change for Hispanics over the last twenty years is their slow but steady growth in the large law firms required to file EEO-1 reports. Although still a relatively small portion of professionals at 2.9 percent, the rate of change over the period was high at 163 percent. This exceeded their growth in obtaining degrees, 148 percent and was much larger than their growth as attorneys in the general work force of 72 percent. See Figure 5.

Figure 5: Comparisons of Degrees Conferred, EEO-1 Employment and
Current Population Survey Data for Hispanics

Image

Asians

As reported above, the growth in Asian attorneys is so rapid that by 2002, the percentage of Asian professionals in Legal Services, 5.3 percent, as reported on the EEO-1 exceeds the percentage of African Americans, 4.4 percent. (Current Population Survey data on the general work force is not available for Asians.) Degrees conferred to Asians also increases during the twenty year study period. In 1982 just 1.3 percent of all law degrees are awarded to Asians but by 2002, they earn 6.5 percent of all degrees. See Figure 6.

Figure 6: Comparisons of Degrees Conferred and EEO-1 Employment Data for Asians

Image

Over the past twenty years the rate of change for the percent of Asians reported as professional by Legal Service firms on their EEO-1 reports is 341 percent. The increase in law degrees earned by Asians is even higher at 400 percent.

Native Americans

Over the past twenty years, the proportion of Native Americans receiving law degrees and reported as professionals by Legal Service firms on their EEO-1 report has increased but still remains less than one percent. By 2002, Native American represented 0.2 percent of the relevant professional work force and 0.7 percent of law degrees conferred.

Summary

Since 1975, the representation of women, African Americans, Hispanics and Asian Americans as professionals in the larger Legal Service firms that are required to file EEO-1 reports has increased substantially. Even greater has been the increase in law degrees earned by minorities. Paralleling the finding regarding increased employment among large law firms are increases in the employment of these groups as attorneys in the general work force.

Law Firm Organizational Characteristics

This section examines some issues related to firm size and geographic location that can be studied with the 2002 EEO-1 data on professionals in private legal service firms. The basic research problem can be posed as follows. There has been a substantial increase in the size of law firms over the last thirty years. As summarized by Chiu and Leicht (1999), "Law firms grew dramatically in size and number. In the early 1960s, there were 38 firms with 50 or more lawyers … In 1991, there were 751 firms with more than 50 lawyers and 13% of all lawyers were employed in firms with at least 50 lawyers, up from 5% in 1980" (p.569).¹⁸ Given this growth in the demand for new lawyers and the increased number of women and minorities graduated from law schools, many observers predicted that larger law firms should have a higher proportion of women and minorities than smaller and medium sized law firms. Yet the empirical results are somewhat mixed.

Among lawyers in private practice, the Payne and Nelson study (2003, Table 2-2) of the Chicago bar in 1995 finds a higher percentage of white women attorneys than African American attorneys working for firms with 100 or more lawyers (44.9 percent and 28.0 percent respectively).¹⁹ For the 1990s cohort of Michigan Law School graduates, Lempert, Chambers, and Adams (2000, Tables 11 and 14) find significant differences in the proportion of white and minority alumni taking first jobs in firms with 151 or more lawyers (55.9 percent and 35.7 percent respectively), but no significant differences between white and minority alumni in holding current jobs in firms with 151 or more lawyers (37.9 percent and 31.0 percent respectively). Chiu and Leicht (1999, p. 569) report that in " … Chiu's (1996) analysis of the 1990 National Survey of Lawyers' Career Satisfaction, women were more likely to work in large law firms than men, but this difference was not statistically significant once years since graduation was controlled." The Chambliss study of elite law firms (1997, Tables 11a and 11b) finds a statistically significant negative relationship between firm size the proportion of female partners, but no statistically significant relationship between firm size and the proportion of female associates.

The 2002 NALP summary data on Women and Attorneys of Color at Law Firms²⁰ suggests that law firm size is more strongly related to the percentage of minority associates than the percentage of women associates. The overall percentage of minority associates in the 2002 NALP survey is 14.3 percent which varies from 10.1 percent in law firms with less then 100 attorneys to 16.9 percent in law firms with 501 or more attorneys. The overall percentage of women associates is 42.4 percent which varies from 40.6 percent in firms with less than 100 attorneys to 43.4 percent in firms with 251 to 500 attorneys.

These relationships are examined in greater detail using the 2002 EEO-1 data.²¹ The following organizational characteristics of law firms are likely to be important factors and are used in this analysis.

1. Total Number of Offices Per Law Firm in the 2002 EEO-1 Survey
2. Total Employment of EEO-1 Legal Professionals in All Offices²²
3. Geographic dispersion as measured by the Total Number of U.S. Census Divisions where the the Firm has Offices²³
4. Firm Cited in Top 100 Ranking of Law Firms by Either Prestige or Partner Profits²⁴
5. Proportion of Total Legal Professions in the Top Ten Legal Markets²⁵
6. Presence of at Least One Firm Office in a U.S. Southern State²⁶

The first three explanatory variables, number of law offices per firm, total number of legal professionals, and total number of census divisions covered by the firm, are highly interrelated.²⁷ To simplify the analysis, we treated the number of offices per firm as a surrogate measure of firm size and created three sub-categories: firms with a single office (319 out of 553 or 57.7 percent), firms with two or three offices (142 out of 553 or 25.7 percent), and firms with four or more offices (92 out of 553 or 16.6 percent). The firms with a single office average 86.9 legal professionals, firms with two or three offices average 204.7 legal professionals, and firms with four or more offices average 454.7 legal professionals.

The Table 1 shows the relationship between the number of offices per firm and other firm characteristics.²⁸ Because the focus on this analysis is on firm characteristics, average or mean of firm percentages are used to better reflect individual firm characteristics.

Table 1: Relationship Between the Number of Offices per Firm and Other Explanatory Variables

The number of law offices per firm is closely associated with the other organizational characteristics. The greater the number of law offices per firm, the greater the proportion of firms ranked among the top one-hundred law firms and the greater the proportion of legal employees located in the top ten legal markets.²⁹ The average percentage of firms ranked in the top one-hundred by prestige and earnings increases from five percent in law firms with a single office to 48.9 percent in firms with four or more offices. Likewise, the average percent of legal employees in the top ten legal markets increases from 37.9 percent in law firms with a single office to 59.3 percent in firms with four or more offices. It also should be noted that law firms with four or more offices are highly likely to have at least one office in a southern state. Essentially it appears that the firms with four or more offices represent relatively large, national (and in some cases, international) law firms, many of whom are well-known and highly regarded within their field. The firms with fewer offices are likely to be smaller, regional firms that serve a more limited client base.

To examine the diversity characteristics of different types of law firms, we analyzed the following proportions:³⁰

• Proportion of Women Legal Professionals in All Offices of a Law Firm
• Proportion of Minority Legal Professionals in All Offices of a Law Firm

Table 2 summarizes both the total percentage and the average percentage of women and minority legal professionals by type of law firm.

Table 2: Professionals by Type of Firm

Total percentages combine all employees in all firms together. The average of percentages is the mean percentage of minorities and women at each individual firm. While these numbers are very similar, the latter is more appropriate for capturing firm behavior. In Table 2, the average percentages for women and for minorities are larger in law firms with four or more offices than in law firms with a single office, but these proportional differences for minorities shows greater statistical significance than the proportional differences for women.³¹ The average percentage of women increases from 37.7 in law firms with a single office to 41.6 in law firms with four or more offices. The probability of observing overall differences in the average percentage of women is about two chances out of one hundred (0.0169).³² Likewise, the average percentage of minorities increases from 9.1 percent in law firms with a single office to 0.13.2 percent in law firms with four or more offices. The probability of observing overall differences in the average percentage of women is about one chance out of ten thousand (33 Thus, minority legal professionals are more likely to be found in the largest law firms and the average percentage of minorities tend to vary more by the type of law firm (measured by number of offices) than does the percentage of women.³⁴

The employment of minorities is more varied across firms than the employment of women even when controlling for number of firm offices.³⁵ This suggests that minority legal professionals might be concentrated in certain firms. Perhaps equally important, these variations for both women and minorities decline when the number of law firm offices increase. The percentages of women and minorities tend to be somewhat more uniform among those firms with four or more offices while the percentages of minorities and women changes from firm to firm when those firms have just one office.³⁶

The relationships between the proportion of minority and women legal professionals in a firm and firm characteristics are examined in greater detail. Regression analyses are computed separately for law firms with one office, two or three offices, and four or more offices. Given the importance of number of offices, this is done to gain insights into what characteristics influence the proportion of minorities and women when number of offices is held constant. Table 3 summarizes the regression results for women.³⁷

Firm characteristics used as explanatory variables appear to have relatively little effect on the proportion of women legal professionals within the different types of law firms. There are no statistically significant relationships,³⁸ for either law firms with two or three offices or law firms with four or more offices. The only statistically significant relationship is represented by a single explanatory variable for law firms with one office. In law firms with one office, there is some indication that women legal professionals are less likely to be found in law firms located in the Southern Census Region (a standardized parameter value of -0.134 and a T probability value of 0.018).

Table 3: Regression Results for Women

The corresponding regression results for minorities are summarized in Table 4.
In contrast to the regression results for women legal professionals, firm characteristics appear to have a more substantial effect on the proportion of minority legal professionals. In law firms with a single office, there is a strong likelihood that minority legal professionals will be associated with law firms concentrated in the top ten legal markets (a standardized parameter value of 0.327 and a T probability value of

Table 4: Regression Results for Minorities

In law firms with two or more offices, there are strong effects from both location in a top ten legal market and a top 100 ranking for either prestige or profits per partner. Minority legal professionals are more likely to be located in law firms with a large proportion of professional employees in the top ten legal markets (a standardized parameter value of 0.230 and a T probability value of 0.013). They are also more likely to be associated with law firms ranked in the top100 law firms by prestige or profits per partner (a standardized parameter value of 0.252 and a T probably value of 0.012).

The ability of firm characteristics to explain the proportion of minority legal professionals is much weaker for firms with four or more offices. Taken as a whole, the examined firm characteristics has limited value in explaining or predicting the proportion of minorities (probability of 0.040). Further, the strongest single effect, proportion of legal employees in the top ten legal markets, is not significantly different from having no effect.

In essence, then, minorities are more likely to be employed in firms with more offices. Their employment increases in firms with two or three offices when these firms exhibit characteristics associated with firms with more offices, location in large markets, and top 100 rankings. For single office firms, location in a large market increases minority employment.

Generally speaking, firm characteristics used in this study appear to have more effect on the proportion of minority legal professionals than the proportion of women legal professionals. In addition, it should be noted that the relative size of a law firm, that is, the total number of legal professionals, has little or no effect on diversity proportions within the different types of law firms as defined by the number of offices. Perhaps the most important result concerns the absence of explained variations in diversity proportions among law firms with four or more offices. The large, nationally known law firms generally have a higher proportion of women and minorities than other types of law firms, and there appears to be substantial amount of uniformity within this group, at least for the variables measured in this study.³⁹

Summary

An examination of the 2002 EEO-1 data on legal professionals in private law firms has several broad implications for civil rights enforcement. In large, national law firms, the most pressing issues have probably shifted from hiring and initial access to problems concerning the terms and conditions of employment, especially promotion to partnership. In smaller, regional and local law firms, questions about the fairness and openness of hiring practices probably still remain, particularly for minority lawyers. The next section will examine some general characteristics of promotion patterns in large private law firms.

Status Within the Firm: Partners and Associates

A major issue in law firms generally is the movement from an associate attorney to a partner. Beckman and Phillips explain,

. . . promotion to partner not only involves the greatest increase in income within the law firm, but the partnership includes membership to a professional elite with access to substantial social and political capital (Nelson 1988). More generally, partners of large corporate law firms are among the elite class in the U.S. (Mills 1956; Smigel 1969; Domhoff 1998). Given the power and influence that accompanies large law firm partnership, women's [and minorities] attainment within law firms has larger societal ramifications for access and opportunities (Hagan and Kay 1995, p. 6).⁴⁰

This promotion takes on special meaning for women and minorities since the decision is often viewed as being subjective and thus subject to non-relevant factors such as race/ethnicity or gender⁴¹ (p. 521). Dixon and Seron describe decision making in law firms, "Partners in firms typically rely on centralized informal collective decision making through consensus of the partners rather than decentralized, formal rule-bound decision making via bureaucratic processes"(p. 389).⁴² The study, Perceptions of Partnership: The Allure and Accessibility of the Brass Ring found gender and race/ethnic disparities in the perception regarding equity in the "opportunity for advancement to partnership". With respect to the partnership decision, when examining survey responses from associates in firms with more than 100 employees, 74.3 percent of male associates felt opportunities were "equally available to all," but only 50.0 percent of women associates felt that way. Similarly, 70.2 percent of non-minority associates in these type of firms felt opportunities for partnership were equal but only 30.8 percent of minority associates have the same perception.⁴³

There are few data sources that allow the comparison of associates and partners. Such detailed data is not collected on the EEO-1 or relevant Current Population Survey or Census data. The National Association of Legal Placement (NALP), however, conducts an annual survey of law firms to obtain information about their work forces which includes race/ethnic and gender data by partner and associate job categories. (A copy of the form used to collect this information can be found in Appendix A.) This published data⁴⁴ was utilized to examine the relationship between associates and partners.

In order to analyze the NALP data it was necessary to enter the data manually from the hard copy directory. Therefore, a sample of firms was utilized. The EEO-1 data base was used to select the sample firms. The sample was drawn from a population of all headquarter facilities in the Legal Service industry filing EEO-1 reports in 2002. There were 1,231 total establishments but just 508 headquarters reports from which the sample could be drawn. Seeking at least a ten percent sample and recognizing that not all firms that file an EEO-1 report would necessarily file a report from the NALP, a random sample of 125 was drawn from the EEO-1 data subset.⁴⁵ Of these 33 were not included in the NALP Directory making the actual sample size, 92.

An odds ratio is computed and tested to examine the relationship between partner and associate.⁴⁶ Comparisons are drawn between the control group of White men and the various groups of interest, women, African Americans, Hispanics and Asians. The odds ratio is based on the odds of White males being partners based on their employment as associates divided by the odds of the group of interest, for example, women being partners based on their employment as associates.⁴⁷ When two groups have the same odds of being selected, the odds ratio will be equal to one. Based on the equation applied here, the more the odds ratio exceeds one the more likely it is that White males will be partners. The lower the odds ratio below one, the more likely that the group of interest (for example, women) will be partners.

The test here is not the probability that a female associate will become a partner (necessary data is not available for that analysis). Rather, it is the chance that, given a group of associates and attorneys in a firm, a particular woman, is a partner. Thus, the longevity of partners and historical lack of women and minority associates may produce odds that are different than a woman or a minority's odds of becoming a partner. Nevertheless, this odds ratio, as applied here provides an insight into the status of women and minorities in these firms. Also given the time required to become partner of five to nine years (p. 528)⁴⁸, there has been some time for women and minorities participating in the large growth of law degrees obtained, as discussed in a prior section, to become partners.

Odds ratios are computed for each of the sample firms. Table 5 displays the results when examining women attorneys. The average number of women and White male associates in the sample firms are nearly identical (37.68 for women and 37.60 for White men). However, the mean number of White male partners far exceeds the mean number of women partners at 12.71 percent. The mean odds ratio for this comparison is 5.330; clearly not even odds for the two groups. There was only one instance where the computed odds ratio was at even odds (1) or lower. The average sampled firm would require another 11 women partners to make the proportion of women partners match the proportion of female associates.

Table 5: White Male Attorneys and Women Attorneys Sample Private Law Firms

One must keep in mind that this simple analysis holds qualifications constant. Associates with ten years experience are handled the same as newly hired associates. Due to the recent increase in women and minorities in the legal profession, one would expect their experiences to be more limited than their White male counterparts. For example, in 1988 the median age (reflecting experience) for male attorneys was 42 but just 34 for women (p. 375).⁴⁹ Further, it does not account for variations in the frequency of partnership decisions. However, these results are not necessarily inconsistent with empirical research that controls for such factors. For example, Hull and Nelson report that,

[C]ontrolling for seniority and a wide range of other potentially relevant variables, women's odds of working as law-firm partners are less than one-third of men's odds. Because firm partners command the most money and prestige in the profession, women occupy a distinctly unequal position among lawyers (p. 250).⁵⁰

Table 6 displays the results for African American attorneys. The mean number of African American associates in the sample firms is 4.413 and the mean number of African American partners is 1.076. The odds ratio does not approach the even odds of one. On average, it would require another African American partner at each sample firm to make the proportion of African American partners match the proportion of African American associates. Of the 92 sample firms, 87 (94.57 percent) have an odds ratio greater than one.

Table 6: White Male Attorneys and African American Attorneys Sample Private Law Firms

Table 7 provides the results of a parallel analysis for Hispanics. The mean number of Hispanic associates is 1.837 compared to the mean number of Hispanic partners of 0.630. The mean odds ratio is, again, far from one at 4.262. For these firms on average, they are less than one Hispanic partner short of being having the proportion of Hispanic partners equivalent to the proportion of Hispanic associates in the sample firms.

Table 7: White Male Attorneys and Hispanic Attorneys Sample Private Law Firms

The average number of Asian associates in the sample firms is 6.4 and the mean number of Asian partners is 0.804. The odds ratio here is 7.313. On average, if each sampled firm added two Asian partners, the proportion of Asian partners would match the proportion of Asian associates in the sample firms. In ten (10.87 percent) of the sample firms, the odds ratio is less than or equal to one suggesting that in ten percent of the firms, Asians have the same or better odds than White males as being a partner. See Table 8.

Table 8: White Male Attorneys and Asian Attorneys Sample Private Law Firms

Because the federal government does not collect data on disabled individuals employed in the legal profession, the NALP data provides a unique opportunity to compare the status of disabled attorneys to non-disabled attorneys. This analysis, summarized in Table 9, resembles those provided for other groups but the comparison group is not White males but all attorneys other than those reported as being disabled. The mean number of disabled associates in the sample firms is 0.0761 and the mean number of disabled partners is 0.120. The odds ratio is very close to one at 1.226. Unfortunately, given the very small sample of disabled attorneys being reported these results have limited value.

Table 9: Disabled and Non-Disabled Attorneys Sample Private Law Firms

Law Professors

In order to place these findings in context, odds ratios were computed for the same groups using data from law schools.⁵¹ Assistant Professors were compared to Associate Professors to parallel the Associate to Partner decision. Table 10 summarizes the results. It was not possible to replicate the exact methodology utilized above because the level of detail in the data does not allow comparison groups based on White males. Therefore, women were compared to men and race/ethnic groups were compared to White professors. Data regarding disability was not available.

Table 10: Odds Ratios for Assistant and Associate Professors of Law

Note that in this sector, all of the odds ratios approach one indicating even odds. This provides an interesting comparison as labor market competition and qualifications for the promotion from Assistant to Associate law professor seems likely to approximate that found in the movement from associate to partner in private law firms.

Summary

Using an odds ratio to compare different gender and race/ethnic groups chances of being partners in a sample of law firms suggests disparities between their odds and those of White men. The group with the lowest probability of being partners is Asians with a mean odds ratio of 7.3. The second lowest group is African Americans (6.9) followed by women (5.3) and Hispanics (4.2). The relatively high standard deviation for these measures, as shown in Table 5 through Table 9, suggests that not all firms within the sample behave the same so that some firms exhibit more equitable rates..

Additional Information

For additional information, visit our web site at http://www.eeoc.gov. Click on STATISTICS and JOB PATTERNS FOR MINORITIES AND WOMEN (https://www.eeoc.gov/statistics/employment/jobpatterns/eeo1) for sample copies of the EEO-1 form, an instruction booklet and aggregate statistics.

 Prepared By:      Office of Research, Information and Planning     U.S. Equal Employment Opportunity Commission     Washington, D.C..    

Methodological Appendix: Comparison of EEO-1 and NALP Surveys

The EEO-1 Survey of Establishments in Private Industry is designed to cover a wide range of industries and job groups. To evaluate the accuracy of the EEO-1 information for law firms, we compared the 2002 EEO-1 data for selected law firms to the 2002-2003 National Association for Law Placement (NALP) Directory of Legal Employers. The analysis in this methodological appendix focuses on two main topics: the extent of agreement or disagreement in estimating the total number of attorneys, and the extent of agreement or disagreement in estimating the proportion of women and minorities. In particular, we compare the EEO-1 professional job group for legal services establishments to the NALP data on attorneys, paralegals, and other professionals. Before examining the results, we briefly summarize the instructions provided to EEO-1 and NALP respondents.

Image

EEO-1 and NALP Survey Instructions

The written EEO-1 survey instructions provide general descriptions of job group positions but they do not provide explicit instructions on the employees of legal services establishments.

Professionals. - Occupations requiring either college graduation or experience of such kind and amount as to provide a comparable background. Includes: accountants and auditors, airplane pilots and navigators, architects, artists, chemists, designers, dietitians, editors, engineers, lawyers, librarians, mathematicians, natural scientists, registered professional nurses, personnel and labor relations specialists, physical scientists, physicians, social scientists, teachers, surveyors and kindred workers.⁵²

For firms with multiple offices, law firms are expected to follow the general EEO-1 instructions for establishments at different locations. That is, employment information on the main headquarters office is reported regardless of size. Non-headquarter offices, with a minimum number of 50 or 100 employees, should provide separate reports for each location. In addition, firms supply a consolidated report, aggregating all offices together including offices with less than 50 total employees.

The NALP instructions include detailed definitions for selected types of law firm employees. For example, senior attorneys are defined as "lawyers who were hired for a partnership track position but did not or have not yet become partners and who remained at the firm (whether or not your firm refers to them by another name) or lawyers who were hired for full-time, non-partner positions." Staff attorneys are defined as "lawyers hired as non-partnership track associates or for a fixed term of employment (sometimes referred to as contract attorneys)." The terms "partner/member," "associate," and "paralegal" are apparently undefined. The information on hours worked is limited to "full-time partnership track associates," but it is not clear whether the demographic information on associates includes part-time employees. The official supplying the information is asked whether the information reflects one office only or multiple offices. It should be noted, however, there is no guidance about the minimum size required for reporting purposes nor is there a specification of what constitutes an "office" (potentially a problematic issue for firms with multiple locations in the same city).

Comparison Sample

To compare the corresponding information from the EEO-1 and NALP surveys, we retrieved all establishments from the 2002 EEO-1 Survey with a Standard Industrial Classification code of "81" (Legal Services), a total of 1,231 establishments. (This sample did not include Hawaii, as race/ethnic data is not collected there.) We restricted the population to 782 establishments by only including those with more than one hundred total employees and more than fifteen professional employees. Finally, we restricted the analysis to headquarters firms only, excluding all auxiliary field units. From this universe of 508 establishments, we drew a random sample of 125 establishments from the 2002 EEO-1 survey. By matching firm names and address, we were able to identify 92 law firms in the 2002-2003 NALP Directory of Legal Employers (using NALP data as of February 1, 2002). This represents 73 percent of the original sample and 18 percent of the EEO-1 reporting firms meeting the criteria for analysis. Since NALP is a fee-charging listing service, designed to provide information to potential job seekers, it seems likely that the unmatched firms either had other means of recruiting attorneys or chose not to hire attorneys in 2002-2003.

Comparative Measures

The EEO-1 survey provides three basic ways of measuring the total number of professional employees: the total number of professional employees at the firm's headquarters, the total number of professional employees in the firm's field units, and the total number of professional employees overall. For present purposes, we concentrated on the first and third measures, the headquarters report (hereafter, HD) and the overall consolidated report (hereafter, CN). We compared the HD and CN figures from the EEO-1 survey to seven potential measures of professional employment in the NALP survey:

1. the total number of associate attorneys⁵³
2. the total number of non-partner attorneys (including associates, of counsel, senior attorneys and staff attorneys)
3. the total number of non-partner attorneys plus non-lawyer professionals (such as economists, accounts, and lobbyists)
4. the total number of non-partner attorneys plus non-lawyer professionals and paralegals
5. the total number of all attorneys (including partners and non-partners)
6. the total number of all attorneys plus non-lawyer professionals (such as economists, accountants, and lobbyists)
7. the total number of all attorneys, non-lawyer professionals, and paralegals.

For each law firm, we computed the absolute difference between the EEO-1 numbers and each of the NALP numbers, producing a total of fourteen comparisons (seven HD measures and seven CN measures). We then identified the comparison that produced the smallest absolute difference and calculated a discrepancy proportion using the appropriate EEO-1 base, either HD or CN.

Appendix Table 1: Hypothetical Example

Appendix Table 1 provides an example of these discrepancy computations for a hypothetical law firm. According to Appendix Table 1, the smallest HD discrepancy is a difference of six employees for the NALP category, all attorneys plus other non-lawyer professionals. The proportional disagreement for the smallest HD discrepancy is 0.038 (6/157). The smallest CN discrepancy is a difference of eleven employees for the NALP category all attorneys plus non-lawyer professionals and paralegals. The proportional disagreement for the smallest CN discrepancy is 0.061 (11/181). Since the HD proportional disagreement of 0.038 is less than the CN proportional disagreement of 0.061, the overall proportional disagreement between the EEO-1 and NALP data sources is 0.038. The closest fit to the EEO-1 HD data would be assigned to the NALP data for all attorneys plus non-lawyer professionals.⁵⁴

Comparative Results for the Number of Law Firm Employees

Appendix Table 2 summarizes the overall proportional disagreement between the EEO-1 and NALP data sources.

Appendix Table 2: Proportional Disagreement Between Data Sources

The median or fifty percent value is 0.036. Approximately one-half of the observations fall within the range of 0.017 (First Quartile) to 0.074 (Third Quartile). The average proportional disagreement, which is strongly influenced by extreme values, is 0.054 with a standard deviation of 0.066. In a preliminary examination of these results, eleven of the law firms with the largest discrepancies were reviewed in detail to determine if there were any systematic reasons for the differences. Many of the differences appear to be consistent with the EEO-1 reporting requirements. For example:

Law Firm A. Disparity appears to be due to the combination of two offices in the sample city for the NALP survey but not the EEO-1 survey. When the EEO-1 reports for two offices are grouped together, the total number of EEO-1 professionals is within 4.6 percent of the NALP report for total attorneys and other professionals combined.

Law Firm B. Although the NALP entry indicates that a collective form was used, the NALP counts appear to come from a single office. If the NALP entry is compared to the corresponding EEO-1 report for a single establishment and partners are eliminated from the NALP entry, the NALP results are within 5.8 percent of the EEO-1 results.

Law Firm C. The revised difference reduces to 6.7 percent if partners reported to NALP are not included.

Law Firm D. Combining associates, of counsel, and other non-lawyer professionals results in an exact match with the EEO-1 data.

Law Firm E. The NALP data and the EEO-1 data can be made consistent under the following conditions. The total number of professionals reported to the EEO-1 appears to equal the total number of Of Counsel, Associates, Staff Attorneys and Other Professionals reported to the NALP if one subtracts 22 technical workers on the EEO-1 report from the 32 other professionals on the NALP report. We suspect that the firm includes technical workers in its NALP report because there is no technician category on the NALP.

The remaining firms with outlying discrepancies either cannot be readily explained or they may be law firms that (incorrectly) classify summer law students as professional employees.

Generally speaking, the total number of professional employees reported to the EEOC is a fairly accurate proxy for the relative size of a law firm. The correlations between the total number of professional employees in the EEO-1 survey and the number of attorneys in the NALP sample range from 0.516 to 0.813. The EEO-1 measure of professional employees has a 0.516 correlation with the NALP measure of total partners, a 0.813 correlation with the NALP measure of total associates, and a 0.781 correlation with the NALP measure of total attorneys (all three correlations have probability values less than 0.0001). While the EEO-1 figures for professional employees is related to attorneys in a firm, the figure lacks reliability as a measure for attorneys. There are three main reasons for this reliability issue. First, some firms, as noted above, hire professionals other than attorneys. While a small proportion of the professional work force, it prevents exact tracking of attorney employment. Second, firms did not appear to be consistent in the manner in which they report partners. In examining data from individual firms it is apparent that some include partners in their EEO-1 reporting and others do not. Third, some firms appear to include paralegals as professional employees on their EEO-1 report.

Appendix Table 3: Matching NALP Employment to EEO-1 Organizational Unit (Report Type)

Perhaps the best way to see the "attorneys plus others" aspect of the EEO-1 data is to examine the discrepancies with the NALP data. Approximately two-thirds of the NALP law firms (68.5 percent) are most closely matched to the EEO-1 headquarters data rather than the EEO-1 consolidated data. This suggests that a majority of NALP respondents are reporting information based on a single location.⁵⁵ See Appendix Table 3.

Appendix Table 4 summarizes the assigned reasons for the EEO-1 and NALP discrepancies. The rows identify the EEO-1 source, either consolidated data or headquarters data, which produces the smallest discrepancy with the NALP information.

Appendix Table 4: Matching Various NALP Job Combinations to EEO-1 Professional Job Group by Organizational Unit

The columns identify the corresponding NALP data source (job combination) that produces the smallest discrepancy with the EEO-1 information. The results can be interpreted in several different ways. For many firms, the total number of professional employees reported to the EEOC closely resembles that total number of attorneys reported to the NALP. Combining all the attorney reasons together (that is, associates only, non-partners, and all attorneys), about two-fifths of the law firms (40/92 or 43.5 percent) report a similar number of employees on both surveys. Within the combined attorney group, the total number of attorneys is the dominant reason for reduced discrepancies (26 of 40), followed by the total number of non-partners (12 of 40). In very few firms (2.2 percent) does the EEO-1 professional category and the NALP associate attorneys category coincide.

In addition to lawyers, a substantial number of law firms appear to include other selected occupations in the EEO-1 professional job group. Non-lawyer professionals, such as accountants, are unevenly distributed across law firms. One-quarter of the firms have no non-lawyer professionals, but the upper ten percent of the firms have 32 or more non-lawyer professionals. The median number of non-lawyer professionals is seven employees per firm. Grouping attorneys and other professionals together, about three-fifths of the law firms (56/92 or 60.9 percent) report a similar number of total employees on both surveys. The remaining law firms appear to classify paralegals as professional employees. The median number of paralegals is 25 employees per firm. The combined category of attorneys, non-lawyer professionals, and paralegals constitute about two-fifths (36/92 or 39.1 percent) of the reasons for reduced discrepancies. These results probably reflect the legitimate disagreements about controversies surrounding the terms "legal secretary," "legal assistant," and "paralegal." In summary, it appears that the EEO-1 professional category is composed predominantly of attorneys with substantial number of law firms including non-lawyer professionals and/or paralegals. About three-fifths (57/92 or 62.0 percent) of the law firms appear to report various combinations of all attorneys, including both partners and associates, to the EEOC and about two-fifths (35/92 or 38.0 percent) report various combinations of associate attorneys to the EEOC.

Diversity Results

This section examines the relationship between diversity proportions for law firms in the EEO-1 survey and diversity proportions for law firms in the NALP survey. Appendix Table 5 summarizes the overall proportion of women and minorities for law firms in the 2002 EEO-1 survey.

Appendix Table 5: Proportion of Women and Minorities Employed Sample versus Total

The first row represents all law firms in the 2002 EEO-1 survey (a total of 1,231 observations), and the second row represents the law firms in random sample with corresponding values in the NALP survey (a total of 92 observations). The proportions of women and minorities from the two EEO-1 sources have similar values, between 0.403 and 0.417 for women and between 0.122 and 0.128 for minorities. This suggests that the random sample accurately reflects the proportion of women and minorities among EEO-1 law firms as a whole.

The next two tables summarize the proportion of women and minorities in the 2002-2003 NALP survey organized by associate and partner attorneys. Appendix Table 6 examines summary data for all firms, while Appendix Table 7 only looks at the sample of NALP reporting firms used in this report.

Appendix Table 7: Partners and Associates for All NALP Reporting Firms

Again, the proportions of women and minorities from the two NALP data sources are in close agreement. Notice that EEO-1 proportions generally track the NALP proportions for associates rather than the NALP proportions for partners or for associates and partners combined together. Comparing population values, the EEO-1 and NALP proportional differences among associates are approximately 0.02, 0.021 (0.403-0.424) for women and 0.015 (0.128-0.143) for minorities. The corresponding proportional differences among partners are 0.240 (0.403-0.163) for women and 0.091 (0.128-0.037). The corresponding proportional differences among associates and partners combined are 0.096 (0.403-0.307) for women and 0.032 (0.128-0.096) for minorities.

Appendix Table 8: Partners and Associates for Sample NALP Reporting Firms

Why should the EEO-1 diversity proportions for professionals appear to be in general agreement with the NALP diversity proportions for associate attorneys? Given that approximately two-thirds of the law firms appear to be reporting EEO-1 professional data for partners and associates combined, why are the EEO-1 diversity proportions substantially higher than the NALP diversity proportions for partners and associates combined? There are no conclusive answers to these questions, but the following tables are instructive.

Appendix Table 9 is restricted to the fifty-nine sampled law firms reporting similar employee numbers on both the EEO-1 and the NALP surveys (defined as proportional disparities of 0.05 or less). It compares the imputed discrepancy categories, derived from the NALP data, to the EEO-1 diversity proportions for women and minorities. The first row represents sampled law firms with associate and non-partner discrepancy explanations (associate only, non-partner, non-partner plus other professionals, and non-partner plus other professionals and paralegals). The second row represents sampled law firms with partner and non-partner discrepancy explanations (all attorneys, all attorneys plus other professionals, and all attorneys plus other professionals and paralegals). That is, the first row excludes partner attorneys, and the second row includes both associate and partner attorneys. The NALP results would suggest that the diversity proportions should be higher for EEO-1 responses that exclude partners than for EEO-1 responses that include partners. Consistent with the NALP results, the diversity proportions for the first row are higher than the diversity proportions are for the second row. The proportions of women are 0.463 and 0.375 respectively, and the proportions of minorities are 0.151 and 0.116 respectively. Since some of the law firms classified as non-partners probably include paralegals and other professionals in their EEO-1 reports, the diversity proportions for non-partner respondents on the EEO-1 survey should also be higher than the diversity proportions for "pure" associates reported to the NALP. This expectation is generally confirmed, especially for the proportion of women. For sampled law firms classified as non-partner respondents, the proportion of women is 0.463 on the EEO-1 survey compared to a proportion of 0.428 among associate attorneys on the NALP survey. Likewise, the proportion of minorities is 0.151 compared to 0.146 among associate attorneys on the NALP survey. These results suggest that discrepancy categories, derived from a comparison of total number of employees on the two surveys, are also plausible (or at least potential) explanations for the diversity proportions in the sampled law firms.

Appendix Table 9: Sampled Law Firms with Proportional Differences of 0.05 or Less

Appendix Table 10 attempts to answer the question, what can the NALP data tell us about the likely distribution of different occupational groups (such as partners, associates, and paralegals) in the EEO-1 professional category. As described earlier, the NALP data reports employment totals for a wide range of occupational groups, but it only reports diversity numbers for selected occupational groups (partners, associates, summer students, and auxiliary attorneys). The EEO-1 survey does not differentiate among different types of professional employees that might be present in law firms. Based on household occupation data from the 2002 Current Population Survey (CPS), we made rough estimates for the missing NALP diversity proportions among non-lawyer professionals and paralegals and then used these NALP diversity estimates to predict the overall diversity proportions among EEO-1 professional employees. Specifically, we proceeded as follows:

1. The sample was limited to law firms with minimal discrepancy proportions, i.e., discrepancy proportions of 0.05 or less (a total of fifty-nine law firms).
2. The EEO-1 diversity proportions were computed using the appropriate baseline assigned by the discrepancy measure, either the proportion of headquarters professionals or the proportion of consolidated professionals.
3. Because of time limitations, the NALP category of senior attorneys, of counsel, etc. was eliminated. Diversity proportions for the NALP category of non-lawyer professionals were estimated by averaging the 2002 CPS data for accountants and economists (an average proportion of 0.570 for women and 0.076 for minorities). Diversity proportions for the NADP paralegal category were estimated by the 2002 CPS data for legal assistants (proportions of 0.822 and 0.181 for women and minorities respectively). The remaining NALP categories were based on the actual NALP sample data for partners and associates.
4. The overall NALP diversity was calculated using the appropriate discrepancy category for the sampled firm. For example, law firms classified as non-partners plus non-lawyer professionals were based on the actual number of NALP women and minority associates plus the estimated number of women and minorities among non-lawyer professionals. In a similar fashion, law firms classified as all attorneys plus non-lawyer professionals and paralegals were based on the actual number of women and minorities among partner and associate lawyers plus the estimated number of women and minorities among non-lawyer professionals and paralegals.
5. We then compared the predicted diversity proportions, obtained from the combination of actual and estimated NALP data, to the actual diversity proportions in the EEO-1 data.

The objective of this exercise was to assess the reasonableness of the NALP data as a proxy for the different types of legal sub-groups likely to be present in the EEO-1 data. If we could show that the NALP diversity estimates closely approximated the EEO-1 diversity proportions, we then have a basis for extrapolating the NALP occupational allocations to the EEO-1 data.

As expected, the NALP diversity estimates, derived from the appropriate legal sub-groups, provide a close approximation to the diversity proportions reported to the EEOC. The EEO-1 proportion of women is 0.4023 compared to an NALP estimate of 0.3826, a difference of 0.0197. The EEO-1 proportion of minorities is 0.1246, compared to an NALP estimate of 0.1152, a difference of 0.0094. Neither of these proportional differences is statistically significant.

Based on the NALP sample, the corresponding EEO-1 submissions are, on average, likely to be composed of the following legal sub-groups in Appendix Table 10. The third column shows the proportion of different legal sub-groups in the NALP sample. The fourth column shows the average proportion of women within these sub-groups. The fifth column combines the information on sub-group proportions and average proportions of women to estimate a weighted average proportion of women for each sub-group. The results suggest that about three-fourths (75.86 percent) of the professional employees reported to the EEOC are either partner or associate attorneys. Associates form the largest legal sub-group in the EEO-1 (42.0 percent), followed by partners (33.8 percent), paralegals (16.9 percent) and non-lawyer professionals (7.3 percent).

Appendix Table 10: Average Proportion of Women by Job

Although these sub-group percentages are only rough approximations, the estimated weighted averages help to explain why the overall EEO-1 diversity proportions tend to track the NALP diversity proportions for associate attorneys. Although non-lawyer professionals and paralegals are the smallest legal sub-groups, they each have a relatively high proportion of women and minorities. Combined together, the weighted average for women among non-lawyer professionals and paralegals (0.1802) is almost equal to the weighted average among associate attorneys (0.1783). Likewise, the weighted average for minorities among non-lawyer professionals and paralegals (0.0434) is almost equal to the weighted average among associate attorneys (0.0600). Thus, on average, the relatively small number of women and minority partners is counter-balanced by a combination of other legal sub-groups that are similar to associate attorneys in terms of overall contributions to the proportion of women. Put another way, the EEO-1 data seems to provide a fairly accurate surrogate measure of diversity among associate attorneys (at least in the current time period), not because the EEO-1 data is a strict measure of associate employment, but because the mixture of "attorneys plus other occupations" produces diversity proportions much closer to the proportions for associate attorneys than the proportions for all attorneys, partners and associates, pooled together.

Summary

Although the relationships between the EEO-1 and NALP surveys are complex and tenuous, two general principles seem evident. First, because the EEO-1 professional category seems to be composed predominantly of attorneys, combined with selected other legal sub-groups, the EEO-1 data provides a fairly accurate index of a relative firm size, but it is not a reliable guide to the actual number of partners and associates in any given law firm. Second, the particular mixture of legal sub-groups covered in the EEO-1 survey appear, on average, to provide a plausible indicator of diversity among associate attorneys, but they do not appear to be representative of diversity among partner attorneys or partner and associate attorneys combined. Finally, the measurement problems reviewed in this appendix suggest that the available information on law firms would be greatly improved if all of the organizations responsible for collecting data on law firms could reach a consensus on the appropriate measurement standards to be applied to law firm surveys. For example, the NALP survey provides detailed information on specific legal sub-groups (such as partners, associates, and senior attorneys), but respondents appear to need more guidance on the appropriate definition of a law firm's "office," especially when multiple offices are located in the same geographic area. Likewise, the EEO-1 survey provides fairly detailed definitions for law firm locations, but respondents appear to need more guidance on how to allocate specific legal sub-groups among the EEO-1 job groups. Hopefully this appendix has suggested various ways in which the available law firm surveys can be compared and contrasted to yield insights that might not be evident from any given survey taken by itself.

SELECTED REFERENCES

Beckman, Christine M., and Damon J. Phillips. "Interorganizational Determinants of Promotion: Client Leadership and the Promotion of Women Attorneys", unpublished manuscript, 2003.

Chambliss, Elizabeth, "Organizational Determinants of Law Firm Integration," The American University Law Review, Volume 46, pp. 695-696, pp. 669-746, 1997.

Dixon, Jo, and Carroll Seron, "Stratification in the Legal Profession: Sex, Sector, and Salary" Law & Society Review Volume 29, Number 3, pp. 381-412, 1995.

Domhoff, William, Who Rules America? Power and Politics in the Year 2000. 3rd Edition. Mountain View: CA: Mayfield Publishing, 1998.

"Gender Equity in the Legal Profession - A Survey, Observation and Recommendations", New York State Bar Association. Hagan, John, and Fiona Kay, Gender in Practice, New York: Oxford University Press, 1995.

Heinz, John P., and Edward O. Laumann, Chicago Lawyers, New York: Russell Sage Foundation and American Bar Foundation, 1982.

Johnson, Alex M., Jr. "The Underrepresentation of Minorities in the Legal Profession: A Critical Race Theorist's Perspective", Michigan Law Review Volume 95, pp. 1005-1062, 1997.

Kay, Fiona, and John Hagan, "Cultivating clients in the Competition for Partnerships: Gender and the Organizational Restructuring of Law Firms in the 1990s", Law & Society Review Volume 35, Number 3, pp. 517-555,1999.

Kay, Fiona, "Flight from Law: A Competing Risks Models of Departures from Law Firms", Law & Society Review Volume 31, Number 2, pp. 301-335, 1997.

Leicht, Kevin T., and Mary L. Fennell, Professional Work A Sociological Approach, Malden, Massachusetts: Blackwell Publishers, 2001. Mills, C. Wright, The Power Elite, Oxford University Press, Inc.,1956.

Nelson, Robert L., and Laura Beth Nielsen, "Cops, Counsel, and Entrepreneurs: Constructing the Role of Inside Counsel in Large Corporations", Law & Society Review Volume 34, Number 2, pp. 457-494, 2000.

Nelson, Robert L., Partners with Power, The Social Transformation of the Large Law Firm. Berkeley: University of California Press,1988.

Nelson, Robert L., "The Futures of American Lawyers: A Demographic Profile of a Changing Profession in a Changing Society", Case Western Reserve Law Review, Volume 44, pp 345-406, 1994.

Schultz, Ulrike, and Gisela Shaw ed., Women in the World's Legal Professions. Portland, Oregon: Hart Publishing, 2003.

Smigel, Erwin O., The Wall Street Lawyer: Professional Organization Man? Bloomington: Indiana University Press,1969.

Sterling, Joyce S., and Nancy J. Reichman, "Recasting the Brass Ring," Forthcoming in Capital University Law Review. "The Global 100 Rankings", The American Lawyer, November 2002 "The Top 100 Most Prestigious Firms - 2002".

The Vault. 2002 http://www.vault.com/nr/lawrankings . Wilkins, David B. and G. Mitu Gulau, "Reconceiving the Tournament of Lawyers: Tracking, Seeding, and Information Control in the Internal Labor Markets of Elite Law Firms", Virginia Law Review Volume 84, pp. 1581-1705, 1998.

Wilkins, David B, "Why Global Law Firms Should Care About Diversity: Five Lessons from the American Experience". European Journal of Law Reform Volume 2 Number 4, pp. 415-438, 2000.

Wilkins, David B., and G. Mitu Gulati, "Why are there so Few Black Lawyers in Corporate Law Firms? An Institutional Analysis", California Law Review , pp. 493- 618,1996.

---

1. Hagan J. and F. Kay, Gender in Practice: A Study of Lawyers' Lives, New York: Oxford Press, 1995, p. 3.

2. Nelson, R. "The Futures of American Lawyers: A Demographic Profile of a Changing Profession in a Changing Society", Case Western Reserve Law Review, vol 44, 1994, pp. 345-406.

3. Johnson, Jr. A., "The Underrepresentation of Minorities in the Legal Profession: A Critical Race Theorist's Perspective", Michigan Law Review, vol 95, February 1997 pp. 1105-1062.

4. Nelson, R, Partners with Power: The Social Transformation of the Large Law Firm, Berkeley: University of California Press.

5. Monique R. Payne and Robert L. Nelson, "Shifting Inequalities: Stratification by Race, Gender, and Ethnicity in an Urban Legal Profession, 1975-1995," 2003, unpublished manuscript.

6.Elizabeth Chambliss, "Organizational Determinants of Law Firm Integration," 1997, The American University Law Review, vol. 46, pp. 669-746.

7. There are, however, several major articles with substantial data on minorities. These include the study of the members of the Michigan Law School classes of 1970-96 undertaken by Richard O. Lempert, David L. Chambers, and Terry K. Adams in "Michigan's Minority Graduates in Practice: The River Runs Through Law School," 2000, "font-family: Shruti" style=Law and Social Inquiry, pp. 395-505 and the study of the New York University Law School classes of 1987-90 undertaken by Lewis A. Kornhauser and Richard L. Revesz in "Legal Education and Entry into the Legal Profession: The Role of Race, Gender, and Educational Debt," 1995, "font-family: Shruti" style=New York University Law Review, vol. 70, pp. 829-964.

8. For a general discussion of the factors affecting law firm attrition and their changes over time, see Rebecca L. Sandefur, 2003, "Attrition from the Legal Profession and Mutable Labor Markets for American Lawyers, 1949-2000," unpublished manuscript prepared for presentation at the Annual Meetings of the American Sociological Association, Atlanta, Georgia.

9. The NALP Foundation for Law Career Research and Education, Keeping the Keepers II: Mobility and Management of Associates, 2003, Washington.

10.For a detailed discussion of attrition among Black associates, including scarce training opportunities and access to good work assignments, see David B. Wilkins and G. Mitu Gulati, "Why are There So Few Black Lawyers in Corporate Law Firms: An Institutional Analysis," 84 California Law Review, May 1996, pp. 493-618.

11.In a discussion of part-time work and career interruptions, Sterling and Reichman quote a women attorney from the Denver area who says, " ... There are very few women who are partners with traditional lives. Very few. And the ones that are there are not succeeding ... [The ones succeeding] they've either got a stay-at-home partner, husband, whatever, they don't have kids. They're the primary bread-winner." Another women attorney says, " ... I mean you just can't be gone a year. If you gone a few months, clients can kind of make due while you are gone; they don't really have to shift their loyalties. If you're gone a year, you know, some of "font-family: Times New Roman" style=them go off to different lawyers." See Joyce S. Sterling and Nancy J. Reichtman, "Recasting the Brass Ring: Deconstructing and Reconstructing Workplace Opportunities for Women Lawyers," forthcoming, Capital University Law Review.

12.Mary C. Noonan, Mary E. Corcoran, and Paul N. Courant, "Pay Differences Among the Highly Trained: Cohort Differences in the Gender Gap in Lawyers' Earnings," unpublished revised manuscript based on presentation at the Population Association of America annual meeting in Atlanta, 2002.

13.Cynthia Fuchs Epstein, and Robert Saute, Bonnie Oglensky, and Martha Gever, "Glass Ceilings and Open Doors: Women's Advancement in the Legal Profession," A Report to the Committee on Women in the Profession, The Association of the Bar of the City of New York, 1995,Fordham Law Review, vol. 64, p. 291-449.

14. See John P. Heinz and Edward O. Laumann, Chicago Lawyers, 1982, Russell Sage Foundation and American Bar Association, and Kathleen E. Hull and Robert L. Nelson, "Assimilation, Choice or Constraint? Testing Theories of Gender Differences in the Careers of Lawyers," 2000 Social Forces, vol. 79:1, pp. 229-264. "font-family: Times New Roman">

15. The methodological appendix provides analyses comparing EEO-1 reports to a sample of law firms. It specifically examines the relationship between the EEO-1 professional job group and more detailed job titles in law firms. While reliability problems are documented, a relationship between the professional EEO-1 job group and the associate job title is suggested. For the purpose of this report, law firms are defined based on the Standard Industrial Classification code for Legal Services. A parallel code exists for the North American Industrial Classification System. The EEO-1 reports used in this report did not include those from Hawaii, as race/ethnic data is not collected there.

16. Current Population Survey data is a national monthly survey of approximately 60,000 households conducted by the Bureau of the Census for the Bureau of Labor Statistics. The data used is "Household Data, Annual Averages, Employed Persons by Detailed Occupation, Sex, Race and Hispanic Origin. The 2002 data is available at www.bls.gov/cps/cpsaat11.pdf. Data for other years was obtained directly from the Bureau of Labor Statistics. Data for 1982 is not available so data from the 1980 Census data is substituted for those figures.

17. Data for women and total for 1982 obtained from 1982 Review of Legal Education, American Bar Assocation and the Law School Admission Council, 1982. Other data regarding total and women degrees conferred from www.abanet.org/legaled/statistics/jd.html. Data regarding minority degrees from www.abanet.org/legaled/statistics/mindegrees.html.

18. Charlotte Chiu and Kevin T. Leicht, "When Does Feminization Increase Equality? The Case of Lawyers," 1999 Law and Society Review, vol.33, p. 557-590.

19. The percentages in private practice (comprising solo practice, small firms, medium firms, and large firms) were recomputed from Table 2-2 by the authors of this report.

20. See the summary table on the website, https://www.nalp.org/2002presenceofwomenandattorneysofcolor

21. Readers are reminded that the EEO-1 survey uses a broad definition of professional employees that covers attorneys as well as other non-attorney legal occupations. For the reasons described in the appendix, it appears that the EEO-1 data on legal professionals provides a fairly representative index of diversity among associate attorneys. Readers are also reminded that because of establishment size limitations, the EEO-1 survey not provide information on solo practitioners or relatively small law firms. Therefore, this report examines relative variations in firm size within a group of medium to relatively large law firms.

22. The total number of EEO-1 legal professionals, for this purpose, represents the sum from all reporting establishments with a common headquarters number and is not necessarily equivalent to the consolidated totals discussed in the methodological appendix.

23. For a list of the states in each of the nine Census Bureau divisions, see the description on the website, www.bls.gov/help/def/la.htm.

24. The list of the top 100 law firms, ranked by profits per partner, was taken from The American Lawyer, November 2002. The list of the top 100 law firms, ranked by prestige, was taken from "The Top 100 Most Prestigious Firms - 2002". The Vault. 2002. Firms on either list were assigned a value of "1", otherwise a value of "0."

25. The top ten legal markets were estimated by aggregating the number of EEO-1 legal professionals by city and then ranking the cities by the proportion of all legal professionals in the 2002 EEO-1 survey. The top ten cities, encompassing approximately 60 percent of all legal professions in the 2002 EEO-1 survey were New York, Washington, DC, Chicago, Los Angeles, Boston, San Francisco, Philadelphia, Houston, Dallas, and Atlanta.

26. Southern states were defined by the Census Bureau South Region covered the states of Delaware, District of Columbia, Florida, Georgia, Maryland, North Carolina, South Carolina, Virginia, West Virginia, Alabama, Kentucky, Mississippi, Tennessee, Arkansas, Louisiana, Oklahoma, and Texas. This variable is used to capture the higher availability of minorities, particularly African Americans in this area.

27. The correlations among these three variables range from 0.739 to 0.820. A multivariate factor analysis, not reported here, suggests that these three variables should be treated as a single factor.

28. The variations in these average proportions are all statistically significant with F probability values of 0.002 or less.

29. These proportions are the mean of each firm's proportion.

30. It should be noted that the diversity percentages, reported here, are summary measures, computed by summing all establishments with a common EEO-1 headquarters number. Since EEO-1 data is not collected from non-headquarter offices with less than 50 total employees, these percentages are not equivalent to the consolidated figures on legal professionals discussed in the appendix.

31. The F-probability values were computed with angular transformations of the proportional diversity values. Angular transformations are designed to achieve a constant error variance, i.e., they reduce the likelihood of a wider range of errors around 0.5 than around 0.01 or 0.99. See, for example, the discussion of variance-stabilizing transformations "font-family: Times New Roman">in Michael "font-family: Times New Roman">O. Finkelstein and Bruce Levin, Statistics forLawyers, 1990, p. 441.

32. Using the Tukey Studentized Range Test, the comparison between firms with a single office and firms with four or more offices (0.039) is statistically significant at the 0.05 level, but none of the other law firm comparisons (i.e., between single offices and two or three offices and between two or three offices and four or more offices) are statistically significant at the 0.05 level.

33. Using the Tukey Studentized Range Test, all of the law firm minority comparisons are statistically significant at the 0.05 level except for the comparison between law firms with two or three offices and law firms with four or more offices.

34. These EEO-1 results are generally consistent with the 2002 NALP results.

35. This is based on an examination of variations in the proportion of minorities and women among different firms, When the mean values of two groups are unequal, variations around the mean are usually measured by a coefficient of variation expressed as a percentage (computed as the standard deviation divided by the mean). For example, in Table 2 the average proportion of women in single office firms is 0.377, the standard deviation is 0.13225 making the coefficent of variation, 0.3504 (0.13225/0.377) or 35 percent. The higher the coefficient, the greater than the relative dispersion around the mean or average value. The coefficients of variation for women range from 20 percent to 35 percent, and the coefficients of variation for minorities range from 46 percent to 85 percent.

36. Specifically, for the percentage of women, the coefficient of variation decreases from 35 percent in firms with a single office to 20 percent in firms with four or more offices. For the percentage of minorities, the coefficient of variation decreases from 85 percent in firms with a single office to 46 percent in firms with four or more offices.

37. The F probability value shows the statistical significance of the overall regression equation. The standardized parameter value shows the relative strength and direction of each explanatory variable ranging from +1 (strong positive effect) to -1 (strong negative effect). The T probability values show the statistical significance of each explanatory variable controlling for the effects of the other variables in the equation. The dependent variables, that is, the variables that are being explained or predicted by the regression equation, are the proportions of women and minority legal professionals in a law firm standardized by angular transformations.

38. Statistical significance level used here is a probability value less than or equal to 0.05

39. The best fitting models in this study, minority proportions for law firms with one office and two or three offices, have adjusted R-squared values of 0.124 and 0.185 respectively. This suggests that there may be other explanatory variables, not available in the EEO-1 survey, that need to be considered (e.g., firms with different types of legal specialties such as bankruptcy, criminal practice, tax law, corporate mergers, etc.). For a major study of elite law firms that uses data from the NALP Directory to measure area of legal specialization, see Chambliss (1997).

40. Beckman, C. and D. Phillips, "Interorganizational Determinants of Promotion: Client Leadership and the Promotion of Women Attorneys", draft manuscript, August 26, 2003.

41. Kay, F. and J. Hagan, "Cultivating Clients in the Competition for Partnership: Gender and the Organizational Restructuring of Law Firms in the 1990's", Law and Society Review, vol 33 no. 3, 1999, pp. 517-555.

42. Dixon, J. and C. Seron, "Stratification in the Legal Profession: Sex, Sector, and Salary", Law and Society Review, vol. 29 no. 3, 1995, pp. 381-412.

43. Perceptions of Partnership: The Allure and Accessibility of the Brass Ring, National Association of Legal Placement, Foundation for Research and Education, 1999.

44. NALP 2002-2003, Directory of Legal Employers, National Association of Legal Placement, 2002.

45. The SAS^® Institute procedure PROC SURVEYSELECT was used to generate the sample. Additional conditions used were that total employment had to be greater than or equal to one hundred and that the total number of professionals reported had to be greater than or equal to 15.

46. Fienberg, S. (1977) The Analysis of Cross-Classified Categorical Data, The MIT Press, Cambridge, pp. 17-18. Also see Agresti, A. (1990) Categorical Data Analysis, John Wiley and Sons, New York, pp. 14-15; Breslow, N. and Day, N. (1980) Statistical Methods in Cancer Research, IARC, Lyon, p. 125; Finkelstein, M. and Levin, B. (1990) Statistics for Lawyers, Springer-Verlag, New York, p. 2.

47. The formula used in computing the odds ratio comparing White men and women is,

where G= Odds ratio for a glass ceiling, M= Male, F= Female, p = partners, a = associates. Parallel equations are used for the other groups.

48. Kay, Fiona M. and John Hagan, "Cultivating Clients in the Competition for Partnership: Gender and the Organizational Restructuring of Law Firms in the 1990's," Law and Society Review, 33 (3) 1999, pp. 517-555.

49. See Nelson R.,1988, cited above.

50. Hull, K. and R. Nelson, "Assimilation, Choice, or Constraint? Testing Theories of Gender Differences in the Careers of Lawyers", Social Forces, September 2000, 79 (1) pp. 229-264.

51. White R., Updated Tables for the 2001-02 AALS Statistical Report Including Revised Historical Data for the 1990-1991 through 2000-01, Association of American Law Schools forthcoming Association of American Law Schools Statistical Report on Law School Faculty and Candidates for Law Faculty Positions, 2001-2002.

52. See Appendix 5, Description of job categories http://www.eeoc.gov/

53.The NALP Directory of Legal Employers reports two sub-totals for the number of associate attorneys, one in an upper-right box (labeled "Demographics"), the other in a matrix containing counts by gender and minorities. We retained both measures and compared each measure separately to the NALP data. For the purpose of computing overall discrepancy statistics, we used the associate sub-total that produced the smallest discrepancy.

54. This discrepancy measure compares the relative proportion of EEO-1 and NALP employees, and, therefore, may be subject to random fluctuations, especially for small numeric differences. Other discrepancy measures, incorporating estimates of measurement and sampling error between the two data sources, could yield different discrepancy assignments. For example, the effective dates of the two surveys are likely to produce different employment totals reflecting the timing of new hires and terminations, particularly in cases of high turnover. We have not attempted to set a threshold for distinguishing between significant and insignificant discrepancy values.

55. NALP respondents were asked whether the information provided was for one law office only or multiple law offices. We compared the NALP responses (collective form equals "N" and "Y") to the designations of headquarter and consolidated discrepancies used in this study. The two measures of office aggregation levels are not strongly related. About one-half of the consolidated designations (48.3 percent) had a collective form value of "N," and about two-fifths of the headquarter designations (42.9 percent) had a collective form value of "Y." The likelihood that differences of this magnitude could have arisen by chance is 0.502 using a two-sided Fisher's Exact Test.

56. See, for example, the discussions under the heading, Definition of a Paralegal, on the website for the National Federation of Paralegal Associates, Inc. (www.paralegals.org).

57.Readers are reminded that the discrepancy reasons were assigned on the basis of numeric differences between the EEO-1 and NALP surveys. In the absence of interviews with the responding officials, there is no conclusive way to determine how law firms allocated employees among the different EEO-1 job groups.

58.The Z score for women is 0.2197 (a two-sided probability value of 0.826), and a Z score for minorities is 0.3294 (a two-sided probability value of 0.742).

59. Readers are reminded that the estimates for the average proportion of women among non-lawyer professionals and paralegals are based on the 2002 CPS data rather than the NALP sample. It also should be noted that the NALP predictions, described above, used the actual diversity numbers for partners and associates in each law firm rather than overall averages.

---

This page was last modified on October 22, 2003.