Participation in Local Unions: A Comparison of Black and White Members Author(s): Michele M. Hoyman and Lamont Stallworth Source: Industrial and Labor Relations Review, Vol. 40, No. 3 (Apr., 1987), pp. 323-335 Published by: Cornell University, School of Industrial & Labor Relations Stable URL: http://www.jstor.org/stable/2523490 Accessed: 12/01/2010 13:33
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=cschool. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Cornell University, School of Industrial & Labor Relations is collaborating with JSTOR to digitize, preserve and extend access to Industrial and Labor Relations Review.

http://www.jstor.org

PARTICIPATION IN LOCAL UNIONS: A COMPARISON OF BLACK AND WHITE MEMBERS
MICHELE M. HOYMAN and LAMONT STALLWORTH*

This article compares the participation of black and white union members in their local unions. Using more detailed measures of union participation than those employed in earlier studies, and focusing on members, not just leaders, the authors find little difference between the extent of participation by blacks and that by whites. This surprising result, which contradicts the finding of previous studies that blacks participate in unions less than whites, holds even with controls for gender, salary, education, number of years as a member, the presence of friends in the union, the strength of a sense of efficacy, confidence in the ability to gain local union office, and the liberalness of attitudes about civil rights.

BLACK

participation in local unions,

though largely ignored by previous research on union democracy, is an important subject for several reasons. First, unions' responsiveness to blacks as members of a minority group can serve as a measure of union democracy alternative to that commonly employed in previous studies, namely, majority participation, or the overall participation of all union members without regard to race (Lipset, Trow, and Coleman 1956; Magrath 1959; Strauss 1977). Minority members' access to their union can be viewed as a key indication of unions' effectiveness in promoting industrial democracy. An analysis of black participation thus can also indicate whether legislation encouraging union democracy, such as the Landrum-Griffin Act of 1959, has succeeded.
*Michele Hoyman is Assistant Professor of' Political Science, University of Missouri-St. Louis, and Lamont Stallworth is Associate Professor of Industrial Relations, Loyola University. The authors thank the National Institute of Mental Health and the Summer Research Fellowship Program of the University of Missouri-St. Louis for funding; John Delaney, William Gordon, and Philip Metzler for research assistance; and William Gould, James Gross, and Arthur Schwartz for comments on an earlier draft.

Second, a study of black participation can provide an indication of how much progress unions have made in guaranteeing the civil rights of their members since the passage of civil rights legislation in the 1960s and 1970s-legislation intended to increase the participation in all American institutions, including unions, of minorities who were previously discriminated against. Third, this study should also help to resolve the debate in the literature over whether blacks participate in their unions less than whites. Both the scant literature on black participation in unions and the more general research on democratic practices in unions point to blacks participating less than whites. But in studies of participation in politics, blacks have been shown to participate more than whites when the blacks perceive a potential for group gains and when the research design includes controls for socioeconomic status. This study also makes several methodological contributions to the analysis of participation in unions. Based on a survey of a random sample of 2,000 union members, the study elicited detailed information on the respondents' participation in a wide variety of union activities, as well as their

Industrial and Labor Relations Review, Vol. 40, No. 3 (April 1987). C)by Cornell University. 0019-7939/87/4003 $01.00

323

324

INDUSTRIAL AND LABOR RELATIONS REVIEW at least some degree of influence over decisions made in the local union. Several previous studies have suggested that blacks participate in union activities less than whites (see, for example, Gould 1977; Hill 1977; and Gross 1962). The few empirical studies in the literature only analyzed highly formal types of participation, however, such as holding office, voting, and attending meetings (Hoyman and Schwartz 1984; Hagburg and Blaine 1967). But there is further evidence that black participation is limited. For example, among the members of the AFL-CIO Executive Board and (an overlapping group) the presidents of international unions, there currently are very few blacks.' Nonetheless, available data suggest that the number of blacks within the ranks of staff, international union leaders, and executive board members is increasing. In particular, there is a significant number of blacks on the executive boards of several international unions, notably the American Federation of State, County, and Municipal Employees, the Transport Workers Union, the United Automobile Workers, the American Federation of Teachers, and the United Steel Workers.2 This trend, along with the increasing activism of the Coalition of Black Trade Unionists, indicates at the very least that the presence of blacks in the American labor movement has become more noticeable in recent years. Both structural and legal characteristics of the U.S. industrial relations system have been cited as barriers to the participation of minorities. Singled out in particular has been the principle of exclusivity (Hill 1977). Under exclusivity, one bargaining agent represents an entire bargaining unit, and
'Of a total of 35 executive board members in 1986, two were blacks: Frederick O'Neal, President of the Associated Actors and Artists of America, and Barbara Hutchinson, Director of Women's Activities of the American Federation of Government Employees. See Hoyman and Schwartz (1984). 2Hoyman and Schwartz (1984). In 1979, for example, seven of the 34 international vice-presidents of the American Federation of Teachers were black. 'Established in 1972, the Coalition of Black Trade Unionists has pledged to work within the labor movement to lobby for the betterment of black workers.

sense of self-confidence and efficacy; their attitudes toward the local union; their political attitudes, such as ideology and partisanship; and their demographic characteristics. We therefore have been able to construct a more encompassing model of union participation than that employed in previous studies, which have addressed only formal participation activities and a much smaller set of potential antecedents of participation. Moreover, the methodology of this analysis ensures that its results will have greater validity than those of earlier studies of union democracy. The typical study of this topic has employed a captive-audience sample of those union members who attend union meetings, whereas ours uses a large random sample. Random sampling avoids the potential self-selection bias inherent in classic captive-audience sampling. We were also fortunate to overcome the problems of access to the research site that have plagued previous union democracy studies. Finally, because of the large size of our sample and the large proportion of blacks in the union studied, we are confident of the validity of our comparisons between blacks and whites in terms of the amounts and types of participation they exhibit. The Literature on Black Participation in Unions Motivating the general membership to participate in a perennial problem for the local union, but motivating minority members to participate is an even greater challenge. Unions are white majority institutions, and as such they may face in their minority members an "outgroup" that feels disenfranchised, even alienated, from the union as an institution. Blacks, in particular, have been shown to exhibit certain characteristics that are associated with low levels of political participation, such as low income, little education, and a weak sense of efficacy (Verba and Nie 1972; Campbell 1960). Generally, a sense of efficacy is the individual's belief that he or she can make a difference. In this study, a sense of efficacy is defined as the respondent's belief that he or she has

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 325 no separate minority representation is allowed. Critics of the exclusivity doctrine judge that it effectively disenfranchises blacks and other minority groups, particularly since a majority vote is required to replace any bargaining agent. Much of the existing literature on black participation addresses the issue of the means to increase black influence in predominantly white unions (see, for example, Kornhauser 1952; Lamm 1975). Among the strategies mentioned as encouraging black participation are sponsorship (white leaders selecting blacks), guardianship (white leaders responding to black members' needs), and black power (blacks running on separatist platforms or as a separate slate opposing that endorsed by the incumbent leadership). (Campbell 1960; Verba and Nie 1972; Miller et al. 198 1).4 A sense of efficacy is especially pertinent to black participation, as blacks tend to have a lower sense of efficacy than whites. There also is evidence that blacks distrust the political system or feel alienated from it (Aberbach and Walker 1970; Miller 1974). Both factors-the low sense of efficacy and distrust of the political system-would be expected to dampen the participation of blacks in their unions. There is, therefore, some controversy over whether blacks participate in unions more or less than whites. The literature on blacks in unions suggests that they participate less, and the literature on blacks in politics suggests the opposite, at least in cases in which blacks identify as a group and in which SES factors are controlled. Of the two views, the former seems more likely to be true in the union context. The history of labor unions is fraught with racial discrimination, manifest in its most overt forms in segregated locals (Hill 1975; Gould 1977; Hill 1977). In fact, the leading duty-of-fairrepresentation case, Steele v. Louisville and Nashville Railroad [323 U.S. 192; 65 S. Ct. 226 (1944)], involved among other of practices the prohibition blacks from union membership. There are still vestiges of discrimination in unions, and it would be reasonable to expect this discrimination to depress black participation rather than elevate it. Since the weight of the evidence points to blacks participating less than whites in unions, and since the literature on blacks in unions is clearly more relevant to our analysis than that on blacks in politics,

The Literature on Black Participation in Politics Explanations for black participation in unions may be sought by examining what motivates individual blacks to participate in the larger political system by voting, holding an office, or campaigning. The traditional model of citizen participation in politics holds that socioeconomic status (SES) is very important in determining participation levels (Campbell 1960; Verba and Nie 1972). Given the traditionally low average SES of blacks, it is not surprising that previous research without controls for SES has shown less political activity by blacks than by whites (Campbell 1960; Verba and Nie 1972). But studies that have controlled for common SES indicators have found that blacks participate more than whites. What has appeared to be a race effect has actually been an SES effect: the average black apparently participates in politics muchmore, in fact, than the average white of comparable education, income, and occupation (Verba and Nie 1972:149-73). This higher participation rate has been attributed to group consciousness among blacks; that is, blacks identify closely with one another's plight and believe that their own participation can bring about a group gain

'The evidence used to support the group-consciousness hypothesis is that some blacks identify themselves as members of an ethnic minority and others do not. Verba and Nie use the frequency with which race is mentioned in responses to open-ended questions as an indication of blacks' sense of group identity. Miller et al. argue, however, that group identification is not the same as group consciousness. They suggest that group consciousness involves both a psychological identification with a social stratum (group identification) and an awareness of the relative status of one's group as necessary prerequisites to participation in collective action as a means to achieve group gains.

326

INDUSTRIAL AND LABOR RELATIONS REVIEW on participation (socialization literature; see, e.g., Niemi 1973). are family members also a. Respondentswhose more of members theunion willparticipate (see, for examthan otherrespondents ple, Niemi 1973). who b. Respondents havefriends in theunion will participatemorethan thosewho do not, given the greater potential for social reinforcement.6 c. The number of years as a memberof the union will be positively related to participation,reflecting the fact that the union itself may be a socializing agent.7 HYPOTrHESIS 5: A numberof attitudinal variableswill bepositivelyrelatedtoparticipation. with a positiveattitudetoward a. Respondents will thecivil rightsmovement participate (Verba and Nie 1972; morethan others Baxter and Lansing 1980; Andersen 1975). whoare satisfiedwith the barb. Respondents gaining processand thegrievanceprocess will participatemorethan thosewho are not (Perline and Lorenz 1970; Spinrad 1960). whoperceivethat union decic. Respondents sions are made democratically-that is, ratherthan indidirectlybythe members rectly by the leaders-will participate more than those who do not (Perline and Lorenz 1970; Spinrad 1960). d. Respondentswith a sense of efficacy8will participate more than those without (Verba and Nie 1972; Campbell 1960).
6We constructed this variable by analogy with the argument in the socialization literature in political science that family and schooling are strong socializing forces in the formation of political attitudes and particularly partisanship. 7In this sample, there was a very high correlation among age, seniority, and the number of years in the union. We therefore performed a number of statistical tests to verify that this covariance was due to the number of years in the union rather than to age or seniority.

we will hypothesize for the purpose of testing our data that blacks participate less than whites. Methodology and Data Our analysis will proceed in three stages. First, we will examine the correlations between race and individual participation indicators to assess whether there are racial differences in union participation, activity by activity. Second, we will compare the overall participation of blacks and whites, using a ten-point participation scale we have developed. Finally, since previous comparisons involve only simple correlations between race and participation indicators, we will present the results of a multiple regression that includes controls both for factors that have normally been included in models of participation (such as SES) and for those variables that are found to correlate most strongly with race in the first stage of analysis. The specific hypotheses to be tested are the following: 1: Blacks will participate less in HYPOTHESIS their unions than whites. HYPOTHESIS 2: Women will participate than less men. (Baxter and Lansing 1980; Sapiro 1984; Klein 1984.) HYPOTHESIS 3: Socioeconomic status(salaryand education) and some demographicvariables (seniority, skill, and ale) will be positively relatedto participation. HYPOTHESIS 4: Socialization and social reinforcementvariableswill have a positive effect

5The literature abounds with support for including SES as a predictor of participation (see, for example, Campbell 1960; Verba and Nie 1972). Higher salaries and greater seniority with the employing agent are predicted to increase participation based on the "stakes" hypothesis, which posits that the greater a union member's vested interest in outcomes on the job, the greater the chances he or she will participate in union affairs. The "stakes" hypothesis can be found throughout the extensive literature on union democracy (see, in particular, Perline and Lorenz 1970). Numerous studies have found that the more active union members are those who are highly skilled, more economically secure, more oriented toward their own ethnic group, and more satisfied with their union. Note that age and skill were later removed from the multiple regression because they added nothing to the model.

8The survey item that measured sense of efficacy was the question, "How much influence do you feel you have over decisions in the local union?" Four responses were possible: "a great amount of influence," "some influence," "very little influence," and "no influence." For the dichotomous version of this variable, "great" and "some" were coded as one and "very little" and "no" were coded as zero.

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 327 Survey The data were collected from a survey administered in the spring of 1980 to 2,000 randomly selected union members. All the members worked for the same government agency at different sites within the state of Illinois. All were represented by the same large, public sector union, which has locals across the state. Although a few militant locals of this union have engaged in strikes in the past, strikes in this particular sector are technically unlawful. The membership is predominantly semiskilled and skilled, white-collar, and with reasonably high levels of education. The response rate was 44.4 percent. Of the 888 questionnaires returned, 12 survey forms were eliminated because they were completed by managers. The resulting sample size was therefore 876 members from 1 14 local unions. The survey instrument contained approximately 90 items. Many of the items asked about participation in the following kinds offormal union activity:holding union office; voting in local, state, or international union elections; voting to authorize a strike; voting to ratify a contract; attending union meetings; and reading the union newspaper. Other questions probed more sporadic or occasional activities, termed here informal union activity, consisting of striking, picketing, or participating in any of the following union-sponsored activities: political, community, or recreation activities or training; and educational, health, or welfare programs. Other questions concerned respondents' attitudes toward the local,9 their political attitudes'0 their family socialization, the number of years they had been in the union, the number of grievances they had filed on their own behalf (rather than on the behalf of some'3Satisfactionwith the way the union handled their grievances; satisfaction with the way the union had negotiated to secure the most recent labor contract; perception of the extent of democracy (amount of power wielded by the members rather than the leaders) in the union; sense of efficacy; and confidence in their ability to gain local union office. '01dentification with a political party and attitudes toward the civil rights movement. "'This item asked whether a member of the respondent's family was a union member.

one else or of the bargaining unit as a whole), and their demographic characteristics. l Characteristics of the Overall Sample As members of a public sector union, the respondents in this sample may differ from a typical sample of union members in a private sector setting. The respondents in this sample probably had more education than the average industrial worker because most were white-collar workers.' And the sample's age distribution may slightly underrepresent young workers compared to the overall industrial population. 4 Of the respondents who indicated their race, 588, or 68.3 percent, were white, and 254, or 29.5 percent, were black. There were eleven Hispanics, one Asian, and five "others" (1.3, .1, and .8 percent of the sample, respectively). Women made up 39.3 percent of the sample. Thus, although the percentage of women was representative of their numbers in the general working population, blacks were overrepresented compared to the general working population, and particularly compared to workers in manufacturing. Furthermore, the sample does not contain as great a range in salary as exists in the industrial population.' We should therefore caution the reader that the study has limited external validity.
'2The demographic characteristics were race, gender, salary, education, and age. 3The educational distribution of the sample was: Years of Completed Education 7 8 9 10 11 12 13 14 15 16 17 Graduate Degree Number of Respondents 2 13 13 30 32 380 128 156 52 28 26 16 Percent of Sample .2 1.5 1.5 3.4 3.7 43.4 14.6 17.8 5.9 3.2 3.0 1.8

"Only 8.2 percent of the sample were under 26 years of age, and a full 45 percent were over 46. '5The minimum annual salary was $14,806 and the highest salary was $21,974. The average salary was $18,236.

328

INDUSTRIAL AND LABOR RELATIONS REVIEW
Table 1. Coefficients of Correlations Between Individual Characteristics and Race.
Coefficient No. of of ResponCorrelation dents

The most likely bias of this sample is that of self-selection: those who responded may have been over-representative of members who were most active in the union. We cannot rule out that possibility, but we believe any such bias is of minor significance. Seven percent of the sample reported they had "ever served" as local union president and 16 percent reported they had "ever served" as local union president, vice-president, secretary, or treasurer. We have no information on the actual numbers of both present and past office holders who are still members of the 114 local unions represented in our sample, but any reasonable allowance for turnover among local officers would suggest that the 16 percent of present and past officers in the sample is not sufficiently large to invalidate our analysis. 16 Characteristics of the Black and White Subsamples As indicated in Table 1, blacks did not differ significantly from whites with respect to seniority, age, education, salary, status as a parent, skill, whether they had friends in the union, or whether they had family members who were members of the union. 17

Variable
SENIORFl ya
AGE')

.05
-.04

844
851

EDUCAlIONC
SALARY PARENTAL. STATUSd
SKILLI,'

.03
.02 .02

846
727 727

GENDER'

.12 . 16**
-.29** .02

MARITALSTATUSg
FAMUMEMII

816 854 850
851

BUD)SUMEM
CONFIDENCE IN ABILITY TO GAIN LOCAL OFFICE CONFIDENCE IN ABILITY TO GAIN INTERNATIONAL OFFICE BAR(;AINING SATISFACTION GRIEVAN(;E SATISFACTION SENSE OF EFFICACY AMOUNT OF DEMOCRACY PARTY I1)ENTIFICATION' CIVIL RIGHTS ATTITUDE LIB.-CONS. ATTITUDE WOMEN'S EQUALITY ATTITUDEk k .06*
.05 .1 3**
- .02**

842
704 715 679 282 837 745 807 841 841 851

.16** .07**
.11 **

.31** .54** .17** .13**

Note: Pearlson's correlation coefficient is used in all IbAnother important reason for our confidence in the representativeness of the sample is the response rate of approximately 45 percent. Such a high response rate to a 12-page survey in which the respondents had no direct interest can be taken as some assurance that there is no large bias. In fact, the union had predicted a response rate of less than 5 percent, since only about 5 percent of the local union presidents responded when the international asked them for a list of their contract demands (a matter in which they obviously had a direct interest). '7The only one of these relationships that is significant is that between race and having friends in the union. We report it here with the other nonsignificant correlations because it is a relatively weak relationship. It should be noted that other studies show dramatic differences between blacks and whites in terms of salary, education, and sometimes seniority; and in the general population, of course, blacks and whites differ greatly in socioeconomic status. The explanation for the divergent finding here may be the employing agency's merit system, which has equalized both hiring and promotion opportunities for blacks and whites. This merit system may result in less discrimination than do other systems in many private employment sites, which have served to create or reinforce wage gaps between the races. cases except SKILL, for which Spearman's rho is used. 'SENIORITY:number of years seniority. bA(;Ewas measured by six categories: under 26, 26-

35, 36-45, 46-55, 56-65, and over 65.
'ED)UCA'I ION: years of education. dA positive coefficient indicates the respondent was a parent. ePearson's r = 12 with N = 816 and sig. = .001. 'Female = 1; male = 0. A positive result indicates that women exhibit more of a particular characteristic than men. 1A dichotomous variable: married and not married. Not married is composed of "separated or divorced," "widowed," and "never married." I'FAMUMEM: a positive sign indicates that a family member is a member of the union. iBUDSUMEN:a positive coefficient indicates having friends in the union. 'A positive coefficient indicates Democrat or Independent. In another version, in which, Republicans and Independents were grouped together, the relationship was slightly weaker (r = .25). kA high score indicates a liberal attitude. *Significant at the .05 level; **significant at the .01 level.

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 329 Neither did blacks differ significantly from whites in their confidence in their ability to achieve local union office. They were, however, more likely to be satisfied with the grievance process, although very few of the sample answered this question (N = 282). Blacks were more likely than whites to be dissatisfied with the union's handling of bargaining over contracts. Contrary to expectations, blacks were more likely to feel efficacious than whites were. They were also more apt to view union decision making as undemocratic, that is, to believe that decisions were made by the leaders rather than directly by the members. Blacks were much more likely to identify themselves as Democrats than were whites, and they were also more liberal in their civil rights attitudes, their attitudes to women's equality, and their general liberalconservative ideology.'8 Results for Individual Participation Indicators Differences in Leadership Experience We now turn to the question of whether blacks differ from whites in the amount or type of their participation in unions. Table 2 displays the differences in participation by race. The most traditional way of measuring union participation is by documenting service in formal leadership positions. Respondents to the questionnaire we employed reported whether they had served in their union as president, vicepresident, secretary, treasurer, delegate, steward, bargainer, trustee, or in some "other" leadership position."
'"The item for civil rights attitudes is a seven-point scale ranging from the response "Civil rights leaders are pushing too fast" to the response "Civil rights leaders are not pushing fast enough." The Pearson's correlation is .54, with N = 841 and sig. = .01. On a standard liberal-conservative self-designation scale, blacks also showed more liberalism than whites; the Pearson's correlation is .17, with N = 841, and sig. = .001. Blacks were also more likely to give a liberal response to the question concerning women's equality; the Pearson's correlation is .13, with N = 851 and sig. = .001. These three items are standard items taken from the National Election Study by the Survey Research Center, University of Michigan. "9Thewording for the leadership items was: "Have Table 2. Correlations Between Race and Different Types of Participation in Union Activities. Number of Correlation ResponCoefficient dents Holding LeadershipOffice - . 15** .11 ** .13*5 -.04 -.09* -.08 -.03

Variable

President

855 855 855 855 853 855 855

Vice-President Secretary Treasurer Bargaining Committee Member Trustee Other

Engaging in Formal Union Activities Voting in Local Union Elections .02 Voting in State Union Elections - . 12* Voting in International Union Elections -.04 Reading the Union .03 Newspaper - .08* Voting to Authorize Strike Voting to Ratify ConLract - .12'* Attending Meetings Political .02 Union Activities Engaging in Irifiormial -..04 Training -.05 Education .01 .03 Picketing Recreation Health, Education, and Welfare
Commtinity

811 662 744 851 839 769 841 833 833 831 831 835 836 837 694

.03 .07 I* .05X' .20* i

Strike

Note: The race variable is dichotomous, with black being positive and white negative. Hence, a negative sign on the coefficients indicates greater participation by whites by blacks. *Significantt at the .05 level; **significant at the .01 level (one-tailed tests).

you ever held a union office or served on a union committee?" Respondents who answered yes were asked how many years they had served in each listed office.

330

INDUSTRIAL AND LABOR RELATIONS REVIEW Once again, as shown in Table 2, the results are mixed. There was no statistically significant difference in the frequency with which blacks and whites participated in five of the informal union activities: political activity, training programs, educational programs, picketing, and recreation programs. There was a significant difference in the amount of their participation in the three remaining activities-strikes; community activity; and health, education, and welfare programs-and in all three cases, blacks were morelikely than whites to engage in the activities in question. Strike activity is, however, the only activity that shows a strongly significant relationship. To summarize the results in Table 2, blacks participated significantly less than whites in two of the seven leadership positions and in three of the seven formal union activities; and they participated significantly less in none of the eight informal union activities. In fact, in three out of the eight informal activities, they participated more than whites. One noteworthy pattern is the increase in black participation as the focus shifts from the leadership positions to the other formal activities, and again as it shifts from the formal activities to the informal activities. It should be noted, however, that there is a fair amount of variation in black participation across the individual activities in each of these categories. Of most interest in these findings is the fact that the black-white differences are not nearly as strong as the previous literature would have predicted. There is much variation by individual activity; many of the differences are only technically significant; and those differences that are strongly significant are few and far between.20 In light of that pattern, and given the large sample size, we believe it is safest to conclude that the average black member participates in most union activities as often or nearly as often as his or her white counterpart.

The results show that blacks were significantly less likely than whites to have served at one time or another as president or as member of the bargaining committee, but significantly more likely to have served as vice-president or secretary. Although blacks held offices in the trustee, treasurer, and "other" categories less often than whites, the magnitude of those differences is nearly zero. Thus, the outlook varies office by office, with black participation being significantly less for only two offices. The most important finding is that the office blacks were least likely to have attained is the critical one of local union president. Since that office is an essential stepping stone to the international staff, this finding may help to explain why blacks continue to be under-represented among the offices of most international unions. Differences in Other Avenues of Formal Participation The other traditional measures of union participation are voting in local, state, and international union elections; voting to ratify the contract; voting to authorize a strike; attending union meetings; and reading the union newspaper. Returning to Table 2, we again find mixed results. Blacks participated significantly less than whites in three of these seven activities: state-level elections, strike votes, and contract votes. For all the other activities, the differences found are not statistically significant, though they favor (with the exception of the international elections) black participation. Differences in Informal Participation Informal forms of participation-those that are sporadic or occasional-have seldom been examined in previous studies of union participation. Some of these activities, such as picketing, striking, and political action, may be central to the union's mission. Specifically, the questionnaire asked about participation in the following union-sponsored activities: political activity; training program; educational programs; community activity; health, education, and welfare programs; picketing; and strike activity.

2OThe strongest correlation in Table 2 is .20 for strike activity. The next strongest is - .15 for race and holding local union office of president.

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 331 Results for a Composite Scale of Participation In the preceding section, we examined each activity singly. To better assess the overall participation of blacks and whites, we constructed a composite scale of union member participation. The Scale A ten-point scale was used, comprising ten items, each assigned a weight of one point. Three of the activities on the scale were formal activities, six were informal activities, and one was leadership activity. Using established techniques for scaling, we checked for multi-dimensionality of the scale, and found that all ten items appeared to represent one underlying dimension.' In choosing the activities from Table 2 to include in the scale, we eliminated the following six because of questionable validity or a low item-to-total correlation: voting in state union elections, voting in international union elections, voting to authorize a strike, strike activity, reading the union newspaper, and union-sponsored recreational activity.22 union meetings; and the six informal activities: participating in various unionsponsored activities-political activity, educational programs, training programs, health, education, and welfare programs, and community activity-and picketing.23 All ten activities were represented by dichotomous variables, with a positive response equal to one and a negative response equal to zero.24 The Results The Pearson's correlation for race and the composite participation scale was - .05, with N = 690 and sig. = . 10. We therefore find no difference between blacks' and whites' overall participation in their local union. Further, when we added the SES variables of salary and education to the model, race still was not significant. The betas for the standardized multiple regression were: -.04 for race, .17 for salary (significant at the .05 level) and .14 for education (significant at the .05 level). (The R2 for the equation is .05 and N = 615.) These results are surprising. Theory would predict either less participation by blacks, based on their demographic characteristics, or more participation by blacks, based on their group consciousness (when SES is controlled).
23The overall alpha of' the scale was .73, and the standardized item alpha was .75. The item-to-total correlations for the scale items are available upon request to the authors. Although the two voting variables and the community activity variable behaved slightly differently from the others, the items generally formed a cohesive single scale. The results of the factor analysis revealed that Factor 1 explained 65.5 percent of the variance and had an Eigen value of 2.65, whereas Factor 2 explained only an additional 21.8 percent of the variance and had an Eigen value of .88. 24The voting items were already dichotomous. Union meeting attendance had to be collapsed from a double-digit absolute number to a dichotomous variable. If the respondent reported having attended one or more meetings in the previous year, his response was coded as one. If the respondent reported nonattendance, his or her response was coded as zero. The item querying reading the union newspaper originally had four possible responses: regularly, sometimes, seldom, and never. A response of "regularly" was coded as one and the other responses were coded as zero for the purpose of making this item dichotomous for use in the scale.

The remaining activities included in the scale were the one-point item for leadership (a single dichotomous variable indicating whether or not the respondent had ever served in one of the local union positions listed at the top of Table 2 or on any committee in a local union); the three formal activities: voting in local elections, voting to ratify the contract, and attending
21All of these measures covaried sufficiently to form a single scale. Tables exhibiting the correlation matrix and the results of a varimax factor analysis are available upon request to the authors. 22The three items of questionable validity, due to a large number of missing data, were voting in state elections, voting in international union elections, and voting to authorize a strike. Specifically, there were 196 missing responses for the first of those activities; 114 missing responses for the second; and 266 missing responses for the third, 250 of which were for respondents who indicated that they had not been union members at the time of the last strike authorization vote. The three items eliminated because of low item-to-total correlations were strike activity (.19), reading the union newspaper (. 17), and unionsponsored recreational activity (.17).

332

INDUSTRIAL AND LABOR RELATIONS REVIEW
Table 3.

We also tested the model with other controls: those attitudinal and socialization variables that were significant in their simple correlation with overall participation, and other control variables normally included in such models, such as gender. The attitudinal controls included were a sense of efficacy, confidence in the ability to achieve union office, and civil rights attitudes; the socialization controls included were having friends in the union and the number of years of membership in the union. As shown in Table 3, the final model includes these five additional and socialization controls, as well as the two SES controls, salary and education. Even when these factors are included, the differences between blacks' and whites' participation remain insignificant. There are several possible explanations for this unexpected finding. The first is that the finding of previous studies-that blacks participate less than whites-may have resulted from examining samples of limited size and ignoring the more commonplace avenues through which the average union member participates in his or her organization (our "informal" activities). In addition, other studies have tended to focus on leaders, not, as this study does, on the average member. This study's finding of no significant difference may therefore hold in other settings. Another possible explanation is that the collinearity between civil rights attitude and race is depressing the amount of influence race appears to have. We retested the model to check for this. Yet again, the race effect was nonsignificant. One qualification is in order. As acknowledged before, the blacks and whites in this sample do not differ from each other in ways, such as salary and education, that distinguish blacks and whites in the larger U.S. population. Our results therefore intimate that in settings where blacks and whites are similar, there is no reason to predict that blacks will participate less than whites. Where they are different demographically, there may be differences in participation; but any analysis of such a setting will need to control carefully for socioeconomic status differences.

StandardizedRegressionCoefficients for the Determinantsof Union Participation, as Measuredby the Composite Scale, for the Total Sample, for BlacksOnly, and for Whites Only.
Total Sample
-.05
-.04 -.03 -.04

Variable
RACE GENDER SALARY EDUCATION FRIENI)SHIPS IN UNION SENSE OF EFFICACY (CONFII)ENCE IN ABILITY TO GAIN UNION OFFICE CIVIL. RIGHTS ATTITUI)E YEARS AS MEMBER OF THE UNION ADIUS'Ei) RR2

Blacks Only

Whites Only

.07* .12* .18* .11*

.01 .06 .07 .14

.09* .16* .21* .10*

.46* .05* .15* .37 697

.25* .11 .27* .16 158

.51* .02 .10* .46 511

N

*Significant at the .05 level.

Notwithstanding this caveat, our finding is still, we believe, a convincing one, for two reasons. First, the simple (or uncontrolled) correlation between race and participation is nearly zero. Second, that correlation is not significantly different from zero even when the controls are added. It appears, then, that race, when compared to other factors, is not an important predictor of participation. Relative Importance of Factors Salary, education, and the feeling of efficacy or influence over local union matters are all positively related to participation of all members. But the two strongest predictors of participation are having friends in the union and having confidence in one's ability to gain a local union office (see Table 3). The finding on confidence in the ability to gain office has several implications. First,

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 333 it suggests that the motivation to participate is somewhat rational. People participate when they expect they can make a difference, rather than because of the intrinsic value of participation. Second, it follows that the local union setting itself is critical. If the local union setting prevents or discourages members from gaining office, it will discourage them from participating; if it offers open access, it will encourage participation. If this conclusion is correct, then the extent of members' participation must depend critically on their local union's distinct political culture. How Does Black Participation Differ from White Participation? To investigate the differences between the correlates of black participation and those of white participation, we applied the composite scale to the responses of the two groups separately. The results for this analysis are also displayed in Table 3. The factors found to covary with black participation are very different from those found to covary with white participation. For example, five of the seven predictors that were found to be significant for the total sample-salary, education, friendships in the union, a sense of efficacy, and civil rights attitude-are nonsignificant for the blacks-only group. In addition, the overall predictive power of the model is lower for blacks than for whites, as indicated by the lower R' for blacks. One factor-the number of years as a union member-is much more predictive of black participation than of participation by the total sample, although it is also predictive for the total sample. The effect of this variable supports the hypothesis that the union serves as a socializing agent. Over time, blacks may grow more comfortable with their union's goals or with their union as a social setting, or both. the differential effect of union socialization on blacks and whites can be explained by the role of blacks as a minority or "outgroup" within the local union. Whites may feel more comfortable from the outset about participating in their local union. Blacks, however, may have previously experienced barriers to participation in unions as well as in other institutions. Thus, they may initially be reluctant to "take on" the institution, either for group gain or for personal gain, by, for example, running for office or attending meetings. But as time goes by and they become established in the local, blacks may begin to see the potential benefits to be won by their participation. Friendships in the union, on the other hand, are much less important to blacks than to whites. It is reasonable to assume, by way of explanation, that the social pressure to participate "because my friends do" would be less among the minority group than among the majority group. Conclusion and Implications The findings of this study are encouraging in light of the questions of black enfranchisement we raised at the outset of this paper. Our most striking, if not startling, finding is how much blacks actually do participate, at least in the union studied here. We did, however, find some blackwhite differences. Blacks participated more than whites in some activities and less than whites in others; and the attitudinal and socioeconomic correlates of black participation were different from the correlates of white participation. There are several important implications of the findings for the individual activities. First, and most important, we believe the prudent conclusion to draw here-based on the mixed results and the relatively low that levels of differences we found-is blacks and whites exhibited little difference, overall, in their participation in these individual activities. Second, of the differences that were found, blacks were somewhat more apt than whites to participate in the informal activities and somewhat less apt than whites to participate in the formal activities. It may be that as formerly disenfranchised members tentatively begin to participate in their unions, they do so first in the less frequent, less formal activities of the union and only later, as they begin to reap benefits from those activities, move on to participate more formally by voting and holding office.

334

INDUSTRIAL AND LABOR RELATIONS REVIEW salary, education, and having friends in the union as white participation is. Analytically, it certainly appears that the factors that motivate participation are different enough between the two racial groups to warrant further study. In particular,research to assess the conditions under which parity of participation exists would be an important contribution to our understanding of this subject. It would also be useful, of course, to determine whether our results hold across a variety of workplaces and unions. In conclusion, our finding of near parity in the extent of participation of black and white union members is one step toward resolving the discrepancy in findings between the political science literature and literature on blacks in unions. Considering the evidence of ongoing discrimination in unions, our result is indeed startling. Whether that result reflects the use of more comprehensive measures of participation than those used in earlier union democracy studies-and the choice of the union member, rather than the union leader, as the unit of analysis-is a question for future research to decide. The question of how equal participation by black and white members affects the union's actual policy making and functioning also remains to be addressed. We hope that the research design employed here will be useful to future researchers in answering both of these questions.

Third, one of the strongest coefficients in Table 2-that showing fewer blacks than whites holding the office of local union president-clouds the otherwise encouraging picture painted by these results. As blacks do participate more in the local union, they may be running into the same barriers they face in corporations and in politics, where whites acquiesce to the participation of minorities and women in midlevel functions and offices but resist allowing them access to the key positions of power: the presidency, membership on the board of directors, or presence in the halls of the U.S. Senate. Our results for the composite scale of participation echo the results for the individual measures, indicating, in fact, no difblack and white ference between participation. That result holds even with the addition of controls for socioeconomic and attitudinal variables. Nonetheless, it does appear that the correlates of black participation are different from those of white participation. One factor that may motivate blacks more than whites is the number of years as a union member. Blacks' feeling of powerlessness as an underrepresented minority in unions may ameliorate with time, thereby reducing the distrust and alienation that stymie their participation. In addition, black participation does not tend to be as dependent on such factors as

REFERENCES Aberbach, Joel, and Jack Walker 1970 "The Meanings of Black Power: A Comparison of White and Black Interpretations of a Political Slogan." American Political Science Review, Vol. 64, pp. 367-88. Andersen, Kristi 1975 "Working Women and Political Participation, 1952-1972." AmericanJournal of Political Science, Vol. 19, No. 3, pp. 439-54. Baxter, Sandra, and Marjorie Lansing 1980 Women and Politics: The Invisible Majority. Ann Arbor: University of Michigan Press. Campbell, Angus, Philip Converse, Warren Miller, and Donald E. Stokes The American Voter. New York: Wiley. 1960 Fulenwider, Claire Knoche Feminism in AmericanPolitics: A Study of Ideo1980 logical Influence. New York: Praeger. Gould, William B. Black Workersin White Unions:Job Discrimina1977 tion in the United States. Ithaca, N.Y.: Cornell University Press. Gross, James A. "NAACP and the AFL-CIO: An Overview." 1962 Negro HistoryBulletin, December.

PARTICIPATION IN LOCAL UNIONS: BLACK AND WHITE MEMBERS 335
Hagburg, Eugene C., and Harry R. Blaine 1967 "Union Participation: A Research Note on the Development of a Scale." Industrial and LaborRelationsReview, Vol. 21, No. 1, pp. 9296. Hill, Herbert 1977 BlackLaborand theAmericanLegal System: Race, Workand the Law. Washington, D.C.: Bureau of National Affairs. 1975 "Black Labor, the National Labor Relations Board and the Developing Law of Equal Employment Opportunity." Labor Law Journal, Vol. 26, No. 4, pp. 207-23. Klein, Ethel 1984 GenderPolitics. Cambridge: Harvard University Press. Kornhauser, William "The Negro Union Official: A Study of Spon1952 sorship and Control." AmericanJournalof Sociology, Vol. 57, No. 5, pp. 443-52. Lamm, Roger 1975 "Black Union Leaders at the Local Level." Industrial Relations, Vol. 14, No. 2, pp. 22032. Lipset, Seymour M., Martin A. Trow, and James S. Coleman 1956 Union Democracy.Glencoe, Ill.: Free Press. Magrath, C. Peter 1959 "Democracy in Overalls: The Futile Quest for Union Democracy." Industrialand LaborRelalions Review, Vol. 12, No. 4, pp. 503-25. Miller, A. M., P. Gurin, G. Gurin, and 0. Malancuk 1981 "Group Consciousness and Political Participation." American Journal of Political Science,

Vol. 25, No. 3, pp. 494-511. Miller, Art 1974 "Political Issues and Trust in Governimnent: 1964-1970." AmericanPolitical ScienceReview. Vol. 68, No. 3, pp. 951-72. Niemi, Richard 1973 "Political Socialization." In Jeanne N. Knutson, ed., Handbook of Social Psychology. San Francisco: Jossey-Bass. Perline, Martin, and V. R. Lorenz
1970 "Factors
Influencing

Participation

in Trade

Journal of Economy Union Activities." American & Society, Vol. 29, pp. 425-37.
Sapiro, Virginia

Political Integration of Women. Urbana: University of Illinois Press. Schwartz, Arthur, and Michele Hoyman 1984 "The Changing of the Guard: The New American Labor Leader." In Annals of the AmericanAcademyof Political and Social Science, Vol. 473, pp. 64-75. Spinrad, William 1960 "Correlates of Trade Union Participation: A Summary of the Literatuie." American Sociological Review, Vol. 25, No. 2, pp. 237-44. Strauss, George 1977 "Union Governmeint in the U.S.: Research Past and Future." IndustrialRelations, Vol. 16, pp. 215-45. Verba, Sidney, and Norman Nie and 1972 Political Participationin America:Democracy Social Equality. New York: Harper & Row.

1984