Social Confidence in Unions: A U.S.-Canada Comparison

Social Confidence in Unions: A U.S.-Canada Comparison

LORENZO FRANGI

University of Quebec in Montreal

MARC-ANTONIN HENNEBERT

HEC Montreal

VINCENZO MEMOLI

University of Catania

Cet article examine l’evolution des niveaux de confiance sociale enversles organisations syndicales au Canada et aux Etats-Unis entre 1982 et2006 en s’appuyant sur les donnees de la World Value Survey (WVS).Partant du constat que ces taux de confiance sont relativementsimilaires au sein de ces deux pays, nous appliquons un modele deregression logistique sur les deux vagues de donnees les plus recentes duWVS (soit celle de 2000 et 2006) afin de mieux comprendre la propensiondes individus a faire confiance aux syndicats dans chaque contextenational en fonction de certaines de leurs caracteristiques sociales etpolitiques. Les resultats montrent des similitudes entre les deux pays(par exemple, les citoyens les plus progressistes et les plus jeunes ontgeneralement une plus grande confiance envers les syndicats), maisegalement des distinctions importantes (a titre d’exemple, seul auCanada les individus repondant a un profil de “classe ouvriere”manifestent une plus grande confiance envers les syndicats).

This article examines changes in levels of social confidence in unions inCanada and the United States between 1982 and 2006 based on ananalysis of the World Value Survey (WVS) data set. It considers whyconfidence rates are similar in the two countries, applying a logisticregression model to the two most recent WVS waves (i.e., 2000 and 2006)so as to bring out the effects of political and social differences on thepropensity of individuals to trust unions in each national context. The

Lorenzo Frangi, School of Management, Organisation and Human Resources Department, Case postale8888, Succursale Centre-ville, Montreal (Quebec) H3C 3P8. E-mail: [email protected]

C© 2014 Canadian Sociological Association/La Societe canadienne de sociologie

Social Confidence in Unions 171

results show similarities between the two countries (e.g., moreprogressive and younger citizens generally have greater confidence inunions), but also important distinctions (e.g., only in Canada doindividuals with a working class profile appear to have greaterconfidence in unions).

UP TO THE 1990s, public opinion analyses helped enhance understandingof the nature of the relationship between unions and society. However,this type of analysis has since been all but abandoned (Rouillard 2009).Our analysis helps revitalize it by studying “social confidence in unions,”an indicator that, with the exception of a few recent studies (Frangi andMemoli 2013, forthcoming), has been largely unexplored in the debate onunionism.

Our analysis specifically compares social confidence in unions in theUnited States and Canada, an indicator providing important insights inthe enduring debate over how unionism in these two countries differs.Based mainly on the evidence of divergent union density rates (i.e., a dra-matic downturn in the United States since the mid-1960s versus a slightdecrease in Canada), many comparative studies have indicated that union-ism is going through a deeper crisis in the United States than in Canada(e.g., Godard 2009; Kumar 1993). However, as demonstrated by Lipsetand Meltz (2004), the social support underlying unionism is much morecomplex than narrow interpretations of union density might suggest. Infact, Lipset and Meltz (2004) specifically demonstrated that union approvalrates among employees do not appear to differ greatly between the UnitedStates and Canada. An analysis of social confidence in unions that consid-ers the opinion of society at large rather than that of employees exclusivelycan thus shed further light on this debate.

Section “Confidence in Unions and Trust” of the article draws on thedebate regarding confidence in institutions to highlight the most impor-tant mechanisms that structure confidence in unions, with a specific focuson the individual’s inclination to trust “others” and the government. Sec-tion “Confidence in Unions and Values in Canada and the United States”describes some fundamental social and individual values in the UnitedStates and Canada that can influence confidence in unions. Section “Dataand Method” presents the data and the logistic regression model used totest the hypotheses. For each country, confidence in unions is thereforeexamined in light of the level of trust and the political orientation andsociodemographic characteristics of the respondents.

CONFIDENCE IN UNIONS AND TRUST

Confidence in unions refers to the relationship of trust between citizensand unions in a given society. As is the case for confidence in many other

172 CRS/RCS, 51.2 2014

institutions (e.g., Kasperson, Golding, and Tuler 2005; Tonkiss and Passey1999), confidence in unions is both highly situational—influenced by thespecific features of unions in a society—and dependent on each individual’sinclinations.

Citizens have confidence in an institution—in this case, unions—insofar as they perceive that it does not engage in opportunistic behav-ior and that their own interests are encapsulated in those pursued bythe institution (Hardin 1999). Some scholars maintain that confidence isthe outcome of the interaction between essentially three different mech-anisms: cognitive, emotional, and identity-based (Laplante and Harrison2008). The cognitive mechanism connotes that the individual’s confidenceorientation is based on a rational calculation of the impact of institu-tional performance on the well-being of the individual or of larger groups(Newton and Norris 1999). The emotional mechanism refers to the individ-ual’s emotional response to specific institutions, while the identity-basedmechanism concerns the perception of shared values.

These three mechanisms interact at different relational levels. In fact,confidence can be structured on a direct personal relationship with the ex-ponents or members of an institution, on a relationship mediated throughother trusted persons, or simply on distant relationships built on opinionsbased on institutional reputation and mass-media information (Warren1999).

As pointed out in many studies on confidence in institutions, confi-dence in unions cannot be examined without considering its relationshipwith the individual propensity to trust “others” (e.g., Lewis and Weigert2012; Putnam 1995).1 As suggested by the empirical evidence from thesestudies, the propensity to trust others is a central inclination that struc-tures confidence in institutions. Moreover, one institution can benefit fromconfidence in another institution perceived by individuals, at least in someinstances, as belonging to the same social sphere. Recent studies involvinga set of different institutions have shown that, in several societies, con-fidence in government has considerable positive effects on confidence inunions (Frangi and Memoli 2013, forthcoming).

In line with these arguments, we hypothesized that individuals whowere more trustful of “others” and individuals who expressed greater confi-dence in government, regardless of which society they belonged to (Canadaor the United States), would also show a greater tendency to express con-fidence in unions (Hypothesis 1).

1. In the wide debate on “trust,” this type of trust is referred to as “generalized trust” (e.g., Galeotti andZizzo 2012; Iglic 2010; Yamigichi and Yamigichi 1994). In public opinion surveys, “generalized trust”is usually surveyed using the question, “Generally speaking, would you say that most people can betrusted or that you need to be very careful in dealing with people?” Despite some criticism, this questionhas been confirmed to provide a good measure of confidence in strangers in different societies (Warren1999).


Considering that social confidence in institutions—in this caseunions—is highly situational, the next section of this article explores somefundamental social and individual values in American and Canadian soci-eties that can influence confidence in unions.

CONFIDENCE IN UNIONS AND VALUES IN CANADAAND THE UNITED STATES

Unionism has evolved in the context of different social values in Canadaand the United States. Differences in social values between the two coun-tries can be traced back to their origins. Lipset (1990) has long emphasizedthat the United States was born of a liberal revolution where individual-istic and laissez-faire traditions prevailed, which explains why no visiblesocialist or labor party has ever emerged there (Lipset and Marks 2000).Canada, on the other hand, had “Red Tories” and a decidedly Europeanconservative lineage, which provided rooted statist and social democratictraditions (Taylor 1982).2 The United States is a characteristically capi-talist or libertarian society, where property and ownership rights and theinterests of capital have always prevailed over collective rights. Canada,on the other hand, has been described as a socialist or social democraticsociety, where the state and collective authority have traditionally engen-dered a stronger sense of obligation (Alston, Morris, and Vedlitz 1996).Support for social democratic institutions in Canada was reinforced in theearly 1960s, in large part due to the influence of Quebec’s Quiet Revo-lution. These historical sociopolitical values have resulted in a balancebetween individualism and collectivism in Canada (Taras, Ponak, andGunderson 2001) while a “possessive individualism” (Macpherson 1964)of Calvinist tradition has largely prevailed in the United States. This hasresulted in a value environment that is more conducive to supporting col-lective rights and worker representation in Canada than in the UnitedStates. From a cross-class perspective, Canadians therefore have more so-cial democratic values than Americans, while the latter prefer freedomover equality (Lipset and Meltz 2004).3

2. In post-Confederation Canada, the “Red Tories” and the “Blue Tories” represented two wings of theProgressive Conservative Party of Canada. More specifically, Red Tories were partisan conservativeswho espoused progressive economic and social ideas and supported the Welfare State. The Red Toriesadvocated moderate economic liberalism combined with varying degrees of social progressivism andhistorically exerted a significant influence within the Progressive Conservative Party. However, theparty’s merger with the Canadian Alliance in 2003, forming the Conservative Party of Canada, changedthe balance of power between Red and Blue Tories within the new party. The Red Tories lost a greatdeal of influence and some of them decided to quit the Conservative Party or even join the Liberal Partyof Canada.

3. According to Lipset and Meltz (2004), Canadians “clearly hold more statist and communitarian values,which are congruent with Tory-social democratic politics and collectivist or group-oriented (mosaic)orientations, while Americans exhibit more antistatist, individualistic and competitive (meritocratic)values, which imply laissez-faire politics” (p. 78). The impact of these different value systems on support

174 CRS/RCS, 51.2 2014

Moreover, a closer look at key industrial relations protagonists showsthat the two societies have different employer attitudes toward unionsand different working class values. American employers are generallystrongly opposed to unions (Fligstein 2001; Freeman and Medoff 1984;Gomez, Lipset, and Meltz 2001). This hostility—which, according to somescholars, has no comparable counterpart in Canada—has also been drivenby extensive media campaigns aimed at portraying labor unions as cor-rupt, undemocratic, unduly powerful, and the major cause of inflation,influencing social perceptions of the legitimacy and usefulness of unions(Kumar 1993). Differences in working class values between Canada andthe United States are even more relevant. In the United States, the work-ing class is fragmented and conservative, and concerned primarily witheconomic gain. The political system, which sustains the dominant po-sition of employers, is never queried (Friedman 1998; Jacoby 1991). Incontrast, the Canadian working class has been more interwoven withsignificant left-wing components and social progressive forces (Palmer1992).

The higher historical working class closeness to unionism—in termsof values and union density—in Canada may be one important underly-ing factor explaining the determinants of social confidence in unions inthe two countries. More specifically, in line with our earlier arguments,we expected that Canadian unions would enjoy greater confidence amongrespondents with a working class profile. In other words, we hypothesizedthat in Canada, citizens with lower incomes and from blue-collar occupa-tions would be more likely to have confidence in unions than other citizens.On the other hand, we expected that these variables would not be signif-icant in terms of discerning between American citizens who expressedconfidence in unions and those who did not (Hypothesis 2).4

Furthermore, confidence in unions can also be defined by individualvalues. Since one substantial difference between the United States andCanada is the political involvement of unions, this point and its possiblerelationship with individual political values are relevant.

In fact, the development of Canadian unionism has been characterizedby more left-wing challenges and left-leaning union leaders (Adams 1995).Canadian unions tend to be more involved politically than their Ameri-can counterparts essentially because of their proximity to the New Demo-cratic Party, founded as the Canadian labor party in 1961 (Goldfield and

for unionism, however, remains the subject of much debate. Several studies have pointed out in thisregard that cultural and values-related dimensions do not appear to have a significant effect on unionmembership (Bruce 1989) while others have shown that differences in Canadian and U.S. values donot explain the variation in unionization rates between these two countries (Lipset and Meltz 2004).

4. It would have been interesting to consider the role played in this relationship by the respondent’s sectorof occupation (i.e., private vs. public). However, while this variable (variable label in the database: x052)was included in the 2006 wave, it was not included in the 2000 wave (i.e., the two waves considered inthis article).


Palmer 2007). In Quebec, this kind of political relationship could also beobserved during the 1970s between the Parti Quebecois and provinciallabor organizations. As pointed out by Grant (2003), a significant num-ber of union officers were party activists, even though there has neverbeen a formal link between the Parti Quebecois and Quebec labor orga-nizations. These political ties between unions and the New DemocraticParty (and the Parti Quebecois in Quebec) have enhanced the proximityof Canadian unions to the political arena. Moreover, thanks to their polit-ical involvement, union leaders do not limit their agency to the firm, butrather promote more far-reaching social and political change as an inte-gral part of union activity (Rose and Chaison 2001). These ties betweenthe union movement and the political sphere, which have marked the de-velopment of unionism in Canada, have to some extent weakened in recentyears.5

In the United States, on the other hand, a business unionism modelhas historically prevailed, and unions have not developed strong rela-tions with the political arena. More specifically, unions in the UnitedStates have never created their own labor or social democratic party,nor have they supported one (Lipset and Meltz 2004). Thus, not onlyhave union leaders rejected any form of partisan politics, but unionshave also become anathema to exponents of nonconservative politicalideology (Sinyai 2006). Union goals have remained limited to the pro-tection and advancement of the economic interests of union members,rather than the pursuit, through political action, of broader reforms forthe larger working class or citizens in general (Friedman 1998; Jacoby1991).

This reasoning led to our third hypothesis, which suggests that, be-cause Canadian unions have historically been more involved in the politicalarena and political parties than American unions, interest in politics wouldbe a significant factor when it came to discerning between citizens who ex-pressed confidence in unions and those who did not in Canada, but not inthe United States. Similarly, due to the stronger links between Canadianunions and left-wing parties versus the greater distance between unionsand party dynamics in the United States, we expected that a progressiveversus conservative personal political orientation would significantly con-tribute to explaining confidence in unions among Canadians but not amongAmericans (Hypothesis 3).

5. The traditional relationships between the political sphere and the union movement in Canada indeedappear to be more tenuous these days. Moreover, a recent body of literature on union renewal offers ananalysis of the changing political ground of the Canadian union movement (Savage 2010). This analysispoints out, in particular, that both the Parti Quebecois and the National Democratic Party have shiftedto the right of the political spectrum, thus further weakening their ties with union organizations(Camfield 2011). The rise of neoliberalism and the at times openly antiunion policies of both the federalgovernment and several provincial governments also appear to have forced unions to “retreat” andrevise their strategies of support for some political parties (Allen 2006; Panitch and Swartz 2003).

176 CRS/RCS, 51.2 2014

DATA AND METHOD

In order to test the hypotheses, we used the World Value Survey (WVS)data set.6 Even though public opinion surveys around the world haveshown marginal interest in measuring public support for unions since the1990s (Rouillard 2009), the WVS collected information about confidence inunions until 2006, not only among workers, but also in broader society. Inorder to provide more recent insight and to consider a wider set of indepen-dent variables regarding the sociopolitical values and sociodemographiccharacteristics of the respondents (since some of these were not includedin earlier WVS waves), we focused on the last two WVS waves (study yearsfor Canada: 2000 and 2006; for the United States: late 1999 and 2006). Therepresentative sample for the last two waves comprised 1,928 respondentsin Canada and 1,449 respondents in the United States.7

The dependent variable, that is, confidence in unions, was consideredto be dichotomous: on the one hand, respondents who expressed “a greatdeal” or “quite a lot” of confidence in unions and, on the other hand, thosewho expressed “not very much” confidence in unions or “none at all.” Sincethe dependent variable was dichotomous, a binary logistic regression modelwas applied.

In order to test the hypotheses, a set of independent variables wasconsidered. More specifically, in order to test the first hypothesis (trust in“others” and trust in government are positively related to confidence inunions in both countries), the model included “trust in others” (dichoto-mous: 0 = cannot be too careful; 1 = most people can be trusted) and“confidence in government” (0 = not very much + not at all; 1 = a greatdeal + quite a lot).

In order to test the second hypothesis (working class profile has a sig-nificant impact on confidence in unions in Canada but not in the UnitedStates), the model incorporated variables related to the respondents’ cur-rent (or last, if any before the interview) occupation and self-positioningon a 10-step income scale.

In order to test the third hypothesis (citizens who are more inter-ested in politics and more progressive have greater confidence in unionsin Canada but not in the United States), the model included a variablemeasuring interest in politics and three variables used to differentiatebetween citizens with a progressive versus conservative political orienta-tion. The latter three variables were measured based on the respondents’self-positioning on 10-step scales: left-right ideology (1 = left; 10 = right),views on income inequality (income should be made more equal = 1, we

6. WVS (www.worldvaluessurvey.org) is a data set on sociopolitical attitudes and values implemented bya worldwide network (more than 50 countries) of social sciences researchers.

7. For further information about the sampling techniques adopted by the WVS in each country, seehttp://www.wvsevsdb.com/wvs/WVSDocumentation.jsp.


need larger income differences as incentives = 10), and views on privateversus state ownership (private ownership of business should be increased= 1; government ownership of business should be increased = 10).

Last, we evaluated the effects of sociodemographic variables. In addi-tion to age, gender, and highest level of education attained, we includedthe language spoken in the respondent’s home. This was particularly im-portant in Canada, in order to evaluate the impact of the French- versusEnglish-speaking tradition regarding confidence in unions (0 = English;1 = French). In the United States, this variable provided indications onthe difference between immigrants and other Americans (0 = English; 1 =other languages). Moreover, drawing on the responsiveness concept, Pow-ell (2005) and Memoli (2013) argue that there is a consistent relationshipbetween personal satisfaction and confidence in institutions. We thereforeincluded general satisfaction with life and household economic satisfactionas control variables.8

RESULTS AND DISCUSSION

Descriptive Analysis

Some important findings emerged from the descriptive analysis of confi-dence in unions in Canada and the United States (Figure 1), our dependentvariable. The percentage of citizens who had confidence in unions wasquite similar in both countries (between 32 and 35 percent in Canada,and between 30 and 39 percent in the United States), and had declined inboth countries between the two last WVS waves. These levels of confidencewere different from the union density rates, and markedly so in the UnitedStates.9

Moreover, the levels of social confidence in unions were also quitesimilar to those found in most European countries. More specifically, theoverall confidence level in both the United States and Canada was lowerthan in Nordic European countries, but not substantially different fromthat in continental European countries; it was nevertheless higher than insome Mediterranean countries, where the unions are considered to have ahistorically higher impact on labor market dynamics than in North Amer-ica (e.g., Hall and Soskice 2001).10

8. These two variables, which were originally cardinal in the data set, were recoded as follows: 0 =dissatisfied, 1 = satisfied.

9. Union density in Canada was 28.3 in 2000 and 27.5 in 2006, and in the United States was 13.4 in 1999and 11.5 in 2006 (OECD data [http://stats.oecd.org/Index.aspx?DataSetCode=UN_DEN]).

10. According to WVS data, the level of confidence in unions is as follows: in Scandinavian countries:Finland (51 percent in 1996, 53 percent in 2000, and 62.6 percent in 2005); Sweden (44.4 percent in1996, 42.5 percent in 1999, and 52.3 percent in 2006); in continental European countries: Germany(37.6 percent in 1997, 37 percent in 1999, and 32.1 percent in 2006); France (34.1 percent in 1999and 38.7 percent in 2006); in Mediterranean countries: Italy (28.7 percent in 1999 and 34.2 percent in2005); Spain (32.2 percent in 1995, 27.4 percent in 1999, and 29.7 percent in 2007).

178 CRS/RCS, 51.2 2014

Figure 1

Confidence in Unions over Time

2530354045

nfide

nt

05

101520

1982 1990 1995 1999/2000 2006

% C

on

Canada USA

Source: World Values Survey.

Table 1 shows the percentage of individuals who expressed confidencein unions according to the independent variables included in the model,country, and survey wave.11 The empirical evidence suggests a few notabledifferences between Canada and the United States. Specifically, Canadi-ans who positioned themselves closer to the left of the ideological scale hada higher level of trust in unions in both waves compared to individualswho have the same political orientation in the United States. Canadianmanual workers also had a greater concentration of people who expressedconfidence in unions in the fifth wave compared to American manual work-ers.

Overall, however, the descriptive analyses brought out few substantialdifferences between the two countries. However, before concluding that“confidence in unions” is another element of similarity between them, ouranalysis should be deepened by examining, for both Canada and the UnitedStates, the effect of the interaction between the set of social and politicalvariables considered.

11. More specifically, respondents were sorted by wave, country, each independent variable, and finally bythe fact of being confident in unions. The independent continuous variables were recoded as dichotomousin order to simplify the reading of the table. We then focused our attention on the most importantcategory for union support as suggested by the literature about unionism (e.g., the variable “self-positioning on the ideological scale” is a 10-point scale and we focus our attention on people whoposition themselves between 1 and 5). Finally, inside this category, we show the percentage of peoplewho declared to be confident in unions.


Table 1

Percentage of Respondents Confident in Unions by Social andPolitical Indicators

Canada (percent) United States (percent)

Trust orientationTrust in others (most people can be trusted)

Fourth wave 36.2 39.5Fifth wave 34.9 27.9

Confidence in government (quite a lot and a great deal)Fourth wave 44.7 53.3Fifth wave 42.6 35.0

Working class profileOccupation (manual worker)


Income category (low level)Fourth wave 65.5 63.8Fifth wave 67.3 73.4

Political orientationInterest in politics (very and somewhat interested)


Self-position on ideological scale (left)Fourth wave 64.9 36.6Fifth wave 70.2 22.4

Private versus state ownership (state ownership should be increased)Fourth wave 43.5 51Fifth wave 42.7 29.3

Income equality (income should be more equal)Fourth wave 64.1 62.5Fifth wave 68.7 74.2

Sociodemographic variablesGender (male)


Age (18–50 years old)Fourth wave 34.2 35.3Fifth wave 31.6 29.9

Language at home (English)Fourth wave 37.6 39.3Fifth wave 39.1 29.1

Education (up to incomplete secondary)Fourth wave 62.3 64.9Fifth wave 67.1 70.7

SatisfactionLife satisfaction (very and somewhat satisfied)


Household financial satisfaction (very and somewhat satisfied)Fourth wave 38.4 37.9Fifth wave 33.8 30.3

Source: World Value Survey (fourth and fifth waves).

180 CRS/RCS, 51.2 2014

Multivariate Analysis

The application of the logistic regression model revealed that the factorsexplaining the fairly similar levels of social confidence in unions, whilealike in some regards, also presented important differences for the twocountries (Table 2).

The model led to interesting findings with respect to the first hy-pothesis (trust in “others” and trust in government are positively relatedto confidence in unions in both countries). In fact, regardless of whetherAmerican and Canadian citizens chose “Most people can be trusted” or“Can’t be too careful,” their choice did not have a significant impact interms of explaining their confidence in unions. Therefore, trust in “others”did not appear to be a driver of confidence in institutions in either coun-try.12 The outcome was different when a less abstract level of trust wasconsidered. In fact, the impact of confidence in government on the depen-dent variable was highly significant and went in the same direction in bothmodels. More specifically, Canadians and Americans who had confidence ingovernment were, respectively, 1.7 and 2.5 times more likely to have confi-dence in unions than those who did not.13 In line with previous studies, itappears that a positive perception of government enhanced the perceptionof unions. Again, this suggests that unions—as intermediates between thestate (first represented by the government) and society—benefit from apositive perception of government. In light of these results, the first hy-pothesis was only partially confirmed.

The empirical evidence concerning the impact of the working classprofile (i.e., our second hypothesis) on confidence in unions merits atten-tion. In Canada, blue-collar workers were 1.8 times more likely to expressconfidence in unions than individuals who had never held a job. More-over, the higher the respondent’s income category, the less likely he or shewas to trust unions. This means that individuals with lower incomes hadmore confidence in unions. These two results suggest that unions enjoy thetrust of individuals with a traditional working class profile. In other words,working class citizens are more likely to perceive that their interests areencapsulated by those of unions.

There are several possible explanations for this. The most importantmay be that unions have taken advantage of the higher union densityin Canada, which may have helped structure working class confidence onpersonal relations with union supporters. In turn, these workers may havehelped spread trust among individuals with a low income. Moreover, the

12. This finding could hide a problem of endogeneity. In order to exclude this possibility, we conductedtwo tests: first, we applied our model only to nonunion members; second, for the whole sample, wecontrolled for the role played by union membership in this relationship. The results in each case madeit possible to exclude the problem of endogeneity.

13. The impact of the independent variables on confidence in unions was estimated in terms of odds ratio.


Table 2

Logistic Regression of Confidence in Unions

Canada United StatesOdds ratio SE Odds ratio SE

Trust orientationTrust in others (0 = cannot be too

careful; 1 = most people can betrusted)

.969 (.103) 1.191 (.150)

Confidence in government (0 = notvery much; 1 = a great deal)

1.69**** (.174) 2.459**** (.296)

Working class profileOccupation (ref = No job)Professional/employer 1.07 (.327) .948 (.437)Nonmanual (white collar) 1.02 (.301) .768 (.355)Manual (blue collar) 1.85** (.525) 1.165 (.535)

Income category (0 = lowest level; 10= highest level)

.917**** (.018) .986 (.024)

Political orientationInterest in politics (0 = not very +

not at all interested; 1 = very +somewhat interested)

1.33*** (.141) 1.157 (.146)

Self-position on ideological scale (1= left; 10 = right)

.920*** (.026) .852**** (.027)

Private versus state ownership (1= private should be increased;10 = government should beincreased)

1.08*** (.025) 1.117**** (.031)

Income equality (0 = incomeshould be more equal; 10 =larger income differences asincentives)

.938*** (.021) .966 (.027)

Sociodemographic variablesGender (0 = male; 1 = female) 1.12 (.115) 1.214 (.146)Age .99*** (.004) .993* (.004)Language at home—Canada (0 =

English; 1 = French)1.22* (.136)

Language at home—United States(0 = English; 1 = Other)

.441* (.185)

Education (1 = noncompletion ofelementary education; 8 = upperlevel tertiary education)

1.05 (.034) .893*** (.031)

Time (0 = fourth wave; 1 = fifthwave)

1.02 (.114) .505**** (.066)

SatisfactionLife satisfaction (0 = dissatisfied;

1 = satisfied)1.16 (.206) .859 (.160)

Household financial satisfaction (0= dissatisfied; 1 = satisfied)

1.42*** (.189) 1.126 (.161)

Pseudo R2 .059 .089LR chi2 (17) 148.29 168.42N 1,928 1,449

Source: World Value Survey (fourth and fifth waves).

Note: *p < .10; **p < .05; ***p < 0.01; ****p < .001.

182 CRS/RCS, 51.2 2014

significance of these two variables in our model lends support to the ideathat there remain some traces of a persistent social democratic orientationamong the Canadian working class (Friedman 1998; Jacoby 1991), whichmay lead to confidence based on perceived shared values. This findingis directly in line with one of the main conclusions drawn by Lipset andMeltz (2004), who observed a higher degree of “class awareness” in Canadathan in the United States. In any case, despite the widespread processesof deindustrialization and production reconfiguration that have emergedin response to global challenges, unions appear to have maintained animportant relationship of trust with the working class in Canada.

In the United States, on the other hand, occupation and income arenot significant factors in terms of explaining confidence in unions. Unlikein Canada, unions do not appear to enjoy the support of the working class;this may be a result of the general difficulty unions have experiencedin organizing workers over the last three decades, which may have ledto a downward spiral of confidence in unions. Moreover, the historicalprevalence of a conservative political orientation and fragmentation withinthe American working class, as argued by Friedman (1998) and Jacoby(1991), may have helped make income and occupation insignificant factorsof confidence in unions. This finding supports the widespread observationof generally distant relations between unions and blue-collar workers inthe United States. In sum, these empirical results confirm our secondhypothesis.

The third hypothesis (i.e., higher impact of interest in politics andof progressive political orientation in Canada than in the United States)was not completely confirmed. On the one hand, as expected, the fact thatCanadian unions had traditional ties to political parties and, more gener-ally, to the political arena meant that interest in politics was a significantfactor of confidence in unions. More specifically, people who were inter-ested in politics were 1.3 times more likely to have confidence in unionsthan those who were not. This positive relationship appears to suggestan overlap between unions and politics in Canada. The historical doublelink between unions and parties, notably the New Democratic Party andthe Parti Quebecois, and the influence wielded by union leaders in thepolitical arena when it comes to promoting social justice issues, attest tothe greater politicization of the union movement in Canada. On the otherhand, also as expected, the relationship between interest in politics andconfidence in unions was not statistically significant in the United States,where a business unionism that has never developed a relationship of ac-count with parties and political dynamics has historically prevailed. Unionleaders have generally been far removed from the political arena, in partbecause union involvement is anathema to political parties. The nonsignif-icance of this variable in the model also supports the hypothesis that thereis limited overlap between unions and the political sphere in the UnitedStates.


However, Canadians and Americans shared some traits with regard tothe influence of political orientation on confidence in unions. In fact, two ofthe three variables introduced in the model to discern between citizens witha progressive versus conservative political orientation were significant oneither side of the border and went in the same direction. More specifically,in both countries, the greater the extent to which citizens identified withideological positions that leaned toward the extreme left, the more likelythey were to have a “great deal” or “quite a lot” of confidence in unions. Sim-ilarly, the further the respondents’ choices were from “private ownershipof business should be increased” (value 1 in the scale) and the closer theywere to “government ownership of business should be increased” (value 10in the scale), the more likely they were to have confidence in unions.

However, there were differences in the two countries with respect tothe respondents’ answers regarding income inequality. In fact, the greaterthe extent to which Canadians approved of the statement that larger in-come differences act as an incentive, the less likely they were to haveconfidence in unions. This relationship was not significant in the UnitedStates, which can be interpreted as an indication that the values of “pos-sessive individualism” prevail widely there.

All in all, taking these three variables into account, it appears that, ingeneral, unions evoke more positive reactions among the most progressivecitizens in both countries, but more consistently so in Canada. This may beexplained on the grounds that citizens with a more progressive political ori-entation feel an allegiance to the norms and values associated with uniongoals, and therefore conclude that unions can be trusted. The general ideaof unions as an equitable social force may have prompted their responsesthat may have been based more on emotion or identity than on a cognitiveassessment of specific institutional performance. Based on these results,the third hypothesis was not completely confirmed. In fact, although thedata corroborate the notion of distant relations between unions and politicsin the United States, the political orientation of citizens had a significantimpact in both countries.

The impact of sociodemographic variables also shed light on differ-ences between Canada and the United States. While the respondents’ gen-der was not significant in either country, age was significant in both. Al-though this finding was stronger for Canada, the older the respondent, theless likely he or she was to have confidence in unions in both countries. Theimplications of this finding are twofold. On the one hand, it is a positivefinding for unions that young people express greater confidence in unions,as this trust could substantially contribute to their future actions and toenhancing union renewal. It suggests, moreover, that even though unionshave been losing members over time, they continue to appeal to youngerpeople. On the other hand, unions enjoy less trust from the people theygenerally target, namely, the larger share of adult workers. An aware-ness among older citizens of the growing limits that unions have faced in

184 CRS/RCS, 51.2 2014

recent decades and the perception that union performance has waned overtime may partially explain this tendency. In other words, it may be thatolder people are aware of the crisis of unionism in both countries, and aretherefore less inclined to trust unions. Further longitudinal analyses areneeded to explore these considerations and confirm whether this finding isdue to a pure age or cohort effect, or a mix of the two.

The language spoken at home was also significant, although not highlyso, in both the United States and Canada. In the United States, people whodid not speak English at home, such as immigrants, were less inclined totrust unions. This result is in line with many previous studies in theUnited States showing that the integration of immigrants into the labormovement is far from automatic and involves a complex social processthat evolves over time and depends notably on how immigrants enteror leave unionized occupations and workplaces (Reitz and Verma 2000).With regard to Canada, only citizens who indicated that English (1,387) orFrench (541) was the language spoken at home were included, separately,in the model, owing to the high social significance of language for Canadiansociety.14

Among the 541 respondents who spoke French at home, 487 wereinterviewed in Quebec. The model shows that Francophone respondentshad greater confidence in unions than Anglophones. This finding under-scores the fact that where unions have historically been more organizedand influential in terms of labor dynamics, such as in Quebec, they enjoyhigher confidence. Moreover, the historical prevalence of more collectivity-oriented values in Quebec compared to other regions in Canada, as well asthe high union proximity to the Parti Quebecois and the crucial experienceof Quebec’s Quiet Revolution, all contribute to explaining the higher levelof trust in unions among people who speak French at home (Grant 2003).

The highest level of education attained by the respondents was notsignificant in Canada. The reverse was true in the United States. Morespecifically, the higher the respondents’ level of education, the less likelythey were to have confidence in unions (odds ratio scores less than 1).This was the only variable in the model that might suggest a degree ofconnection between unions and the working class in the United States.

In both the Canadian and the American models, we introduced a timevariable representing the two survey years. This variable was not signif-icant in Canada, but was significant in the United States. Thus, between1999 and 2006, Americans became less likely to express confidence inunions. This confirms that confidence in unions is on a downward slopein the United States. Further, analyses are needed to test whether thisdecline represents the first step in a negative spiral or was merely a con-tingent finding.

14. People who spoke neither English nor French at home were excluded. However, the model was alsotested including these excluded citizens, and the relationships did not change.


Finally, the effect of the respondents’ financial situation and life sat-isfaction on their confidence in unions brought out important differencesbetween the two countries. Although there was no significant relationshipbetween personal life satisfaction and confidence in unions in either theUnited States or Canada, household financial satisfaction, on the otherhand, was highly significant in Canada but not in the United States. Infact, Canadians who indicated that they were satisfied with their family’sfinancial situation were 1.4 times more likely to have confidence in unionsthan those who indicated that they were dissatisfied. Thus, in Canada,there was a positive link between household financial satisfaction andconfidence in unions. Given the union’s role as an active agent for salarybargaining, the higher union presence in both the public and private sec-tors in Canada, compared to the United States, may partially explain thisfinding.

CONCLUSION

Analyses of the changing situation of unionism in industrialized countrieshave tended to draw a rather bleak picture of its future prospects. Theoften fatalistic discourse that has emerged, mainly from the analyses ofunion density, has sometimes concluded that unionism is in the midst ofa crisis of social legitimacy (Carroll and Ratner 1995; Kane and Marsden1988). In this regard, this article focused on confidence in unions as abroader measure of public support for unions and compared two nationalcases, the United States and Canada.

The findings highlight two fundamental observations. Primarily, inthe most recent years of the survey, in both countries, it was not onlyunion members who expressed confidence in unions. In fact, more than oneof three American and Canadian citizens expressed confidence in unions.This result, in itself, constitutes both good news and bad news for the unionmovement. It is good news insofar as it shows that union organizationscontinue to enjoy an important fund of public confidence. It also allows for adegree of optimism regarding the growth potential of the union movement,particularly in the United States, since our analysis demonstrates thatthere is a vast pool of individuals who appear to adhere to the valuesof unionism without benefiting from union representation. On the otherhand, it clearly points to the important institutional and legal obstaclesstanding in the way of the unionization of workers in the United Statesand making it impossible to satisfy the desire of some of these workers tobe collectively represented.

Second, the analysis conducted in the context of this article clarifiedthe influence of certain sociodemographic characteristics and the politi-cal and trust orientation of the respondents on their propensity to haveconfidence in unions. While Canada and the United States generally en-joy a relatively similar level of social confidence in unions, there were

186 CRS/RCS, 51.2 2014

differences as well as similarities between the two countries with respectto the effect of these respondent characteristics.

As for the differences, our results suggest, in particular, that as inter-est in politics and personal economic satisfaction rise among Canadianscitizens, so does their confidence in unions. These two relationships werenot significant in the United States. Our results also show that, whileindividuals with a working class profile in Canada tend to have greaterconfidence in unions, this relationship could not be established for theUnited States.

As for the similarities, the most interesting finding was that suggest-ing that young people were more inclined to have confidence in unionsthan older people, in both Canada and the United States. In line withmany international studies suggesting that young people continue to havepositive attitudes toward unionism (Gomez, Gunderson, and Meltz 2002),this finding is promising for the labor movement since it raises the possi-bility of greater unionization among these citizens, who may one day jointhe labor movement when given the opportunity. Nevertheless, given thehistorically lower level of unionization among young people in both Canadaand the United States, this finding also supports the conclusions of otherstudies demonstrating that the union density differential between youngand adult workers is due to supply-side constraints rather than a lowerdesire for unionization on the part of the young.

References

Adams, R. 1995. Industrial Relations under Liberal Democracy. Columbia, SC: University ofSouth California Press.

Allen, B. 2006. “Inside the CAW Jacket.” New Socialist 57(July-August):18–20.

Alston, J.P., T.M. Morris and A. Vedlitz. 1996. “Comparing Canadian and American Values: NewEvidence from National Surveys.” American Review of Canadian Studies 26(3):353–64.

Bruce, P.G. 1989. “Political Parties and Labor Legislation in Canada and the U.S.” IndustrialRelations 28(2):115–41.

Camfield, D. 2011. Canadian Labour in Crisis. Reinventing the Workers’ Movement. Halifax andWinnipeg: Fernwood Publishing.

Carroll, W.K. and R.S. Ratner. 1995. “Old Unions and New Social Movements.” Labour 35:195–221.

Fligstein, N. 2001. The Architecture of Markets: An Economic Sociology of Twenty-First CenturyCapitalist Societies. Princeton, NJ: Princeton University Press.

Frangi, L. and V. Memoli. 2013. “Unfolding the Growing Confidence in Latin American Unions:A Longitudinal Analysis.” Argomenti 38:25–46.

Frangi, L. and V. Memoli. Forthcoming. “Social Confidence in Brazilian Unions: A LongitudinalAnalysis.” Latin American Perspectives.

Freeman, R.B.and J.L. Medoff. 1984. What Do Unions Do? New York: Basic Books.


Friedman, G. 1998. State-Making and Labor Movements. The United States and France, 1876–1914. Ithaca, NY: ILR Press.

Galeotti, F. and D.J. Zizzo. 2012. “Trust and Trustworthiness with Singleton Groups.” WorkingPaper series, Centre for Behavioural and Experimental Social Science (CBESS) 12-03,School of Economics, University of East Anglia, Norwich, UK.

Godard, J. 2009. “The Exceptional Decline of the American Labor Movement.” Industrial &Labor Relations Review 63(1):82–108.

Goldfield, M. and B.D. Palmer. 2007. “Canadian Labour Movements: Uneven Developments.”Labour 59(1):149–77.

Gomez, R., M. Gunderson and N. Meltz. 2002. “Comparing Youth and Adult Desire for Unioni-sation in Canada.” British Journal of Industrial Relations 40(3):521–42.

Gomez, R., S.M. Lipset and N. Meltz. 2001. “Frustrated Demand for Unionisation: The Caseof the United States and Canada Revisited.” Discussion Paper 492, Centre for EconomicPerformance, London School of Economics and Political Science, London, UK.

Grant, M. 2003. “Quebec: Toward a New Social Contract–From Confrontation to Mutual Gains?”Pp. 51–96 in Beyond the National Divide, Regional Dimensions of Industrial Relations,edited by M.J.R. Thompson and A.E. Smith. Montreal and Kingston: McGill-Queen’s Uni-versity Press.

Hall, P.A. and D. Soskice, eds. 2001. Varieties of Capitalism: The Institutional Foundation ofComparative Advantage. Oxford: Oxford University Press.

Hardin, R. 1999. “Do We Want Trust in Government?” Pp. 22–41 in Democracy & Trust, editedby M.E. Warren. Cambridge: Cambridge University Press.

Iglic, H. 2010. “Voluntary Associations and Tolerance: An Ambiguous Relationship.” AmericanBehavioral Scientist 53(5):717–36.

Jacoby, S. 1991. “American Exceptionalism Revisited: The Importance of Managers.” Pp. 173–200 in Masters to Managers: Historical and Comparative Perspectives on American Em-ployees, edited by S. Jacoby. New York: Columbia University Press.

Kane E. and D. Marsden. 1988. “The Future of Trade Unionism in Industrialized MarketEconomies.” Labour and Society 13:109–24.

Kasperson, R.E., D. Golding and S. Tuler. 2005. “Social Distrust as a Factor in Siting HazardousFacilities.” Pp. 29–50 in The Social Contours of Risk, vol. 1, edited by J.X. Kasperson andR.E. Kasperson. Trowbridge, UK: Cromwell Press Ltd.

Kumar, P. 1993. From Uniformity to Divergence: Industrial Relations in Canada and the UnitedStates. Kingston: Queens University, IRC Press.

Laplante, N. and D. Harrison. 2008. “Conditions for the Development of Trust between Managersand Union Representatives in a Context of Innovation.” Industrial Relations 63(1):85–107.

Lewis, J.D. and A.J. Weigert. 2012. “The Social Dynamics of Trust: Theoretical and EmpiricalResearch, 1985–2012.” Social Forces 91(1):25–31.

Lipset, S.M. 1990. Continental Divide: The Values and Institutions of the United States andCanada. New York: Routledge.

Lipset S.M. and G. Marks. 2000. It Didn’t Happen Here: Why Socialism Failed in the UnitedStates. New York: W.W. Norton.

Lipset, S.M. and N.H. Meltz. 2004. The Paradox of American Unionism. Why Americans LikeUnions More Than Canadians Do, But Join Much Less. Ithaca, NY: Cornell UniversityPress.

188 CRS/RCS, 51.2 2014

Macpherson, C.B. 1964. The Political Theory of Possessive Individualism: Hobbes to Locke.London: Oxford University Press.

Memoli, V. 2013. “Responsiveness.” Pp. 233–49 in La qualita della democrazia in Italia, editedby L. Morlino, D. Piana and F. Raniolo. Bologna: Il Mulino.

Newton, K. and P. Norris. 1999. “Confidence in Public Institutions: Faith, Culture and Perfor-mance.” Paper presented at the annual meeting of the American Political Science Associ-ation, September 1–5, Atlanta, GA.

Palmer, H. 1992. Working Class Experience: Rethinking the History of Canadian Labour, 1800–1991. 2d ed. Toronto: McClelland & Stewart.

Panitch, L. and D. Swartz. 2003. From Consent to Coercion: The Assault on Trade Union Free-doms. 3d ed. Aurora: Garamond.

Powell G.B. 2005. “The Chain of Responsiveness.” Pp. 62–76 in Assessing the Quality of Democ-racy, edited by L. Diamond and L. Morlino. Baltimore, MD: Johns Hopkins UniversityPress.

Putnam, R. 1995. “Bowling Alone: America’s Declining Social Capital.” Journal of Democracy6(1):65–78.

Reitz, J.G. and A. Verma. 2000. “Immigration, Ethnicity and Unionization: Recent Evidence forCanada.” Industrial Relations 43(4):835–54.

Rose, J.B. and G.N. Chaison. 2001. “Unionism in Canada and the United States in the 21stCentury: The Prospects for Revival.” Industrial Relations 56(1):34–65.

Rouillard, J. 2009. L’experience syndicale au Quebec. Ses rapports avec l’Etat, la nation etl’opinion publique. Montreal, QC: VLB Editeur.

Savage, L. 2010. “Contemporary Party-Union Relations in Canada.” Labour Studies Journal35(1):8–26.

Sinyai, C. 2006. Schools of Democracy: A Political History of the American Labor Movement.Ithaca NY: ILR Press.

Taras, D.G., A. Ponak and M. Gunderson. 2001. “Introduction to Canadian Industrial Relations.”Pp. 1–24 in Union-Management Relations in Canada, 4th ed., edited by D.G. Taras, A.Ponak and M. Gunderson. Toronto, ON: Addison Wesley Longman.

Taylor, C. 1982. Radical Tories. Halifax, NS: Formac Publishing Company.

Tonkiss, F. and A. Passey. 1999. “Trust, Confidence and Voluntary Organizations: Better Valuesand Institutions.” Sociology 33(2):257–74.

Warren, M.E., ed. 1999. Democracy & Trust. Cambridge: Cambridge University Press.

Yamigichi, T. and M. Yamigichi. 1994. “Trust and Commitment in the United States and Japan.”Motivation and Emotion 18(2):129–66.

Documents

Social Confidence in Unions: A U.S.-Canada Comparison