Worker Voice in America: Is There a Gap between What Workers Expect and What They Experience?

Abstract

This article is the fifth in a series to celebrate the 70th anniversary of the ILR Review. The series features articles that analyze the state of research and future directions for important themes this journal has featured over many years of publication.

The decline in unionization experienced in the United States over the past 40 years raises a question of fundamental importance to workers, society, and the field of industrial relations: Have workers lost interest in having a voice at work, or is there a gap between workers’ expectations for a voice and what they actually experience? And if a “voice gap” exists, what options are available to workers to close that gap? The authors draw on a nationally representative survey of workers that both updates previous surveys conducted in 1977 and 1995 and goes beyond the scope of these previous efforts to consider a wider array of workplace issues and voice options. Results indicate that workers believe they should have a voice on a broad set of workplace issues, but substantial gaps exist between their expected and their actual level of voice at work. Nearly 50% of non-union workers say they would vote for a union, compared to approximately one-third in the two prior national surveys, which points to continued interest in unions as a voice mechanism. Additionally, the authors find significant variation in the rates of use of different voice options and workers’ satisfaction with those options. The results suggest that a sizable voice gap exists in American workplaces today, but at the same time, no one voice option fits all workers or all issues.

Keywords

worker voice voice gap unionization rate workplace issues Worker Voice Survey Current Population Survey

Whether workers have sufficient voice at work to affect their working conditions and pay is one of the fundamental questions addressed in the field of industrial relations. The lack of voice has led to waves of unionization drives historically, with union density in the United States peaking at 35% in 1945 and now at 11%. During this period of union decline, inequality has returned to the historic highs of the 1920s, the nature of work and the workforce have changed, and new forms of voice at work have emerged—raising the question of whether workers now have sufficient options to shape their work environment and pay.

In this article, we consider two primary questions. First, do US workers currently face a “voice gap”? (By this, we mean the difference between how much say they expect to have at work versus how much they actually have.) Second, has the voice gap changed over time? We address these questions using a new nationally representative survey of the US workforce to assess the extent to which contemporary workers expect to have a voice on a broad array of workplace issues, the size of the voice gap experienced on these issues, and the differences across demographic groups and between those in standard and non-standard employment relationships. We also compare the current situation with previous surveys to help us assess changes.

We also investigate workers’ experiences with the growing array of new options that provide workers with a voice on workplace issues. In doing so we pay particular attention to changes in the level of interest in a more traditional option: union representation. We compare interest in unions now with expressed interest in the 1990s (as assessed by Freeman and Rogers 1999) and the 1970s (as captured in an earlier Department of Labor survey and studied by Kochan 1979 and Quinn and Staines 1979). Finally, we suggest directions for future research and policies needed to identify viable paths for filling the gap between the level of voice at work that workers say they expect versus what they actually experience.

The Concept of Worker Voice: Historical and Contemporary Considerations

The term “worker voice” has been used in various ways historically and currently. Hirschman (1970) provided a generic definition of voice as an effort directed at a higher authority to achieve a change in practice. Employment relations scholars use this generic definition but adapt it depending on their frame of reference for understanding the interests at stake in employment relationships (Fox 1966; Budd and Bhave 2008). As Barry and Wilkinson (2016) noted, those using a unitary frame of reference assume that workers’ and employers’ interests are congruent, and therefore the goal of worker voice is to elicit “positive” actions or “organizational citizenship” behaviors. Doing so will improve individual, group, or organizational outcomes that will potentially also enhance commitment, engagement, trust, and job satisfaction (Organ 1988; Marchington, Boxall, Purcell, and Wright 2007; Morrison 2011; Klaas, Olson-Buchanan, and Ward 2012; Bashshur and Oc 2015). Workers’ voice serves joint goals.

By contrast, those following a pluralist tradition of employment relations view employers and workers as having conflicting as well as shared interests. The concept of worker voice under this tradition is rooted in a democratic ethos articulated by Webb and Webb (1897). Pluralists argue that workers should have the right and ability to assert their interests individually or collectively to influence the conditions under which they work. Our use of the term “worker voice” is embedded in this pluralist tradition, which recognizes that workers want to identify with and contribute to organizations that share their values and interests. Voice may therefore mix individual and collective efforts to improve organizational processes and performance with efforts to assert worker interests that are in conflict with employers’ or other parties’ interests at work. Collective voice is most commonly exercised through collective bargaining, in which trade unions negotiate the terms of employment with employers (Commons 1913; Fox 1975; Kochan 1980; Budd 2004; Ackers 2007). Other channels for worker voice, however, have developed in US workplaces, as well as in other contexts, and we consider these as well.

Whither the Voice Gap

Previous surveys reported evidence of two types of voice gaps: between the percentage of workers who expected to have influence over some aspects of work but did not have it and between the percentage of non-union workers who wanted union representation but did not have it. Both the 1977 survey and the 1995 survey examined how many non-union workers would prefer to be represented by a union if given a choice, which Freeman and Rogers (1999) labeled the “representation gap.” Results were quite stable across the time periods of the two surveys: Approximately one-third of the non-union, non-managerial workforce indicated they would vote to unionize if given the opportunity (Kochan 1979; Freeman and Rogers 1999). The 1995 survey conducted by Freeman and Rogers also provided data on the gap between the amount of say or influence workers felt they should have on a set of workplace issues, and their actual level of say or influence. Their results showed a sizable gap.¹ We use the term “voice gap” to include both the representation gap and this workplace influence gap.

Given the central role that voice and representation play in the field of employment relations, it is somewhat surprising how little theory can be drawn on for anticipating when voice gaps will be evident or how they may shift over time. We examine possible explanations for why we might see a gap today, or why previously documented gaps may not be present now, before turning to an empirical assessment of the current realities.

Several arguments have emerged to explain why a voice gap may not be evident today. US workers have faced substantial changes in employment relationships and working conditions over the past 40 years (Barley and Kunda 2006; Wartzman 2017). One body of research suggests that perhaps workers have adjusted their expectations to fit with current economic and workplace realities so that workers no longer believe they ought to have a voice on workplace issues. Or, they may have become conditioned to take increased insecurity and instability as the new norms for work (Newman 1999: 69; see also Meyer 1995; Heckscher 1996; Kalleberg 2013; Sharone 2013). Other research suggests that most workers accept these changes as being driven by “the market,” as a disembodied force, rather than being chosen by their employers. If current working conditions reflect broader forces that employers cannot change, then workers may have internalized acceptance of limited voice as a form of “learned helplessness” (Peterson, Maier, and Seligman 1993). This belief would make efforts to exert voice less relevant and could lead to a decrease in what is viewed as an appropriate level of say.

Workers may also have internalized meritocratic norms that assume “winners” will get what they need from employers (Smith 2002), or the idea that they are “free agents,” even as employees. From this perspective, workers may not expect to have a say in their current workplaces but instead expect that they would need to move on to find more appealing work conditions if they are dissatisfied or their contributions are not recognized. Pugh (2015), for example, found that workers expect little or nothing from their employers but hold high expectations of themselves. This “one-way honor code” means workers demand hard work, dedication, and cheerful compliance of themselves but do not expect job security or voice in return, hoping only for a paycheck. These broad cultural or structural changes may mean that workers do not generally believe it is appropriate for them to have a say in determining working conditions. Furthermore, younger workers, who have been exposed only to the current ideologies, may be less likely to expect or demand more say. The idea here is that the voice gap may have declined because expectations for what is appropriate have shifted.

A second argument for why there may no longer be a sizable voice gap is that employers and workers have reached a satisfactory set of arrangements; that is, with the development of new human resource policies and systems, a gap no longer exists between what workers believe is appropriate regarding their say at work and what they experience on their jobs (Foulkes 1980; Guest 1987; Fiorito 2001; Machin and Wood 2005). A variety of internal, firm-provided processes, such as ombudsman systems (Rowe 1987), non-union grievance procedures (Lewin 1987), and affinity or identity groups linking individuals of the same race, gender, religion, or sexual orientation (Portes and Sensenbrenner 1993; Creed and Scully 2000), may provide effective channels that satisfy workers’ interest in voice and due process. In some workplaces, formalized participation processes and worker–employer committees (Freeman and Rogers 1999) may help workers feel they can weigh in effectively. Supervisors may be chosen or trained to welcome and address worker concerns more than in the past (Detert and Burris 2007; Detert and Treviño 2008).

Finally, it may be that alternative, independent forms of worker voice are now filling the void created by union decline. A variety of alternative worker advocacy initiatives have emerged in recent years that are attempting to provide voice options that are independent of employers yet do not seek to establish collective bargaining relationships. For example, new channels for worker advocacy include online petitions, such as those offered by Coworker.org (Heckscher and McCarthy 2014; Arvins, Larcom, and Weissbourd 2018); online ratings of employer practices, such as Glassdoor or Turkopticon (Benson, Sojourner, and Umyarov 2015); movements, such as the Fight for $15 (Rolf 2016); industry- or occupation-based workers’ centers, such as the Restaurant Opportunities Center United or the National Domestic Workers Alliance; or demographic-based associations, such as immigrant worker centers (Fine 2006; Milkman 2011; Cordero-Guzmán 2015). To date, however, no systematic evidence documents how aware workers are of these options or how much use they are making of them.

Alternatively, there may be reasons to believe the voice gap still exists and may, in fact, have increased. The decline in union membership from 28% in 1977, to 15% in 1995, to 11% in 2017 (using Current Population Survey figures) is one reason to expect a voice gap increase. Workers may still want voice at work but the decline in union representation means they do not have the power or resources to demand a say. Two broad interpretations surround the relationship between declining union representation and interest in worker voice. First, it may be that unionization is less likely today in the face of increasing employer power and opposition to unions (Bronfenbrenner 1998), the high hurdles involved in navigating the union election processes (Ferguson 2008), the constrained forms of representation allowed under labor law (Estlund 2010; Kochan 2011), and the low likelihood that a union organizing drive will reach a given workplace (National Labor Relations Board 2018). Yet, interest in workplace voice may still be high, and so we might see a growing representation gap that indicates an unmet demand for union representation (Farber and Krueger 1992). Indeed, a number of smaller polls and surveys carried out over the years have reported some growth in interest in, and support for, unions (Lipset and Meltz 2004; Freeman 2007). Second, it may be that declining unionization, as well as the changes reviewed above, means many workers no longer think of union representation as a primary voice mechanism and no longer feel a need for unions. In that case, we would expect to see both decreased interest in joining a union over time and perhaps a closed or very small voice gap today.

We use our 2017 national survey to assess the current state of voice in the United States and consider what explains that gap. We then compare that to findings from the two prior surveys in 1977 and 1995 to assess change over time.

Data and Measures

Data

Our 2017 survey of voice at work included a nationally representative sample of 3,915 American workers from the National Opinion Research Corporation’s (NORC) AmeriSpeak Panel (NORC 2017). We repeated some questions asked in the 1977 and 1995 surveys but expanded the scope to cover a broader array of workplace issues, such as work scheduling; concerns about respect, discrimination, and harassment; and newer options for voice, such as ombudsman systems, online petitions, and protests (see section below on Measures for a complete list of questions). The questions in our Worker Voice Survey were generated by our study team and were refined and pre-tested by NORC.

NORC’s AmeriSpeak panel consists of a nationally representative sample of households whose members agree to be contacted by NORC for the various surveys it conducts. For this survey, respondents were screened to include those who were 18 years or older, were currently working for pay, and were not upper-level managers, owners of businesses that employed others, or family members of owners.² Only one worker per household (randomly selected if more than one eligible AmeriSpeak participant resided in the household) was selected for participation. Those surveyed workers were invited to participate between April 19 and May 29, 2017. The survey was available in English and Spanish, and participants completed the survey on the website or by phone. Participants earned AmeriSpeak credit valued between $3 and $5 for completing the survey. In total, 94% of those who passed the screening requirements for inclusion in the study completed the survey. NORC calculated sampling weights to ensure the final sample accurately reflected the characteristics³ of the workforce as reported in the March 2016 Supplement of the Current Population Survey (CPS). Details on the sampling and weighting procedures are available on request.

Table 1 and Appendix Table A.1 report the individual demographic characteristics of our sample as weighted by NORC on race, gender, age, education, and income with comparisons to the Current Population Survey Annual Social and Economic Supplement (CPS ASEC) in 2016. Descriptive statistics for the remainder of this article use these weighted data whereas the multivariate models do not.⁴ Of our sample, 8% self-identified themselves as temporary employees, contract employees, or independent contractors,⁵ a number that approximates the 2017 Current Population Survey Contingent Worker Supplement estimate⁶ (Bureau of Labor Statistics 2018).

Table 1.

Demographic and Financial Characteristics of Workers in CPS Annual Social and Economic Supplement (ASEC) versus Worker Voice Survey (WVS) Sample

Variable	CPS ASEC (2016)(%)	WVS (2017)(%)
Female	47	47
Married (1 = yes)	56	48
Self-employed	6	7
Union representation^a	11	18
Age (years)
18–34	33	37
35–49	32	31
50–64	29	28
65+	6	5
Race
White	65	64
Black	11	11
Asian, Other	7	5
Hispanic	17	17
Two or more races	1	3
Education
No high school diploma	6	7
High school graduate or equivalent	27	27
Some college	29	33
College degree	23	34
Advanced degree	14	14
Region
Northeast	18	18
Midwest	22	23
West	24	22
South	36	37
Hours worked per week (all jobs)
1–10	2	2
11–20	6	6
21–34	9	11
35–40	59	48
41–50	15	22
51+	9	11
Household income (dollars)
< $30,000	14	21
$30,000–$49,999	15	19
$50,000–$74,999	19	20
$75,000–$124,999	27	26
$125,000+	24	14
Primary job earnings ^b (dollars)
< $30,000	40	41
$30,000–$49,999	24	26
$50,000–$74,999	18	18
$75,000–$109,999	10	10
$110,000+	8	6
Industry
Natural resources, mining and utilities	3	2
Construction	7	5
Manufacturing	10	9
Wholesale trade	3	2
Retail trade	11	11
Transportation and warehouse	4	9
Finance, insurance and real estate	7	7
Information	2	2
Professional and business, scientific, and technical	8	6
Administrative and support and waste management and remediation services	5	3
Educational services	9	9
Health care and social assistance	14	17
Arts, entertainment, and recreation	2	2
Accommodation and food services	7	5
Other services (religious, repair/maintenance, personal service industries, etc.)	5	4
Public administration	5	7
N	53,611	3,644

Sources: Analysis of Worker Voice Survey data (based on NORC AmeriSpeak sample) and the Center for Economic and Policy Research’s extracts of the 2016 CPS Annual Social and Economic Supplement (ASEC) and 2016 CPS Outgoing Rotation Group (ORG) data.

Notes: For each data set, the sample is restricted to those workers age 18+ who are employed and working for pay.

Because the union question is asked of only a subset of CPS ASEC respondents, we instead used the 2016 CPS ORG sample (n = 163,718) to estimate the union membership rate. Note that self-employed members are included in this rate, as opposed to those in the union membership rates published by the Bureau of Labor Statistics.

For the CPS ASEC, this reflects any wage or salary income and is not necessarily limited to one’s primary job.

Although most demographic and workplace characteristics are similar between our sample and the CPS ASEC sample, note these two differences. Since, by design, our sample excludes senior managers, owners, or family members of owners, it contains a higher percentage of low-income workers than in the national population. For similar reasons, 17.6% of the respondents in our Worker Voice Survey report they are “represented by a union or professional association” in their job as compared to 11.5% of the CPS Ongoing Rotation Group (ORG) employed sample who report being represented by a union or employee association (Bureau of Labor Statistics 2017). Although part of this difference is attributable to our exclusion of higher-level managerial employees, the wording of our question “union or professional association” also differs from the CPS question about “union or employee association.” Our wording may include members of professional associations that lobby on behalf of members but do not engage in collective bargaining.

Measures

A key subset of the questions in our survey replicate questions asked in two prior national surveys that addressed some aspects of worker voice: the 1977 Quality of Employment Survey (Kochan 1979; Quinn and Staines 1979) and the 1995 Worker Representation and Participation Survey conducted under the direction of Richard Freeman and Joel Rogers (1999). Appendix Table A.2 lists the specific questions and items used to measure the following constructs in our survey.

Expected (Appropriate) and Actual Say

Freeman and Rogers (1999) measured how much voice (operationalized as how much say or influence) workers indicated they “ought to have” on a variety of workplace issues (which we label “expected say” or “appropriate say” below) and how much say or influence respondents actually had on their jobs. From these two questions, they derived a “voice gap” estimate.

We followed a similar procedure to measure workers’ views on how much say they “ought to have.” However, we expanded the set of issues (see discussion below) to better reflect contemporary employment relations. Respondents in our survey were asked how much say or influence they believe they ought to have over an array of issues affecting their work. This measure captures workers’ views of what is sensible in terms of their input at work; we see it as revealing workers’ views of the appropriate social contract, specifically the balance between management and workers’ say at work.

We chose issues at three levels of the employment relationship that prior research indicates are important to worker voice and welfare (Kochan, Katz, and McKersie 1986). These issues include 1) personal and workplace issues, such as safety, respect, dealing with abuse or discrimination, control over how to do one’s work, and scheduling of work hours; 2) personnel or collective bargaining issues, such as compensation, benefits, job security, promotion, and training; and 3) higher-level organizational issues or strategies involving technology, quality of products or services provided, and employer values.⁷ For each issue, respondents rated on a 5-point scale (5 = unlimited say; 4 = a lot of say; 3 = some say; 2 = little say; 1 = no say) the level of voice they expect at work. Using the same 5-point scale, survey respondents also rated how much say or influence they actually have over the same array of issues discussed above.

Voice Gap

Voice gap is measured as the difference between workers’ Expected Say and their Actual Say at work on each issue mentioned above. In order to make our analyses tractable, we construct three aggregated measures of voice gap: the mean of voice gap across workplace/personal issues, personnel/collective bargaining issues, and organizational strategy issues, respectively.⁸

Union Support

Both prior surveys asked whether workers would vote for or against the union if a union representation election was held at their workplace. We replicated that question in the survey to provide comparative data on this issue. Union Support is a dummy variable that equals 1 if the respondent would vote for the union; 0 if voted no.

Use and Satisfaction with Voice Options

Freeman and Rogers (1999) also asked about voice or representation options in addition to unions. They focused on employee participation committees in response to the public policy debates over this issue that were underway at that time (Commission on the Future of Worker-Management Relations 1994). We chose to further expand the array of voice options to better reflect those that are offered by some firms or that are promoted by some worker advocacy organizations today. We asked about respondents’ use of a broad array of options that might be available within one’s organization, such as asking for assistance from supervisors, coworkers, or ombudsmen; filing a grievance or complaint; or participating in a joint worker–management committee. Note that the internal channels include both formal procedures and informal interactions with supervisors or coworkers. We also asked about options that involve independent channels, such as unions, occupational associations, protests, strikes, and several newer forms of worker voice that have emerged in some settings, such as online forums and petitions. Use of Voice Options is a group of dummy variables that equals 1 if the respondent has used the specific voice option, and 0 otherwise.

Those who have used each type of voice option evaluated their Satisfaction with Voice Options on a 5-point scale (5 = extremely satisfied; 4 = very satisfied; 3 = somewhat satisfied; 2 = not very satisfied; 1 = not satisfied at all).

Results

Do Workers Still Want a Voice, or Have They Lowered Their Expectations?

We begin by presenting the data on the extent to which workers currently expect to have influence over a wide range of contemporary workplace issues. Two things stand out in the responses shown in Figure 1. First, a clear majority of contemporary workers expect to have a voice on how they work, their conditions of employment, the quality of the products or services they help produce or deliver, and the values their organization stands for. Across the full range of issues, fewer than 7% of respondents indicate they ought to have “no say,” and approximately 20% or fewer report that they should have “little” or “no say” on any of these matters.

In addition, workers’ views regarding how much of a voice they should have varies across issues. The distributions show that workers recognize the need to share influence with other parties, presumably their supervisors and managers, on compensation and benefits, promotions, hours and schedules, and on strategic issues, such as organizational values and use of technologies. Even here, however, a substantial minority of approximately 40% report that they expect a lot or unlimited say in some areas. Moreover, more than 50% believe they should have a lot or unlimited say on issues affecting their personal safety, freedom from abuse and discrimination, and respect on the job. Thus, results show no support for the argument that contemporary workers lack interest in or have diminished expectations for having a meaningful voice over workplace issues.

Is There a Voice Gap, or Have Employers Provided Adequate Channels for Input?

To address the question of whether there is a voice gap, we first describe workers’ actual say (as they report it) on a variety of issues and then compare expected and actual say. The results in Figure 2 show that actual voice varies across workplace problems. Indeed, differences in the magnitude of actual say varies more across these issues than do differences in the magnitude of expected say. Workers have the least influence over their benefits, compensation, and promotion opportunities. For example, 62% indicate that they have no or little say over their benefits, and 59% indicate that they have no or little say over compensation. By comparison, only 18% of workers indicate they have no or little say over workplace safety issues.

The results presented in Table 2 also indicate that sizable voice gaps exist across the full set of workplace issues measured in the survey, and the voice gap is the largest among traditional collective bargaining issues. Between 50 and 62% of the respondents report a voice gap for benefits, compensation, promotion, job security, technological change, and protections against harassment. Voice gaps for all the other issues vary between 35% (control over how they do their job) and 49% (access to training) of the respondents. The average distance between expected say and actual say regarding benefits is 1.07, meaning that the average difference is a full response category (e.g., expect a lot of say but have some say or expect some say but have little say).

Table 2.

Voice Gap: Average Appropriate Say, Average Actual Say, Average Voice Gap, and Share of Workers Experiencing a Voice Gap, by Workplace Issue

Workplace issue	Average appropriate say	Average actual say	Average voice gap	Percentage experiencing voice gap (%)
Workplace/personal
Respect toward employees	3.80	2.98	0.83	54
Harassment protections	3.88	3.10	0.79	52
Discrimination protection	3.78	3.07	0.70	49
How to improve work	3.49	2.88	0.60	48
Ability to resolve problems	3.60	3.12	0.48	43
Time to do job	3.34	2.90	0.44	41
Ability to organize schedule	3.23	2.84	0.39	39
Choose how to do job	3.39	3.10	0.30	35
Average of workplace/personal issues	3.56	3.00	0.56	45
Personnel/bargaining
Benefits	3.25	2.18	1.07	62
Compensation	3.29	2.28	1.01	62
Opportunities for promotion	3.25	2.37	0.88	57
Job security	3.48	2.67	0.82	55
Access to training	3.46	2.80	0.67	49
Ability to perform job safely	3.90	3.37	0.53	45
Average of personnel/bargaining issues	3.44	2.62	0.82	55
Organizational Strategy
How new technology impacts job	3.16	2.44	0.72	52
Employer values	3.30	2.66	0.65	50
Quality of products	3.37	2.79	0.58	47
Average of organizational strategy issues	3.27	2.63	0.65	50

Notes: Average appropriate and actual say can take the value range [1,5]. Average voice gap can take the range of [–4,4] and is calculated over all workers, including those with zero or negative voice gaps. Percentage experiencing voice gap represents the share of respondents who report having a positive voice gap, i.e., they report having say on an issue that is lower than the say they think is appropriate.

The 1995 and 2017 surveys allow for comparisons of similarities and differences in the voice gap on seven of the common issues included in those surveys. Caution is in order, however, as the wording of the questions varied somewhat in the two surveys and the sample differs as well. For the 1995 survey, Freeman and Rogers (1999) screened out lower-level managers, public-sector workers, and workers at firms with 24 or fewer employees because, they argued, such groups might prefer different institutions altogether.

As shown in Table 3, similar gaps were reported across years for five of the seven issues; in other words, we see similar percentages who have less say than they want and the gap is of similar magnitude for many of the issues captured by our survey and by Freeman and Rogers (1999). The exceptions are benefits and pay, for which the Worker Voice Survey measures substantially smaller voice gaps. For benefits, in particular, this reflects a lower expectation for having a say in the current period, which might reflect the reality of the decline of pensions and changes in health insurance coverage. Workers also have lower expectations on training in the current period, perhaps because of the decline of internal labor markets.

Table 3.

Comparing the Freeman and Rogers (F & R) Representation Gap to Worker Voice Survey (WVS) Voice Gap, by Workplace Issues (percentage)

	Appropriate say		Actual say		Average gap		Individual gap
Workplace issue	Percentage of workers for whom it is very important to have a lot of influence		Percentage of workers who said they had a lot of direct influence and involvement		Difference between the columns “wanting influence” and “having influence”		Percentage experiencing representation or voice gap
	F & R	WVS	F & R	WVS	F & R	WVS	F & R	WVS
1. Deciding what kind of benefits are offered to employees	60	39	6	12	54	27	83	61
2. Deciding how much of a raise in pay the person in your work group should get	41	39	6	12	35	27	76	61
3. Deciding what training is needed for people in your work group or department	62	51	29	27	33	24	53	48
4. Deciding how to work with new equipment or software, if that has ever been needed	52	37	28	16	24	21	46	51
5. Setting goals for your work group or department	55	53	32	27	23	26	43	47
6. Setting safety standards and practices	55	74	35	49	20	25	45	37
7. Deciding how to do your job and organize the work	76	48	57	37	19	11	31	33
8. Setting work schedules, including breaks, overtime, and time off	42	40	30	30	12	10	47	37

Notes: The scale of Freeman and Rogers’s 1995 survey (4-point scale) differs from our 2017 survey (5-point scale). For appropriate say, Freeman and Rogers’s survey asks, “How important would it be to you to have influence in a certain issue?” (1 = very important; 2 = somewhat important; 3 = not too important; 4 = not important at all). For actual say, Freeman and Rogers’s survey asks, “How much direct involvement and influence do you have in a certain issue?” (1 = a lot of direct involvement and influence; 2 = some direct involvement and influence; 3 = only a little direct involvement and influence; 4 = no direct involvement and influence). To make the two surveys comparable, we combined “unlimited say” and “a lot of say” categories of the appropriate say and actual say questions in our 2017 survey. Percentage experiencing representation or voice gap represents the share of respondents who report having say on an issue that is lower, by any size, than the say that they think is appropriate. Worker voice survey questions can be found in Appendix Table A.2.

We also used the 2017 data to investigate who experiences the greatest voice gap, considering demographic background, employment status, and union status. Here we grouped the issues into the three categories introduced earlier: workplace/personal issues, personnel/collective bargaining issues, and organizational strategy issues.⁹ We find that women reported a consistently larger gap compared to men across all issues, as evidenced by the positive and statistically significant coefficients in Table 4. In the case of workplace/personal issues, for instance, the 0.093 coefficient translates into a 0.11 standard deviation higher gap for women than for men. Additional analyses show that women believe they ought to have more say, as compared to men, but were no more likely to have that say, creating a gender difference in the voice gap.

Table 4.

OLS Regressions of Voice Gaps on Different Sets of Issues

Variable	Workplace/Personal	Personnel/Bargaining	Organizational strategy
Age	−0.000	0.003***	0.002*
	(0.001)	(0.001)	(0.001)
Female	0.093***	0.094***	0.125***
	(0.030)	(0.033)	(0.035)
Education (ref: High school)
No high school diploma	−0.026	0.056	0.047
	(0.097)	(0.106)	(0.113)
Some college	0.013	0.053	0.066
	(0.044)	(0.048)	(0.051)
BA or above	0.046	0.056	0.108*
	(0.048)	(0.052)	(0.055)
Race and ethnicity (ref: White)
Black	0.038	0.036	0.017
	(0.040)	(0.044)	(0.047)
Hispanic	−0.031	−0.081*	−0.043
	(0.042)	(0.046)	(0.049)
Asian, Other	−0.146**	−0.229***	−0.178**
	(0.062)	(0.067)	(0.071)
Two or more races	0.213***	0.142*	0.133
	(0.074)	(0.080)	(0.086)
Represented by union	0.146***	0.011	0.099**
	(0.037)	(0.040)	(0.043)
Type of employment (ref: Regular full-time)
Regular part-time	−0.093**	−0.161***	−0.059
	(0.044)	(0.048)	(0.051)
Temporary employee	−0.074	0.029	0.007
	(0.089)	(0.097)	(0.104)
Contract employee	0.172**	0.078	0.022
	(0.077)	(0.084)	(0.090)
Independent contractor	−0.126	−0.284***	−0.158
	(0.087)	(0.094)	(0.101)
Tenure at current employer (years)	0.002	0.003	0.001
	(0.002)	(0.002)	(0.002)
Primary job earnings (ref: < $30,000)
$30,000–$50,000	0.137***	0.125***	0.165***
	(0.038)	(0.042)	(0.044)
$50,000–$75,000	0.038	−0.004	0.057
	(0.045)	(0.049)	(0.052)
$75,000–$110,000	−0.002	−0.067	0.064
	(0.055)	(0.060)	(0.063)
> $110,000	−0.073	−0.088	−0.021
	(0.067)	(0.073)	(0.078)
Observations	3,476	3,475	3,461

Notes: Standard errors in parentheses. All models include region of residency, occupation, industry, and establishment size controls. Dependent variables are the average voice gap across issues in each of the three categories. Personnel/bargaining issues include compensation, benefits, training opportunities, promotion opportunities, job safety, and job security. Workplace/personal issues include scheduling, time to do work, how to do job, how to improve work, resolve problems affecting ability to do job, discrimination protections, harassment protections, and respect toward employees. Organizational strategy issues include how new technologies affect job, quality of employers’ products or services, and the basic values the employer stands for.

p < 0.1; **p < 0.05; ***p <0.01.

By contrast, few differences exist between respondents in the larger race and ethnic groups, though respondents who were not white, black, or Hispanic¹⁰ experienced lower voice gaps across the board. Somewhat surprisingly, after adjusting for other factors, union members have a larger voice gap on both personal issues (such as respect, safety, and protection from harassment and discrimination) and organizational strategy issues (such as use of technology and organizational values) than do non-union workers. For organizational strategy issues, union workers’ expected say is higher than what non-union workers expect. This finding suggests that union workers believe it is appropriate for them to weigh in on how the organization pursues its goals, whereas unionized employers normally will see these as “management rights” that lie outside the mandatory (legally required) scope of bargaining. For personal issues (including protection from harassment and discrimination), union members’ expected say is similar to non-union workers—so the larger gap reflects union members’ reporting less actual say than others do. This outcome is consistent with results obtained in the 1977 Quality of Employment Survey (see Kochan 1980: 376). Then and now, unions have apparently not been effective in meeting the expectations union members have for gaining a voice on these issues.

Older workers report a significantly larger voice gap for two of the three workplace issues. Moderate earners, making $30,000 to $50,000, report more of a voice gap than their lower-income counterparts. Moderate earners are more likely to believe they ought to have a say in workplace issues than do workers in the lowest income category, that is, their expectations regarding voice are greater. The voice gap for collective bargaining and personnel issues is significantly lower for independent contractors, compared to regular full-time workers. Part-time workers also have smaller voice gap on both collective bargaining and personal issues compared to full-time workers. Taken together, the results reported from Tables 2 to 4 indicate that American workers continue to experience a sizable voice gap. These data do not support the hypothesis that workers are satisfied with their voice at work.

Are Workers Interested in Joining a Union or Not?

As noted earlier, unions have declined precipitously from when the 1977 Quality of Employment Survey was conducted and also from 1995 when the Freeman and Rogers survey was conducted. This decline raises the question of whether workers still have an interest in being represented by a union. The answer from our survey is yes, and considerably more so today than reported in the two prior surveys.

We replicated a question from both the 1977 and the 1995 surveys asking whether the non-union workers in the sample would join a union if a vote was held on their job. Although our questions are comparable across years, the sample does differ a bit. Freeman and Rogers screened out public-sector workers and workers at firms with 24 or fewer employees and did not ask any managers about union interest. We retain lower-level managers, public-sector workers, and those working in small firms in the 2017 sample (as did the 1977 study). Given the Janus v. AFSCME decision, in particular, the opinions of public-sector workers regarding unions are relevant for those unions’ survival, and we cover a broader array of possible voice options that might be more relevant to workers in small firms.

As shown in Figure 3, the 1977 and 1995 surveys produced nearly identical levels regarding workers’ union support: About one-third of the non-union, non-managerial workforce said they would vote to unionize if given the opportunity to do so. By 2017, that percentage had increased to almost half of respondents—48% when we limit the analysis to non-union respondents and 49% when we consider non-managerial, non-union respondents.¹¹ Even when we restrict the 2017 sample to be more comparable to that of the 1995 study, we still find union support seems to have increased since the previous surveys.¹² Moreover, a strong majority (83%) of currently unionized workers said they would vote for a union again. This number was slightly lower than the 90% of union members who reported they would do so in the 1995 Freeman and Rogers survey. (This question was not asked of union members in the 1977 survey.) Thus, despite the decline in union representation and the political and policy hurdles for organizing today, interest in joining a union has increased considerably.

To explore potential reasons for this increase in interest in union representation, we replicated as closely as possible the multivariate analyses of the demographic determinants of the union vote in the 1977 and 1995 studies and compared them to results from the current survey. For the 2017 results, we display three specifications that include a base estimation, then add respondents’ average voice gap in the next model, and finally add respondents’ view of unions’ effectiveness in the final model.

As shown in the odds ratios reported in Table 5, results across the three surveys are remarkably similar. Across the three time periods, minority workers are consistently and strongly more interested in joining a union than are white workers. In 2017, for example, for two otherwise identical workers, the odds that minority workers would vote for unions were 2.4 to 3.5 times greater than the odds were for their white counterparts (see Farber and Saks 1980 for similar estimates on race). By comparing models (4) and (5), we can see that at least a portion of the black–white gap in union support is explained by the fact that minorities perceive unions to be more effective than do white respondents; the coefficient for black respondents drops by about one-third when perceived union effectiveness is included in the model. Across all three time periods, lower income workers are also more likely to vote to join a union than would their higher income counterparts.

Table 5.

Logit Regression: Non-Union Workers Who Would Vote for a Union in 1977, 1995, and 2017 (odds ratios)

	1977	1995	2017	2017 (+voice gap)	2017 (+voice gap, union effectiveness)
Variable	(1)	(2)	(3)	(4)	(5)
Age	0.998	0.990	0.988***	0.987***	0.996
	(0.00811)	(0.00651)	(0.003)	(0.004)	(0.004)
Female	0.972	1.611***	0.870	0.822**	0.871
	(0.217)	(0.249)	(0.083)	(0.080)	(0.093)
Education (ref: High school)
No high school diploma	0.797	2.100***	1.491	1.482	1.381
	(0.215)	(0.582)	(0.467)	(0.474)	(0.478)
Some college	0.662*	0.949	1.079	1.054	1.079
	(0.162)	(0.154)	(0.147)	(0.147)	(0.164)
BA or above	1.136	0.626**	1.205	1.167	1.332*
	(0.347)	(0.125)	(0.177)	(0.175)	(0.219)
Race and ethnicity (ref: White)
Black	4.598***	3.584***	3.418***	3.475***	2.333***
	(1.545)	(0.726)	(0.448)	(0.467)	(0.342)
Hispanic			2.303***	2.402***	2.030***
			(0.305)	(0.325)	(0.303)
Asian, other race/ethnicity, or two or more races			2.535***	2.687***	2.500***
Asian, other race/ethnicity, or two or more races			(0.382)	(0.416)	(0.428)
Asian		2.481***
		(0.711)
Other race/ethnicities	2.371
	(1.515)
Annual salary (ref: < $30,000)
$30,000–$50,000	1.075	0.677*	0.913	0.848	0.813*
	(0.246)	(0.137)	(0.099)	(0.094)	(0.099)
$50,000–$75,000	0.494**	0.253***	0.708***	0.681***	0.655***
	(0.165)	(0.118)	(0.094)	(0.092)	(0.098)
> $75,000	0.331**	0.515	0.476***	0.462***	0.483***
	(0.155)	(0.218)	(0.073)	(0.073)	(0.083)
Tenure (ref: Been at job less than a year)
Been at job 1–3 years	0.943	0.885	0.851	0.823	0.938
	(0.237)	(0.166)	(0.117)	(0.116)	(0.146)
Been at job 3–5 years	1.159	0.801	0.807	0.767*	0.805
	(0.341)	(0.177)	(0.122)	(0.119)	(0.138)
Been at job 5–10 years	0.924	0.662**	0.633***	0.615***	0.679**
	(0.271)	(0.137)	(0.094)	(0.093)	(0.113)
Been at job 10+ years	0.460**	0.843	0.568***	0.548***	0.609***
	(0.155)	(0.194)	(0.084)	(0.082)	(0.101)
Likelihood of finding alternative job (ref: Very likely)
Fairly likely	1.051	1.032	1.118	1.078	0.942
	(0.217)	(0.163)	(0.131)	(0.130)	(0.125)
Not too likely	0.996	1.252	1.655***	1.441***	1.394**
	(0.240)	(0.214)	(0.207)	(0.186)	(0.199)
Establishment size (ref: < 100 employees)
100–499 employees	1.094	0.919	0.976	0.882	0.781**
	(0.272)	(0.174)	(0.102)	(0.095)	(0.094)
500–999 employees	1.210	0.904	0.998	0.865	0.661**
	(0.449)	(0.215)	(0.159)	(0.142)	(0.118)
1,000+ employees	1.432	0.912	1.016	0.941	0.749
	(0.412)	(0.160)	(0.176)	(0.167)	(0.145)
Average voice gap				1.847***	2.104***
				(0.111)	(0.142)
Average rating of union effectiveness					2.113***
					(0.105)
Constant	0.315*	1.319	3.546***	2.903***	0.292***
	(0.189)	(0.804)	(1.384)	(1.159)	(0.135)
Observations	674	1,203	2,743	2,742	2,527

Notes: Standard errors in parentheses. All models control for industry, occupation, and region. All models are unweighted. 1995 Survey: (a) For hourly paid workers, annual salary = weekly salary × 52; (b) likelihood of finding alternative jobs is a categorical variable, equals 1 if the respondent is very confident in finding alternative jobs, 2 if somewhat confident, 3 if not too confident, 4 if not confident at all. For the purpose of parallel analysis, we grouped 3 and 4.

***

p < 0.01; **p < 0.05; *p < 0.1.

In 2017, younger workers were slightly more likely to want to join a union than were older workers, which contrasts with the argument that younger workers may have accepted employer decisions as necessary reflections of broader “market” forces, or that they have been exposed to cultural expectations that individual workers, rather than collectives, are responsible for their situation at work. Indeed, this finding is consistent with recent research showing that most new membership gains are made up of younger workers (Schmitt 2018). Those with a four-year college degree or higher (controlling for other characteristics) are more interested in union representation than are those with a high school degree (or even less education), albeit this contrast is statistically significant at only the 10% level in one specification (column (5)). With income and occupational variables included in this equation,¹³ this coefficient may be capturing the effects for those with a college degree who are in low-wage occupations, that is, the more underemployed segment of recent college graduates. In addition, although women were more likely to vote for a union in 1995, in 2017 the association between women and union support is statistically insignificant. However, women are more supportive of unions in the 2017 data before occupational controls are introduced—suggesting that women are clustered in occupations with higher relative support for unions (e.g., health care practitioners, teachers, food service workers, and so forth). In all three surveys, those who believe it would be hard to find an equivalent job in the external labor market are somewhat more interested in union representation than are those with better external prospects, although these coefficients are statistically significant in only the 2017 survey.¹⁴

In 2017, those with larger voice gaps (averaged across all issues) are significantly more interested in union representation than are those with lower voice gaps.¹⁵ Finally, those who see unions as more effective voice mechanisms for addressing their top priority concerns are more likely to want union representation than those who see unions as less effective in addressing their priority concerns. In a similar vein (but with somewhat different measures), the 1977 study found that workers who experienced higher job dissatisfaction and those who viewed unions as instrumental for improving working conditions were more likely to vote for a union (Kochan 1979).

Taken together, the similarity in the results across these time periods suggests that the increased percentage of workers interested in joining a union is likely caused by an increase in the number of non-union workers with characteristics and working conditions that have consistently predicted union interest. That is, workers have not changed their views on unions; instead it seems that the decline in unions results in a larger number of unorganized workers who are very similar to prior generations who supported union representation.

Thus, the decline in the number of workers joining unions cannot be attributed to lack of interest in union representation. To put these findings in perspective, if all of the non-union workers who have a desire to join a union had the opportunity do so, union membership could increase by approximately 58 million workers, essentially quadrupling the number currently represented by a union, which would raise union density to 54%. Of course, that is not realistic given the demonstrated difficulty of organizing under the current law and the effectiveness of employer resistance to organizing efforts, a point we will return to in the final section of this article.

Are New Forms of Worker Voice Filling the Voice Gap?

Are the options provided by employers and/or those emerging independently filling the void left by union decline and meeting the interests and needs of workers today? To answer this question we first conducted a factor analysis (reported in Appendix Table A.3) of respondents’ assessment of the effectiveness of the full array of options included in the survey. The voice options cluster into two groups. One group captures independent options, such as unions, occupational associations, petitions, protests, and strikes. The other group clusters around internal options that are facilitated or at least implicitly supported by employers, such as talking with a supervisor, conferring with people like themselves, and utilizing grievance and ombudsman processes.

We assess these options by examining the use of, and satisfaction with, internal and independent options. As shown in Table 6, when faced with workplace issues, the vast majority turn first to their supervisors and coworkers for assistance. These two options are available and used by 60 to 70% of the respondents. Use of the other channels then falls dramatically to less than 20%. For example, only 6% of the sample have participated in strikes to address issues at work.

Table 6.

Percentage of Workers Who Used Each Voice Channel

Voice channel	Percentage (%)
Supervisor	71
People like you	64
Joint committee	17
Union	16
Grievance	15
Occupation association	15
Ombudsman	13
Petition	10
Online rating	10
Demographic association	10
Protest/rally	7
Strike	6

Notes: Based on Worker Voice Survey question q4: “In order to deal with workplace issues at your primary/current workplace, have you ever decided to [use voice mechanism]?” Sample restricted to those with valid answers that included “Yes” or “No.”

Table 7, columns (1) and (2), show the logit equation predicting workers’ use of any of the internal or independent voice mechanisms.¹⁶ To assess how characteristics affected the number of each set of mechanisms used, we ran ordinary least squares (OLS) models for the counts of internal mechanisms and independent mechanisms used.¹⁷ Women are approximately 1.5 times as likely as men to use an internal voice mechanism, and there is no statistically significant difference in their usage of any independent mechanisms. People who have attended college are more likely to use both internal and independent voice mechanisms (albeit, the estimates are imprecise for the number of mechanisms used). Black workers are significantly more likely than are white workers to use independent voice mechanisms, being 40% more likely to use any of them. Union members are more likely to use internal and independent voice mechanisms than are non-union workers.

Table 7.

Logit and OLS Regressions: Use of Internal and Independent Mechanisms

	Logit (Used any = 1; used none = 0)		OLS (Number of mechanisms used)
	Odds ratio	Odds ratio	Coefficient	Coefficient
	Internal mechanisms^a	Independent mechanisms^b	Internal mechanisms^a	Independent mechanisms^b
Variable	(1)	(2)	(3)	(4)
Age	1.003	1.002	−0.001	−0.001
	(0.004)	(0.004)	(0.002)	(0.002)
Female	1.506***	1.117	0.186***	0.099**
	(0.178)	(0.111)	(0.046)	(0.048)
Education (ref: High school)
No high school diploma	0.649	1.199	0.037	0.019
	(0.203)	(0.400)	(0.148)	(0.153)
Some college	1.523***	1.691***	0.222***	0.162**
	(0.243)	(0.261)	(0.068)	(0.070)
BA or above	1.441**	1.438**	0.076	0.089
	(0.252)	(0.239)	(0.074)	(0.077)
Race and ethnicity (ref: White)
Black	0.940	1.419***	0.109*	0.278***
	(0.152)	(0.187)	(0.063)	(0.066)
Hispanic	0.743*	0.954	−0.046	0.027
	(0.117)	(0.137)	(0.066)	(0.069)
Asian, Other	0.584**	0.912	0.083	0.121
	(0.132)	(0.189)	(0.098)	(0.102)
Two or more races	0.951	0.936	−0.023	0.074
	(0.269)	(0.219)	(0.111)	(0.115)
Represented by union	1.340*	9.860***	0.454***	1.274***
	(0.213)	(1.239)	(0.057)	(0.060)
Tenure at current employer (years)	1.028***	1.012**	0.016***	0.013***
	(0.009)	(0.006)	(0.003)	(0.003)
Primary job earnings (ref: < $30,000)
$30,000–$50,000	1.401**	1.163	0.120**	−0.051
	(0.207)	(0.150)	(0.059)	(0.061)
$50,000–$75,000	1.327	1.538***	0.228***	0.116
	(0.237)	(0.227)	(0.069)	(0.071)
$75,000–$110,000	1.140	1.448**	0.199**	0.121
	(0.248)	(0.262)	(0.084)	(0.087)
> $110,000	1.325	2.368***	0.276***	0.220**
	(0.368)	(0.506)	(0.103)	(0.107)
Likelihood of losing job (ref: Not at all likely)
Not too likely	1.043	1.166	0.081*	0.078*
	(0.119)	(0.113)	(0.045)	(0.046)
Fairly likely	1.132	2.303***	0.223***	0.430***
	(0.231)	(0.387)	(0.084)	(0.086)
Very likely	2.091**	3.099***	0.417***	0.520***
	(0.645)	(0.696)	(0.111)	(0.115)
Likelihood of finding alternative job (ref: Very likely)
Fairly likely	0.741*	0.755**	−0.146**	−0.183***
	(0.115)	(0.097)	(0.060)	(0.062)
Not too likely	0.798	0.948	−0.155**	−0.161**
	(0.132)	(0.128)	(0.063)	(0.065)
Average effectiveness, by mechanism
Internal mechanisms	1.476***		0.264***
	(0.105)		(0.028)
Independent mechanisms		1.751***		0.372***
		(0.101)		(0.028)
Observations	3,256	3,246	3,256	3,246

Notes: Standard errors in parentheses. All models include region of residency, occupation, industry, employment category, and establishment size controls.

Internal mechanisms include: conversation with supervisor, advice from people like you, advice from ombudsman, file a grievance at the workplace, and join employee–manager committee.

Independent mechanisms include: join a union, sign petition, rate employer on online community, join occupation association, join demographic association, join protest or rally, and go on strike with others.

p < 0.1; **p < 0.05; ***p < 0.01.

Finally, and not surprisingly, those who view internal options as more effective are more likely to use them, and those who view independent mechanisms as more effective are more likely to use those. Taken together these results indicate that considerable variation exists in the voice options that workers choose, based on their demographic and occupational status and their degree of experience with unions.

Turning from use to satisfaction with the voice mechanisms, among the subsample that used them, none of the voice options receives high satisfaction ratings. As shown in Table 8, supervisors, unions, and coworker options receive the highest satisfaction rankings and ratings (an average of 3.08 to 3.05, meaning just above “somewhat satisfied”) followed by occupational associations and strikes. The lowest satisfaction rating was given to grievance processes. Again, differences exist between groups of workers in how they rated their satisfaction with various mechanisms. Notably, union-represented workers rated unions as their most satisfactory option (mean of 3.17) compared to it being the seventh highest in satisfaction among workers who are currently not represented by a union (mean of 2.84). Union members also rate occupational associations higher, and supervisors and petitions somewhat lower, than do their non-union counterparts.

Table 8.

Mean Satisfaction Ratings across Voice Mechanisms, by Union Membership Status

Voice channel	All workers	Non-union	Union	Statistical significance
Supervisor	3.08	3.12	2.92	***
	(0.025)	(0.029)	(0.055)
People like you	3.05	3.07	3.00
	(0.021)	(0.023)	(0.049)
Ombudsman	2.75	2.75	2.75
	(0.059)	(0.071)	(0.108)
Grievance	2.68	2.64	2.75
	(0.060)	(0.074)	(0.103)
Joint committee	2.88	2.92	2.78
	(0.052)	(0.062)	(0.095)
Union	3.07	2.84^a	3.17	***
	(0.052)	(0.095)	(0.062)
Petition	2.72	2.82	2.54	*
	(0.072)	(0.086)	(0.121)
Online rating	2.74	2.69	2.85
	(0.074)	(0.086)	(0.144)
Occupation association	2.97	2.85	3.12	**
	(0.057)	(0.070)	(0.095)
Demographic association	2.92	2.86	3.04
	(0.062)	(0.073)	(0.119)
Protest/rally	2.75	2.68	2.81
	(0.085)	(0.124)	(0.117)
Strike	2.97	2.88	3.05
	(0.092)	(0.116)	(0.142)

Notes: Standard errors in parentheses. Based on Worker Voice Survey question q5: “Thinking about when you decided to [use voice mechanism] in order to address a workplace issue, how satisfied were you with the result?” Sample restricted to those with valid answers on the Likert scale (Not satisfied at all = 1, Not very satisfied = 2, Somewhat satisfied = 3, Very satisfied = 4, and Extremely satisfied = 5).

Answering this question implied having previously used this mechanism so in the case of non-union workers rating unions, we assume that these workers previously had experience with or were in unions.

p < 0.1; **p < 0.05; ***p < 0.01.

Table 9 presents regressions predicting workers’ satisfaction with their use of internal and independent voice mechanisms. We constructed the dependent variable by averaging the values of satisfaction across the internal and independent options.¹⁸ The resulting dependent variables we analyze here are therefore continuous and normally distributed, making the OLS regression appropriate for the analysis. Women are more satisfied with independent mechanisms as compared to men; the coefficient 0.168 translates into a nearly one-fifth standard deviation above the mean value for men. Compared to white workers, Hispanic workers and workers of other races are significantly less satisfied with their use of both internal and independent voice mechanisms, whereas black workers are much more satisfied with independent mechanisms as compared to their white counterparts. Union-represented workers are less satisfied with their use of internal voice mechanisms, but more satisfied with the use of independent voice mechanisms. In the case of column (4), being represented by a union was associated with more than a one-third standard deviation increase in satisfaction with independent mechanisms. Workers with higher incomes are more satisfied with internal voice mechanisms, but perhaps surprisingly, workers in larger firms are less satisfied with internal voice channels. Compared to regular full-time employees, regular part-time workers are more satisfied with their use of internal voice mechanisms (not shown in Table 9). Although the results suggest independent contractors are more satisfied with their use of internal and independent voice mechanisms, the results should be treated with caution given the absence of most types of internal voice mechanisms in most independent contractors’ workplaces as well as the small sample size of independent workers in the survey. Once again, these results point to considerable heterogeneity in satisfaction with the different options. The implication is that “no one-sized shoe” voice option fits all workers or circumstances.

Table 9.

OLS Regression: Satisfaction with Internal and Independent Mechanisms

	Average satisfaction with internal mechanisms^a		Average satisfaction with independent mechanisms^b
Variable	(1)	(2)	(3)	(4)
Age	−0.004***	−0.004***	−0.002	−0.002
	(0.001)	(0.001)	(0.002)	(0.002)
Female	0.041	0.070*	0.131***	0.168***
	(0.031)	(0.036)	(0.050)	(0.057)
Education (ref: High school)
No high school diploma	−0.054	−0.107	−0.052	−0.157
	(0.110)	(0.126)	(0.178)	(0.208)
Some college	−0.023	0.016	−0.101	−0.074
	(0.050)	(0.055)	(0.086)	(0.096)
BA or above	0.058	0.067	−0.140*	−0.115
	(0.049)	(0.060)	(0.084)	(0.105)
Race and ethnicity (ref: White)
Black	−0.056	−0.071	0.181***	0.154**
	(0.045)	(0.048)	(0.066)	(0.071)
Hispanic	−0.082*	−0.121**	−0.140*	−0.180**
	(0.049)	(0.052)	(0.079)	(0.084)
Asian, Other	−0.193***	−0.225***	−0.160	−0.147
	(0.072)	(0.076)	(0.109)	(0.118)
Two or more races	−0.200**	−0.235***	0.000	−0.071
	(0.084)	(0.088)	(0.138)	(0.149)
Represented by union	−0.126***	−0.069	0.278***	0.324***
	(0.039)	(0.044)	(0.050)	(0.059)
Establishment size (ref: < 10 employees)
11–499 employees		−0.142***		−0.040
		(0.043)		(0.078)
500–1,999 employees		−0.180***		−0.084
		(0.063)		(0.103)
2,000+ employees		−0.137**		−0.081
		(0.065)		(0.104)
Type of employment (ref: Regular full-time)
Regular part-time		0.237***		0.152
		(0.054)		(0.095)
Temporary employee		0.159		−0.065
		(0.115)		(0.175)
Contract employee		0.021		−0.006
		(0.096)		(0.140)
Independent contractor		0.327***		0.416**
		(0.108)		(0.187)
Tenure at current employer (years)		−0.004*		−0.003
		(0.002)		(0.003)
Primary job earnings (ref: < $30,000)
$30,000–$50,000		0.015		−0.162**
		(0.047)		(0.076)
$50,000–$75,000		0.075		−0.017
		(0.054)		(0.083)
$75,000–$110,000		0.093		0.046
		(0.066)		(0.102)
> $110,000		0.124		−0.117
		(0.080)		(0.124)
Constant	3.214***	3.218***	2.772***	2.902***
	(0.082)	(0.133)	(0.138)	(0.217)
Observations	3,242	2,925	1,357	1,217

Notes: Standard errors in parentheses. All models include region of residency, occupation, and industry controls.

Internal mechanisms include: conversation with supervisor, advice from people like you, advice from ombudsman, file a grievance, and join employee–manager committee.

p < 0.1; **p < 0.05; ***p < 0.01.

Discussion

Today’s workers expect to have a voice on the full spectrum of issues affecting how they work, how they are personally treated, their compensation and working conditions, the values their organization stands for, and the products or services they help produce or deliver. A gap continues to exist, however, between the amount of influence workers expect and what they experience. Although the gap occurs across all of the issues we examined, it is largest for compensation, benefits, promotions, and job security.

When confronted with a problem at work, more than 60% of workers have turned to their supervisors or coworkers. Still, fewer than 20% of workers have used any of the other options. This finding suggests that these other options are not available to most workers, that they are available but workers are not aware of them, or that these options are not perceived to be useful to most workers today. We found considerable variety in the use of, and satisfaction with, these options across different groups. For example, some demographic groups, such as women and black workers, tend to be more satisfied with their use of independent voice mechanisms. Further analysis not shown here (available from the authors on request) found ratings of the effectiveness of different options also varied across issues.

A majority of US workers today still see unions as a desirable channel to exercise voice. In fact, a substantially larger fraction of the non-union workforce would join a union today (48% of non-union workers) than would have done so in the past (about one-third in 1977 and in 1995). More than 80% of those currently represented by a union would vote to continue union representation. Moreover, the same general factors predict interest in joining a union now as in the past: Interest is greater among nonwhites, low-income workers, those who have larger voice gaps, and those who worry they could not find equivalent jobs in the external labor market if they needed to do so.

Thus, there continues to be a large unmet demand for union representation among American workers. Yet, it is unlikely that this unmet demand will be satisfied under current conditions given the legal requirement of obtaining majority support for union representation within a specified bargaining unit before any individual worker gains representation and given the success of employer opposition to organizing through this process. Either labor law changes will be needed, or employer opposition will need to be overcome. Alternatively, unions will need to pursue organizing strategies that overcome the constraints of both labor law and employer opposition. Worker advocacy groups and other independent voice options have grown in recent years because they are not encumbered by the limits of the organizing model and the legal processes needed to achieve collective bargaining status. Yet to date, these alternatives have reached only a small fraction of the American labor force. This may change as advocates of these newer voice channels learn from early experiences and/or as existing unions experiment with new or similar strategies on their own or in coalition with other labor advocates. But at the present, a large number of workers appear to have no independent option available for exercising their voices on the issues that they believe they ought to be able to influence.

Moreover, the data suggest that today “no one-sized shoe” fits all workers or all issues in play in employment relationships. Some workers are more likely to use and to be more satisfied with internal options provided by employers; others use and are more satisfied with independent options provided by unions or worker advocates independent of employers. Many workers see internal options as effective for some issues and independent options as important for other issues. This point is particularly important because it suggests the value of developing and diffusing multi-option systems of voice and representation. By contrast, labor law now limits internal forms of worker voice that violate bans on employer-supported or dominated labor organizations. Many employers strongly resist and suppress workers’ efforts to form unions or to engage in other options for exercising an independent voice. Many unions, in turn, see internal voice mechanisms as efforts to undermine or avoid union representation. The data from our survey suggest that many workers do not share these distinctions that are salient in law or in practice. Therefore, many workers might respond favorably to systems that mix these options together in an effective fashion.

Comparing these contemporary results with prior surveys demonstrates that voice gaps have persisted for a long time but, in the specific instance of the unmet demand for union representation, have grown considerably. The range of innovative efforts to address this persistent problem is growing, yet our results make it clear that considerable work remains to be done to close the voice gaps present for many at work today.

Conclusions and Future Directions

We see our survey data as providing only a broad overview of the current state of worker voice and the options for closing the voice gaps identified. More intensive analyses of options offered inside firms and those being pursued by various worker advocacy organizations and/or unions are clearly needed. For example, given the increased interest in union representation, it would be useful to develop a better understanding of ways to make unions more accessible, what forms of union representation would be most attractive to prospective members, or what workplace or labor market services would workers most value (i.e., be willing to pay for) from unions. To turn the interest in unions into an increase in union membership and representation may require shifting to an organizing model that does not require obtaining support of a majority in a specific work or occupational setting (Morris 2005) and one that does not lose members if or when they leave a union-represented job or employer (Budd 2010; Kochan 2011). Given the evidence that workers tend to look to the most available of internal options—supervisors and coworkers—more efforts to improve the availability and/or the quality of other internal options, such as joint committees, ombudsman services, or grievance procedures, appear to be warranted.

Given the findings that suggest a “one-sized shoe” does not fit all groups or issues, another priority for further research should be to seek a better understanding of how different options can be provided as complements in a system of voice and representation that gains and sustains workforce trust (Rowe 1987; Lipsky 2015). Is there some complementary mix of internal and independent options that would serve the workforce better than does the current situation in which most employers favor internal options and seek to avoid independent options, whereas unions see internal channels as employer-dominated efforts to substitute for, or be competitive with, union representation?

We hope the results reported here motivate others to address these and other questions they raise in search of ways to close the voice gaps American workers continue to experience today.

Footnotes

Appendix

Table A.3.

Factor Loadings of the Effectiveness of Voice Mechanisms, by Issue

Mechanism	Independent mechanisms	Internal mechanisms
Effectiveness on issue 1 ^a
Supervisor		0.7011
People like you		0.6694
Ombudsman		0.6367
Grievance		0.6539
Joint committee		0.6605
Union	0.6403
Petition	0.5741	0.5433
Online rating	0.5607
Occupation association	0.6541
Demographic association	0.6401
Protest/rally	0.9184
Strike	0.9197
Effectiveness on issue 2 ^a
Supervisor		0.7513
People like you		0.7195
Ombudsman	0.5417	0.5730
Grievance		0.6467
Joint committee	0.5178	0.6174
Union	0.7370
Petition	0.7414
Online rating	0.6997
Occupation association	0.7363
Demographic association	0.7235
Protest/rally	0.8599
Strike	0.8259

Notes: Because the variables of interest are not continuous but ordinal-categorical, a polychoric correlational matrix was employed in the factor analysis and the rotational method was done using the varimax method with Kaiser Normalization. The Kaiser Criterion is a reliable test for significance if the averaged extracted communality (1-Uniqueness) is equal to or greater than 0.60 and the sample size is 250 observations or more (Yong and Pearce 2013), both of which are met with our data. Factor loadings < 0.5 are suppressed. The Cronbach alpha reliability coefficients on internal and independent mechanisms (across both issue questions) are 0.90 and 0.94, respectively, which indicates a satisfactory internal consistency.

Issue 1 and issue 2 are randomly selected from those issues to which respondents reported “unlimited” or “a lot” of appropriate say. These factor analysis results are consistent with those performed on individual issues. The results are available upon request.

For information regarding the data and/or the computer programs utilized for this study, please address correspondence to tkochan@mit.edu.

1

Unfortunately, the 1977 survey did not ask how much say or influence workers actually experienced on their jobs so we cannot make voice gap comparisons with that time period.

2

Our sampling strategy did not screen out managers below this high level. Our coding of reported occupations revealed that most of the managers in the sample are what would be called frontline managers, that is, managers of departments.

3

In combination with the AmeriSpeak base panel weight, NORC adjusted for nonresponse bias in the Worker Voice Survey by adjusting the weight with a raking ratio method to the employed adults age 18+ population totals along the following sociodemographic characteristics: age, sex, education, race/ethnicity, and Census Division.

4

Regressions using the weighted data produced no substantive differences from the unweighted results.

5

Another 4% of our sample identified themselves as both standard full-time or part-time employees and as self-employed. We retain these workers in the standard full- or part-time categories in the analysis that follows to be conservative about our understanding of those in alternative work arrangements.

6

According to the Bureau of Labor Statistics, in 2017, there were 10.6 million independent contractors (6.9% of total employment), 2.6 million on-call workers (1.7% of total employment), 1.4 million temporary help agency workers (0.9% of total employment), and 933,000 workers provided by contract firms (0.6% of total employment).

7

We created these groupings on conceptual grounds and not on the basis of distinct clusters derived from a factor analysis. A factor analysis showed that all 17 issues clustered on a single factor with an Eigen value of 7.69. No second distinct factor emerged. This finding suggests that workers tend to see these issues as components of a single interrelated system of workplace practices, employment conditions, and experiences.

8

Workplace/personal issues include scheduling, time to do work, how to do job, how to improve work, resolve problems affecting ability to do job, discrimination protections, harassment protections, and respect toward employees. Personnel/bargaining issues include salary, benefits, training opportunities, promotion opportunities, job safety, and job security. Organizational strategy issues include how new technologies affect job, quality of employers’ products or services, and the basic values the employer stands for. Alpha coefficients measuring the reliability of these indices are 0.861, 0.834, and 0.713, respectively.

9

The resulting dependent variables are therefore continuous and normally distributed, making the OLS regression appropriate for the analysis.

10

The data we received from NORC combined those respondents identified as Asian (3.65% of full sample), and those who marked “other” without further specification (1.6%). In this analysis (but not in earlier tables), we have combined the Asian and “other” respondents with those who report two or more races (3.5%).

11

All estimates of union support also exclude self-employed workers. Note, again, that the 1995 survey excluded public-sector workers and workers at small firms, and managers were not asked about their union support (: 69) whereas the 1977 survey and our 2017 survey included those workers.

12

When we attempt to apply the same restrictions as Freeman and Rogers, we find union support rates between 46% and 48% (depending on whether we exclude public administration workers, those at small firms, or also exclude workers in education because many of those workers are public employees).

13

The full equations showing the occupational variables in this table and equations to follow are available from the authors on request.

14

Difficulty of finding alternative jobs is coded here as a set of dummy variables indicating if a person reports being very likely to find alternative job, if a person is fairly likely to find an alternative job, or if the person is not likely to find a job with comparable benefits and salary.

15

There was little variation in how the three groups of issues individually affected union support.

16

Internal mechanisms include: conversation with supervisor, advice from people like you, advice from ombudsman, file a grievance, and joint employee–manager committee. Independent mechanisms include: join a union, sign petition, rate employer on online community, join occupation association, join demographic association, join protest or rally, and go on strike with others.

17

Because these variables are counts with many zeros and are not normally distributed, the ideal model to run would be the zero-inflated negative binomial (ZINB) regression, but this model did not converge for the internal mechanism count variable. The ZINB results for independent mechanisms were not substantively different from those findings from the logit and OLS models shown.

18

Alpha coefficients are 0.882 and 0.938 for internal and independent issues, respectively. Note that the question of satisfaction was answered by respondents only if they had indicated that they had used a given voice mechanism.

References

Ackers

Peter

. 2007. Collective bargaining as industrial democracy: Hugh Clegg and the political foundations of British industrial relations pluralism. British Journal of Industrial Relations 45(1): 77–101.

Arvins

Jeremy

Larcom

Megan

Weissbourd

Jenny

. 2018. New forms of worker voice in the 21st century. Accessed at http://iwer.mit.edu/wp-content/uploads/2018/01/New-Forms-of-Worker-Voice-IWER.pdf (March 2018).

Barley

Stephen R.

Kunda

Gideon

. 2006. Gurus, Hired Guns, and Warm Bodies: Itinerant Experts in a Knowledge Economy. Princeton, NJ: Princeton University Press.

Barry

Michael

Wilkinson

Adrian

. 2016. Pro-social or pro-management? A critique of the conception of employee voice as a pro-social behaviour within organizational behaviour. British Journal of Industrial Relations 54(2): 261–84.

Bashshur

Michael R.

Burak

. 2015. When voice matters: A multilevel review of the impact of voice in organizations. Journal of Management 41(5): 1530–54.

Benson

Alan

Sojourner

Aaron

Umyarov

Akhmed

. 2015. Can reputation discipline the gig economy? Experimental evidence from an online labor market. SSRN Scholarly Paper ID 2696299. Rochester, NY: Social Science Research Network.

Bronfenbrenner

Kate

. 1998. Organizing to Win: New Research on Union Strategies. Ithaca, NY: Cornell University Press.

Budd

John

. 2004. Employment with a Human Face: Balancing Efficiency, Equity, and Voice. Ithaca, NY: Cornell University Press.

Budd

John

. 2010. When do US workers first experience unionization? Implications for revitalizing the labor movement. Industrial Relations: A Journal of Economy and Society 49(2): 209–25.

10.

Budd

John W.

Bhave

Devasheesh

. 2008. Values, ideologies, and frames of reference in industrial relations. In Blyton

Paul

Bacon

Nicolas

Fiorito

Jack

Heery

Edmund

(Eds.), The Sage Handbook of Industrial Relations, pp. 92–112. Thousand Oaks, CA: SAGE Publications.

11.

Bureau of Labor Statistics. 2017. Union Members Summary. Accessed at https://www.bls.gov/news.release/union2.nr0.htm (January 2018). Economic News Release. Washington, DC: US Department of Labor.

12.

Bureau of Labor Statistics. 2018. Contingent and Alternative Employment Arrangements Summary. Accessed at https://www.bls.gov/news.release/conemp.nr0.htm (January 2018). Economic News Release. Washington, DC: US Department of Labor.

13.

Commission on the Future of Worker-Management Relations, U.S. 1994. The Dunlop Commission on the Future of Worker-Management Relations – Final Report. Federal Publications, December.

14.

Commons

John R.

1913. Labor and Administration. New York: Macmillan.

15.

Cordero-Guzmán

Héctor R

. 2015. Worker centers, worker center networks, and the promise of protections for low-wage workers. WorkingUSA: The Journal of Labor and Society 18(1): 31–57.

16.

Creed

W. E. Douglas

Scully

Maureen A.

2000. Songs of ourselves: Employees’ deployment of social identity in workplace encounters. Journal of Management Inquiry 9(4): 391–412.

17.

Detert

James R.

Burris

Ethan R.

2007. Leadership behavior and employee voice: Is the door really open? Academy of Management Journal 50(4): 869–84.

18.

Detert

James R.

Treviño

Linda K.

2008. Speaking up to higher-ups: How supervisors and skip-level leaders influence employee voice. Organization Science 21(1): 249–70.

19.

Estlund

Cynthia

. 2010. Regoverning the Workplace: From Self-Regulation to Co-Regulation. New Haven, CT: Yale University Press.

20.

Farber

Henry S.

Krueger

Alan B.

1992. Union membership in the United States: The decline continues. NBER Working Paper No. 4216. Cambridge, MA: National Bureau of Economic Research.

21.

Farber

Henry S.

Saks

Daniel H.

1980. Why workers want unions: The role of relative wages and job characteristics. Journal of Political Economy 88(2): 349–69.

22.

Ferguson

John-Paul

. 2008. The eyes of the needles: A sequential model of union organizing drives, 1999–2004. Industrial and Labor Relations Review 62(1): 3–21.

23.

Fine

Janice

. 2006. Worker Centers: Organizing Communities at the Edge of the Dream. An Economic Policy Institute Book. Ithaca, NY: Cornell University Press.

24.

Fiorito

Jack

. 2001. Human resource management practices and worker desires for union representation. Journal of Labor Research 22(2): 335–54.

25.

Foulkes

Fred K.

1980. Personnel Policies in Large Nonunion Companies. Englewood Cliffs, NJ: Prentice Hall.

26.

Fox

Alan

. 1966. Industrial Sociology and Industrial Relations: An Assessment of the Contribution which Industrial Sociology Can Make Towards Understanding and Resolving Some of the Problems Now Being Considered by the Royal Commission. London: H.M. Stationery Office.

27.

Fox

Alan

. 1975. Collective bargaining, Flanders, and the Webbs. British Journal of Industrial Relations 13(2): 151–74.

28.

Freeman

Richard B.

2007. Do workers still want unions? More than ever. Accessed at https://www.epi.org/publication/bp182/ (November 2017). Washington, DC: Economic Policy Institute.

29.

Freeman

Richard B.

Rogers

Joel

. 1999. What Workers Want. Ithaca, NY: Cornell University Press/ILR Press.

30.

Guest

David E.

1987. Human resource management and industrial relations. Journal of Management Studies 24(5): 503–21.

31.

Heckscher

Charles C.

1996. White-Collar Blues: Management Loyalties in an Age of Corporate Restructuring. New York: Basic Books.

32.

Heckscher

Charles C.

McCarthy

John

. 2014. Transient solidarities: Commitment and collective action in post-industrial societies. British Journal of Industrial Relations 52(4): 627–57.

33.

Hirschman

Albert O.

1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press.

34.

Kalleberg

Arne L.

2013. Good Jobs, Bad Jobs: The Rise of Polarized and Precarious Employment Systems in the United States, 1970s to 2000s. New York: Russell Sage Foundation.

35.

Klaas

Brian S.

Olson-Buchanan

Julie B.

Ward

Anna-Katherine

. 2012. The determinants of alternative forms of workplace voice: An integrative perspective. Journal of Management 38(1): 314–45.

36.

Kochan

Thomas A.

1979. How American workers view labor unions. Monthly Labor Review 102(4): 23–31.

37.

Kochan

Thomas A.

1980. Collective Bargaining and Industrial Relations: From Theory to Policy and Practice. Homewood, IL: R.D. Irwin.

38.

Kochan

Thomas A.

2011. Rethinking and reframing U.S. policy on worker voice and representation. ABA Journal of Labor & Employment Law 26(2): 231–48.

39.

Kochan

Thomas A.

Katz

Harry C.

McKersie

Robert B.

1986. The Transformation of American Industrial Relations. New York: Basic Books.

40.

Lewin

David

. 1987. Dispute resolution in the nonunion firm: A theoretical and empirical analysis. Journal of Conflict Resolution 31(3): 465–502.

41.

Lipset

Seymour Martin

Meltz

Noah M.

2004. The Paradox of American Unionism: Why Americans Like Unions More Than Canadians Do, But Join Much Less. Ithaca, NY: Cornell University Press.

42.

Lipsky

David B.

2015. The future of conflict management systems. Conflict Resolution Quarterly 33(S1): S27–S34.

43.

Machin

Stephen

Wood

Stephen

. 2005. Human resource management as a substitute for trade unions in British workplaces. Industrial and Labor Relations Review 58(2): 201–18.

44.

Marchington

Mick

. 2007. Employee voice systems. In Boxall

Peter

Purcell

John

Wright

Patrick

(Eds.), The Oxford Handbook of Human Resource Management, Chapter 12. Oxford, UK: Oxford University Press.

45.

Meyer

G. J.

1995. Executive Blues: Down and Out in Corporate America. New York: Franklin Square Press.

46.

Milkman

Ruth

. 2011. Immigrant workers, precarious work, and the US labor movement. Globalizations 8(3): 361–72.

47.

Morris

Charles J.

2005. The Blue Eagle at Work. Ithaca, NY: Cornell University Press/ILR Press.

48.

Morrison

Elizabeth W.

2011. Employee voice behavior: Integration and directions for future research. Academy of Management Annals 5(1): 373–412.

49.

National Labor Relations Board. 2018. Election Reports. Accessed at https://www.nlrb.gov/reports-guidance/reports/election-reports (March 2018).

50.

Newman

Katherine S.

1999. Falling from Grace: Downward Mobility in the Age of Affluence. Berkeley: University of California Press.

51.

[NORC] National Opinion Research Corporation. 2017. AmeriSpeak Panel. University of Chicago: National Opinion Research Corporation.

52.

Organ

Dennis W.

1988. Organizational Citizenship Behavior: The Good Soldier Syndrome. Lanham, MD: Lexington Books.

53.

Peterson

Christopher

Maier

Steven F.

Seligman

Martin E. P.

1993. Learned Helplessness: A Theory for the Age of Personal Control. Oxford, UK: Oxford University Press.

54.

Portes

Alejandro

Sensenbrenner

Julia

. 1993. Embeddedness and immigration: Notes on the social determinants of economic action. American Journal of Sociology 98(6): 1320–50.

55.

Pugh

Allison J.

2015. The Tumbleweed Society: Working and Caring in an Age of Insecurity. New York: Oxford University Press.

56.

Quinn

Robert P.

Staines

Graham L.

1979. The 1977 Quality of Employment Survey: Descriptive statistics with comparison data from the 1969–70 and the 1972–73 surveys. Ann Arbor: University of Michigan Press.

57.

Rolf

David

. 2016. The Fight for Fifteen: The Right Wage for a Working America. New York: New Press.

58.

Rowe

Mary P.

1987. The corporate ombudsman: An overview and analysis. Negotiation Journal 3(2): 127–40.

59.

Schmitt

John

. 2018. Biggest gains in union membership in 2017 were for younger workers. Accessed at https://www.epi.org/publication/biggest-gains-in-union-membership-in-2017-were-for-younger-workers/ (February 2018). Washington, DC: Economic Policy Institute.

60.

Sharone

Ofer

. 2013. Flawed System/Flawed Self: Job Searching and Unemployment Experiences. Chicago: University of Chicago Press.

61.

Smith

Vicki

. 2002. Crossing the Great Divide: Worker Risk and Opportunity in the New Economy. Ithaca, NY: Cornell University Press/ILR Press.

62.

Wartzman

Rick

. 2017. The End of Loyalty: The Rise and Fall of Good Jobs in America. New York: Public Affairs.

63.

Webb

Sidney

Webb

Beatrice

. 1897. Industrial Democracy. London: Longmans, Green, and Company.

64.

Yong

An Gie

Pearce

Sean

. 2013. A beginner’s guide to factor analysis: Focusing on exploratory factor analysis. Tutorials in Quantitative Methods for Psychology 9(2): 79–94.