Civic Organizations and the Political Participation of Cross-Pressured Americans: The Case of the Labor Movement

Abstract

Civic associations underpin American democracy. How can politically cross-cutting associations engage members who hold divergent viewpoints amidst increasing partisan polarization and nationalization of politics? I examine this question in the context of labor unions, studying how unions engage members who hold conservative views at odds with some of the union’s political actions. Using original surveys of local union presidents, members, and non-members along with in-depth interviews in selected local unions, I show how local union leaders can foster norms of participation among politically cross-pressured members. Norms of participation increase conservative members’ perceptions of political representation and engagement in politics, including participation in the union’s political action committee and support for union political mobilization. These findings have implications for understanding civic associations and participation in an era of political division, as well as the role unions continue to play in politics.

INTRODUCTION

Civic associations underpin American democracy, teaching citizens politically relevant skills and information, bringing citizens into contact with others from diverse backgrounds, and offering opportunities for political involvement (Han 2014; Han, McKenna, and Oyakawa 2021; Skocpol 2003; Verba, Schlozman, and Brady 1995). How can politically cross-cutting associations engage members who hold divergent viewpoints amidst increasing partisan polarization and nationalization of politics?

I examine this question in the context of the labor movement, studying how unions engage conservative and Republican members. Unions provide an important case for studying civic association and political action. They are one of the only associations that organizes and represents individuals in individuals’ identities as workers, and which represents the economic interests of working- and middle-class Americans in politics (Ahlquist 2017; Verba, Scholzman, and Brady 1995). Unions have historically fostered political activity among lower-paid, less-educated workers who might not otherwise participate in politics (Macdonald 2019; Rosenfeld 2014; Verba, Schlozman, and Brady 1995; but see Yan 2023). And even as their membership has declined, unions remain important political actors (Ahlquist 2017). This is particularly true for public sector unions, which despite legal cutbacks retain about a third of members across the states and are often a central interest group in local and state politics (Finger and Reckhow 2022; Hartney 2022; Moe 2011). The recent wave of strikes, including recent teacher strikes in traditionally conservative states, underscores the continued relevance of public sector unions.

Unlike other groups representing wealthy individuals or interests (like businesses), unions depend more heavily on mass membership for clout. But political polarization and nationalization have meant that unions now face the challenge of engaging more conservative or Republican members. Those conservative and Republican members may be uncomfortable with the political positions taken by unions, especially national union federations, which are increasingly identified with the Democratic party and with left-leaning causes (e.g., Finger and Reckhow 2022; Newman and Skocpol 2023; but see Zoorob 2019).

This challenge is particularly acute for teachers unions, with increasingly left-leaning members and leaders (Moe 2011, 87–98). This challenge has also been intensified by laws and court rulings that allow employees to reap benefits of unionization without paying dues (i.e., “right-to-work” laws). The 2018 Janus v. ASCFME Supreme Court decision applied right-to-work to all public sector unions, making the question of how unions can engage politically cross-pressured workers all the more salient. This is especially true since conservative organizations are capitalizing on Janus to persuade members to drop their union involvement.¹ Little research sheds light on how unions will counter such pressures, because of the substantial empirical challenges involved in studying membership and political education and engagement within individual unions. Typical national surveys used to study political participation rarely include information on the specific unions to which members belong, preventing scholars from understanding how variation in union practices might relate to union engagement (see also Mosimann and Pontusson 2017; Yan 2023).

In this paper, I take advantage of uniquely rich data on one statewide union—of educators—in one relatively conservative, long-time right-to-work state (Iowa), which includes original representative surveys of union members and non-members and a survey of local union leaders. (Local unions, typically representing school districts or community colleges, are the main unit for organizing members and collective bargaining; there are just over 400 such locals in Iowa.) Matched member–leader survey data, collected in 2018–19, permit me to examine variations in member political attitudes and participation and how they relate to the practices deployed by union leaders. A separate 2023 survey further tests hypotheses about the relationship between local union culture and members’ attitudes toward, and willingness to contribute to, union political activities. Building on the correlations identified in the survey data, I then conducted in-depth interviews with local union leaders to examine how specific practices deployed by local leaders shaped the political participation of ideologically cross-pressured members.

Both the survey and interview evidence point to the importance of local union leaders in establishing cultures that increase conservative and Republican teachers’ engagement in unions. By culture, I mean norms about participating in the union and the role politics plays in the union set by leaders (Ahlquist and Levi 2013; Han, McKenna, and Oyakawa 2021; Stepan-Norris and Zeitlin 2002; Voss and Sherman 2000). In a union with high participatory culture, teachers felt a strong norm for supporting the union, even if their political views or partisanship sometimes clashed with the union. Moreover, unions with strong participatory cultures also encouraged all members—including cross-pressured members—to engage in the union’s political activities, by making clear the connection between members’ political participation and their identity as educators and the connection between union political advocacy and local issues. Additionally, the federated structure of the union permitted local unions to talk about politics in ways tied to workers’ local concerns, rather than pitched national debates.

These findings offer contributions to several fields of study on collective action, civic organizations, and the labor movement. Most directly, they speak to research on how associations, including unions, affect the political attitudes and behaviors of members. I find that unions do not only exert an effect on the aggregate political views and participation of members, but that unions can do so for members who might otherwise be politically unaligned with the union. In addition, my study identifies specific mechanisms through which unions transmit values and norms to members. This argument builds on work by John Ahlquist and Margaret Levi (2013), as well as other labor historians and sociologists (e.g., Newman and Skocpol 2023; Stepan-Norris and Zeitlin 2002; Voss and Sherman 2000), documenting the ways that unions might alter the political values and attitudes held by their members depending on varying cultures and leadership styles (see also Mosimann and Pontusson 2017). My contribution is to connect variation in union culture to the political socialization that happens within unions, especially for cross-pressured members.

Beyond the labor movement, this paper speaks to a longstanding literature documenting how civic organizations serve as sites of political education and mobilization for Americans. My findings underscore how social relationships between members of organizations matter for political participation (Ganz 2009; Gerber, Green, and Larimer 2008; Han 2014; 2016; Han, McKenna, and Oyakawa 2021; Sinclair 2012), showing how certain organizational practices can change the calculus that ideological or partisan-cross-pressured individuals consider when deciding to contribute their time and resources to an organization.

Last, this paper speaks to debates over polarization and nationalization of politics in the United States, documenting how individual associations affect ideological, partisan, and other divisions in society (Baldassarri 2011). Political observers have bemoaned the segregation of Americans across party lines and especially the rise of strong partisan identities (Mason 2018), as well as the nationalization of politics (Hopkins 2018). Past research has also suggested that participating in politically cross-cutting social networks can be demobilizing (Mutz 2002). Because it organizes members on the basis of the workplace and work identities—historically crossing income, racial, and partisan lines—unions have united workers who might not otherwise have interacted closely with one another (Mondak and Mutz 2001; Mutz and Mondak 2006). The evidence here suggests important variation in the extent to which civic organizations can appeal to a politically diverse membership, bridging differences in partisan identity. This paper also implies that membership in cross-cutting civic organizations need not be demobilizing. And it shows how federation, long a critical feature of civic associations (Skocpol, Ganz, and Munson 2000), may play an important role in helping labor unions and other groups mitigate the challenges presented by nationalization and polarization.

THE ROLE OF ORGANIZATIONAL NORMS AND POLITICAL PARTICIPATION

Much of the literature on civic engagement has focused on the individual-level decisions that citizens make when choosing to participate in politics (Clark and Wilson 1961; Olson 1965; Verba, Schlozman, and Brady 1995). That approach stresses the costs and benefits facing individuals and has generated enduring insights about the individuals most likely to participate as well as the conditions that lend themselves to greater participation. At the same time, it left open important questions about the social context in which individuals decide to participate in politics—for instance, how social interactions and identities can change individuals’ motivations for engagement.

A subsequent wave of scholarship, on which this paper builds, has instead considered how civic organizations, including the choices and behaviors of their leaders, can shape their members’ political views and actions (Ganz 2009; Han, McKenna, and Oyakawa 2021; Skocpol 2003; Skocpol, Ganz, and Munson 2000). Not all organizations are equally equipped to spur greater activism, however, and their effectiveness depends on their leaders and structure. Hahrie Han and collaborators (Han 2014; 2016; Han, McKenna, and Oyakawa 2021), for instance, have documented how organizations that nurture stronger relational ties between their members are more likely to inspire activism.

These two approaches to studying civic engagement—individual-level decisions and organizational structures—need not be substitutes. As this paper will document, leaders of effective organizations can maximize the participation of individuals by fostering an internal set of norms and expectations that change the calculus of those members.

One important subject of research on organizational activism has been the labor movement. While never as widespread as in other rich democracies, the American labor movement has historically been one of the most important mass-membership political associations (Dark 1999; Lichtenstein 2002; Skocpol 2003). Even today, with union membership at around 10% of workers, unions remain one of the largest membership-based organizations representing the economic interests of working- and middle-class Americans (Rosenfeld 2014; Schlozman, Verba, and Brady 2012). Unions also remain highly politically active and, despite an unfavorable legal context, remain a potent force shaping elections and policymaking—especially in the case of teachers unions (Hartney 2022; Moe 2011).

Existing research using a variety of methods has documented how membership and participation in unions can increase politically relevant knowledge, change workers’ political positions on issues like trade and inequality, and boost the likelihood of participating in politics (e.g., Feigenbaum, Hertel-Fernandez, and Williamson 2019; Kim and Margalit 2017; Macdonald 2019; Rosenfeld 2014; but see Yan 2023).

The puzzle I take up is not whether unions can change the political behavior of their members. Rather, it is whether and under what conditions unions can foster political engagement among workers who might otherwise feel cross-pressured by their affiliation with the organization. By cross-pressured workers, I mean conservative and Republican workers who might feel conflicted about their support for an organization that overwhelmingly supports Democratic candidates and left-leaning issues.

Why might such workers feel cross-pressured? To be sure, since the New Deal, many unions have been aligned closely with the Democratic party coalition (Schlozman 2015). But this relationship has only tightened over time as the two parties have moved further from one another in their positions on labor and economic policy, mirroring broader polarization of interest groups (Crosson, Furnas, and Lorenz 2020; Dark 1999). While not all unions have increasingly supported left-leaning issues or candidates, most have and this is especially the case for teachers unions, arguably one of the most important segments of the labor movement (Moe 2011). The National Education Association (NEA), for example, is the largest labor union in the country, and from 1996 to 2023, the NEA has donated 93% of its campaign contributions to Democratic candidates. National representatives of the NEA have also staked out liberal positions on a range of policy issues—not just those immediately related to schools—such as on universal health care, voting rights, and immigration reform.²

For all these reasons, we might expect that teachers who identify as Republicans or political conservatives might be reluctant to both join their unions and participate actively in politics through their unions. At the same time, I hypothesize that not all unions will face equal difficulty in reaching out to cross-pressured members. State and federal law creates one powerful incentive for teachers, even conservative ones, to join and participate in the organization (Feigenbaum, Hertel-Fernandez, and Williamson 2019; Hartney 2022). Important as external factors are for spurring member activism, in this paper I consider internal characteristics of organizations that might prompt cross-pressured teachers to join and participate in union-led political activities.

More specifically, I focus on culture within local unions. A large body of research has underscored that leaders can shape the culture of organizations—the norms, principles, and values that constitute a shared identity and establish rules about how members ought to behave. In turn, culture can play an important role in determining how organizations operate, including how members construct and order their preferences and choose to act (Kreps 1990; Schein 2017). In the context of the labor movement, scholars have documented that leaders play a crucial role in fostering particular cultures (Stepan-Norris and Zeitlin 2002). But for union cultures to persist over time, there must be mechanisms of communicating their underlying principles and values from leaders to members and that enforce those principles on members and leaders alike (see especially Ahlquist and Levi 2013). These mechanisms bear important similarities to the norms that Ismail White and Chryl Laird (2020) document in predominantly Black social institutions that align conservative Black Americans with the Democratic party. These mechanisms of internal culture also can be thought of as changing the calculus faced by individual members to organizational engagement by shifting the solidary benefits (e.g., social rewards or pressures) they face using the typology of organizational involvements produced by Clark and Wilson (1961).

Given this research, I investigate whether different organizational cultures can explain why politically cross-pressured members might be more likely to participate in union-led political activities. I hypothesize that local leaders have the ability to foster union cultures that encourage participation by their members in politics. I anticipate they can do so by setting a norm of political engagement and participation, including communicating to members why and how political participation is aligned with members’ and local unions’ priorities—even if members might hold broader political views at odds with the union as a whole, especially the national union. In this way, unions can take advantage of their federated structure, with local unions affiliated with state and national associations but able to adapt their messages and practices to local contexts. I hypothesize that local unions that are able to tailor requests around politics not to hot-button national issues or debates but rather to local needs and issues will be especially successful in engaging cross-pressured members.

One challenge in exploring the effect of local union culture on member behavior is that culture and norms are difficult to directly observe and measure systematically across different unions. In this paper, I take two complementary approaches. In the quantitative analysis, I capture observable practices for communicating, institutionalizing, and modeling norms of participation over time. In the qualitative analysis, I use interviews with local union leaders to understand in more detail the mechanisms of generating and transmitting norms through these different practices, as well as broader practices that may be more difficult to systematically measure across locals.

The first mechanism I study for establishing norms around participation occurs when educators first have the opportunity to join the union, through new member education and orientation. This is a critical time for establishing members’ expectations about what the union stands for and their roles and responsibilities within the union (see, e.g., LiUNA N.d.). An older generation of research backs up the importance of early socialization experiences into unions, stressing that both formal and informal socialization can help build members’ sense of commitment to, and identification with, the union (see especially Clark et al. 1993; Gallagher 1989). In light of this research on member orientation and socialization, I focus especially closely on the presence of new member orientations within local unions—and how these orientations complement later communication between the union and members about politics and political issues, including regular publications like newsletters (see also Newman and Skocpol 2023 on the importance of newsletters for local union culture). In Supplementary Material S6, I include an excerpt from one local newsletter that demonstrates how newsletters can connect political calls to action to local-specific concerns and place those calls alongside calls for charitable donations to support victims of a recent local storm as well as social activities. These excerpts all underscore the importance of participating in local union activities to advance a sense of community and common set of values.³

Hypothesis 1: Local unions that conduct new member orientations, and especially new member orientations that establish a norm of political participation and member engagement, will be more likely to foster political participation among conservative members.

Hypothesis 2: Local unions that communicate more regularly with members, especially in ways that establish a norm of political participation and member engagement, will be more likely to foster political participation among conservative members.

Last, I consider whether leaders themselves model the behaviors that they are trying to encourage across their members. As other scholars have emphasized, organizations where leaders themselves model commitments to group norms are more likely to see their members internalize and follow those norms themselves (Ahlquist and Levi 2013; Ganz 2009). In the context of political participation, I anticipate that locals where leaders are more involved politically themselves, especially in public ways, will be more likely to inspire political participation among their cross-pressured members.

Hypothesis 3: Local unions where leaders are more active in politics will be more likely to foster political participation among conservative members.

To be clear, I do not expect that orientations, newsletters, and leadership modeling are the only means of establishing and propagating local union norms around political participation. Instead, I view these as proxies for practices and activities that unions may deploy to build strong norms of political participation. In the qualitative section of the paper, I explore these broader practices in more detail.

THE CONTEXT: IOWA TEACHER UNIONS

This study draws on an ongoing collaboration with the Iowa State Education Association (ISEA), the state’s teacher union affiliated with the NEA. ISEA’s membership is mainly elementary and secondary school teachers but also includes some school secretaries, paraprofessionals, custodians, and instructors at postsecondary institutions, like community colleges. The association represents around 30,000 members out of a potential membership pool of around 50,000 employees across slightly over 400 local unions. These local unions are the main unit in my analysis. Like in other states, teachers account for one of the largest and most politically active unions in Iowa: the ISEA’s political action committee (PAC) is the second largest non-party contributor to state elections in recent cycles. And, like other states, the ISEA is closely tied to the Democratic party. From 1998 to 2022, fully 97% of its PAC contributions flowed to Democratic candidates, according to OpenSecrets.

The ISEA makes for an informative case study for studying the political mobilization of cross-pressured workers in unions. By focusing on a single statewide union of one type of workers, I can hold constant variation in the overall statewide political climate (including laws and institutional structure) as well as the reasons that workers might have to join, allowing me to isolate the effects of variation in local associational structures. Most importantly, there are good reasons to think that Iowa represents an especially challenging environment for unions to engage cross-pressured members in politics. For one thing, the state tilts conservative, and in the 2016 presidential election, Donald Trump won the state by over five percentage points and in 2020 by over eight points. For another, Iowa’s laws are unfriendly to public sector unions. Cross-pressured members wanting union benefits thus have weaker reasons to support the union, let alone participate in its political activities. Understanding how local unions in this challenging environment can engage conservative or Republican members thus carries important lessons for other contexts, especially as unions deal with the post-Janus landscape. Indeed, the state has long been right-to-work, and as interviews with union leaders revealed, that has meant that the union must come up with strategies to appeal broadly to more teachers, including conservative teachers. In turn, this makes Iowa a good case to understand what those strategies might be as compared to states that more recently transitioned to right-to-work after Janus that might not have such well-developed tactics for reaching conservative workers.

According to representative surveys of ISEA members and non-members I describe below, there is important variation in the ideological and partisan composition of their membership. While well over half (57%) of members identify as politically liberal, 25% report being politically conservative (including 14% who identify as somewhat conservative, 9% who identify as conservative, and 1% who identify as very conservative). According to the same survey, a slightly lower percentage of members identify as Republicans (21%). About two-thirds of members identify as Democrats (65%) and the remaining members as Independents (14%). A 2023 survey of educational professionals and staff who were eligible to join the union but who were not members showed that these non-members were less likely to identify as Democrats than were union members (see below for more details). Of these non-members, about 42% were Democrats, 23% were Independents, and 34% were Republicans.

DATA AND METHODS

To test the hypotheses developed above, I draw on a mixed-methods design (similar to, but distinct from, Lieberman 2005). I first tested the role of local union orientations, communications, and leader political activism—signals of broader norms of participation—using surveys of ISEA members and leaders in a large-N statistical analysis and then shifted to a model-testing set of small-N paired interviews, selected to be similar except for the presence of local union political culture.

An original survey of ISEA members, fielded online in May 2018, received 1,904 responses for a response rate of 7%. The survey asked questions related to members’ perceptions of, and involvement in, the union, as well as member demographics. Importantly, participating members were broadly representative of the overall union membership along many important demographic characteristics, including political interest and attitudes about the union (see Supplementary Material S1).

The ISEA local presidents’ survey, fielded online in January 2019, received 154 responses out of 436 for a response rate of 35% (see Supplementary Material S1 for balance checks). The leader survey asked a variety of questions about practices and culture within each association, focusing on the president of each local union. Presidents are the key actors of interest in understanding how local unions are structured and operated. I complemented the matched member–leader survey data with internal administrative records from the ISEA, including on member participation in the ISEA PAC in 2018, which I will use as one outcome. (Survey instruments are in replication materials; see Hertel-Fernandez 2024.)

Last, to further test my hypotheses, especially in the post-COVID-19 context, I partnered with the NEA to access a representative survey of educational professionals and staff they fielded in Iowa in February 2023. The survey was emailed to 45,961 ISEA members and eligible non-members; 2,891 respondents completed the survey for a response rate of 6% (this includes 1,439 current union members). Respondents were very similar to non-respondents on available demographic characteristics; nevertheless, I applied survey weights produced by the NEA to match targets on gender, age, and partisanship.

While the survey evidence can point to the broad relationships that exist between local practices that may signal norms of participation and cross-pressured member political involvements, they cannot reveal the specific mechanisms that may underpin those relationships or the broader practices that local unions may be using to establish and propagate norms of participation that I cannot measure systematically in the surveys. To further probe the connection between local culture and cross-pressured member political participation, I conducted semi-structured, qualitative interviews with eight leaders from local associations chosen because they were similar except for the presence or absence of participatory practices, like new member orientations and regular newsletters (see Supplementary Material S7 for more details).

TESTING LOCAL UNION CULTURE IN THE 2018–19 SURVEY AND ADMINISTRATIVE DATA

Using the matched 2018–19 survey data, the first outcome I consider is an indicator of whether members participated in the ISEA statewide PAC in 2018, an important commitment that helps the union engage in electoral politics. This measure is meaningful because in Iowa members must affirmatively contribute a portion of their wages to the PAC (i.e., there is not a reverse check-off; see Supplementary Material S5 for the form). Like other unions, the ISEA’s statewide PAC makes contributions on the basis of recommendations from a committee of union members. Those members are appointed by local unions across 19 regions in the state, and membership is balanced between registered Republicans and Democrats. Candidates receive ISEA PAC support based on either their legislative voting records or interviews and questionnaires administered by the PAC.

Individual PAC contributions are an outcome that is substantially and theoretically important. Past research has shown that an important element of teachers’ unions political strength are PAC war chests—especially because they signal to politicians that unions have a large, engaged grassroots base (Hartney 2022; see also DiSalvo 2015; Moe 2011). In addition, a methodological advantage to the PAC contribution measure is that it is readily comparable across all members and validated by ISEA internal data (i.e., not self-reported). During the period of this study, 78% of members contributed to the PAC, but there was significant variation by the ideology and partisanship of members. Of the sample of respondents to the 2018 survey, 83% of self-identified liberals contributed to the PAC, while only 71% of conservatives did. Similarly, 72% of self-identified Republicans contributed to the PAC compared to 82% of Democrats.

The key explanatory variables in the 2018–19 analysis come from the leader survey. To test the first hypothesis, the leader survey asked whether local unions conducted regular orientations. Of the local leaders responding to my survey, 31% reported that they conducted new member orientations “regularly,” 32% reported that they conducted new member orientations “sometimes,” and 37% reported that they either did not conduct such orientations or were not sure if their local did so. I then created a binary variable for regular orientations.

To test the second hypothesis about regular communication from the local union to members, the leader survey asked presidents whether their local unions communicated to their members through regular newsletters. These were less common than orientations, with only 17% of local leaders reporting that they had regular communication with their members in this way.

The third hypothesis was whether locals with more politically engaged leaders were more likely to encourage participation among their members, especially conservative members. To test this item, I rely on an item from the leader survey that asked presidents whether they did any of the following things in politics in the past two years: trying to persuade others to vote for candidates or parties, participating in political meetings, putting up campaign signs or posters, volunteering for political campaigns, donating to political candidates or parties, or contacting elected officials. I then created an additive index ranging from zero to six.

It is important to acknowledge that all three measures are reported by local union presidents. To verify reports were accurate, I shared aggregated reports with statewide union leaders responsible for supporting different regions, who confirmed that the distribution of new member orientations and newsletters qualitatively fit their understanding of where these practices were held. Still, I could not verify each individual response.

I estimate ordinary least squares (OLS) regressions with political participation or member views as outcomes, and as the main explanatory variables the presence of different measures of local union political culture, member ideology, and their interaction. Equation 1 details the specification. $ {Y}_{il} $ is the outcome of interest, with i indexing individuals and l indexing union locals. Culture is the measure of local union participatory culture: regular orientations, regular newsletters, or local leader activism. Ideology is a set of dummy variables for member political ideology, including indicators of liberal, moderate, and conservative. To estimate the effect of culture for conservative members, I include an interaction between the measure of local union culture and member ideology dummies, with the liberal dummy being the excluded category. This specification permits me to estimate how member political views or participation varies for conservative members in union locals with and without various indicators of culture. In replication materials additional results 1, I show the same specifications using member partisanship, rather than ideology, and find similar results. Across some models, I also include a set of individual demographic and political controls (D) as well as a set of local union-level controls (U).

(1) $$ {Y}_{il}={\displaystyle \begin{array}{l}{\beta}_0+{\beta}_1{Culture}_l+{\beta}_2{Ideology}_i+{\beta}_3{Culture}_l\\ {}\times {Ideology}_i+{\beta}_4{D}_i+{\beta}_5{U}_l+{\varepsilon}_{il}\end{array}} $$

At the individual member level, I control for gender, race (white/non-white), age (in quartile dummies), length of membership (in quartile dummies), self-reported job satisfaction (on a one-through-five scale), and dummies for members’ occupation (K-12 teacher, non-teacher school staff, postsecondary instructors or staff, or others). Crucially, I control for members’ recent ISEA PAC giving in 2015–16 and 2016–17, representing two years of lagged outcomes.

At the local union level, I control for the size of the local’s bargaining unit (in individuals), the leader’s political ideology on a seven-point scale (ranging from very liberal to very conservative), and the average ideology of the overall local union membership (again on a seven-point scale). These variables help address the concern that the overall ideology of a local union’s membership or leadership is driving my results. Importantly, I account for the different reasons that members might have for joining their local union—attempting to address selection effects of conservatives into particular unions—by drawing on a leader survey item that asked presidents which arguments they used in recruiting new members, including the following appeals: professional development, joining a community of professionals, legal protections, voice in school, voice in politics, financial benefits, or pride and solidarity. This variable helps address the concern that conservative members might select into locals for different reasons that might shape their political participation aside from local union culture. In addition, I control for the broader political climate by including the share of votes received by 2018 House Democratic candidates at the county level. I also control for a key reason that locals might have a well-developed orientation: their relationship to their school administration (as reported by leaders on a 1–5 scale); more favorable relationships mean that the union is more likely to be able to use school property and/or time for an official orientation for new hires and members. Supplementary Material S3 summarizes these variables.⁴ Across all regression models, I cluster standard errors by union local.

Figure 1 summarizes the key quantity of interest across the different regression specifications: the predicted difference between conservative ISEA members’ PAC contribution rates in locals with and without various indicators of local union culture. For example, for regressions studying regular orientations, this quantity shows the difference in the predicted PAC contribution rate for conservative union members in locals with and without regular orientations.⁵ Positive values indicate that conservative members are more likely to contribute to the PAC in the presence of a stronger local union culture; negative values indicate the opposite.

Figure 1. Difference in ISEA PAC Contributions for Conservative Members, by Local Union Culture

Note: The figure plots the predicted difference in PAC contribution rates for conservative ISEA members in locals with and without strong local union political culture (i.e., with and without regular orientations, regular newsletters, or more active local leaders). Positive values indicate that conservative members are more likely to contribute to the PAC in the presence of a stronger local union culture. 95% confidence intervals are shown. For each measure of union culture, I show regression specifications with no controls, individual controls, local controls, and individual and local controls. Replication materials additional results 1 document full regression results.

For each measure of local union culture, I estimate four regressions, with varying individual or local controls (see replication materials additional results 1 for full regression results). The top left panel tests hypothesis 1, regarding the role of new member orientations. Supporting hypothesis 1, I find that conservative members in locals with regular new orientations are more likely to contribute to the ISEA PAC than their counterparts in locals without such orientations.⁶ Across the various specifications, I find that conservative members in locals with regular orientations are about 20 percentage points more likely to contribute to the ISEA PAC than their conservative counterparts in locals without regular orientations. These are meaningful differences; indeed, the liberal-conservative difference in PAC contributions virtually disappears in locals with regular new member orientations.

In the top right panel, I test hypothesis 2, regarding the presence of regular newsletters.⁷ Unlike hypothesis 1, I find less consistent evidence that conservative members contribute to the ISEA PAC at higher rates in locals with regular newsletters. In some models, I find that conservative members are about 16 percentage points more likely to participate in the PAC in locals with regular newsletters compared to locals without regular newsletters, but the estimates are unstable across different specifications.

Last, the bottom left panel tests hypothesis 3, regarding local leaders’ own political activism.⁸ I find noisier evidence that conservative members contribute to the ISEA PAC at higher rates when local leaders are themselves more politically active. For these models, I estimate the difference between a local leader reporting no political acts and the maximum number of political acts (6). In these models, conservative members are about 20–40 percentage points more likely to participate in the PAC in locals where their local leaders are more active, moving from the least to the most active leader. Together, Figure 1 offers strong support for the role of local union culture in conservative members’ willingness to participate in politics, especially as measured through regular orientations and local leaders’ own activism.

In Figure 2, I test another set of outcomes. Instead of examining PAC contributions, I use a survey item that asked members to indicate on a five-point scale “How well does the ISEA represent your interests in the following areas?” Those areas included “state spending on education” and “other state policies aside from education.” I anticipate a stronger union culture toward participation should help conservatives feel better represented by their union in politics—for both education- and non-education-related issues.

Figure 2. Difference in Perceptions of Union Political Representation for Conservative Members, by Local Union Culture

Note: The figure plots the predicted difference in conservative ISEA members’ views of how well the union represents them in politics in locals with and without regular orientations; the left plot examines views of union representation in politics for education spending and the right plot shows views of union representation on non-educational political issues. Positive values indicate that conservative members are more likely to feel represented in politics in the presence of regular member orientations, especially for non-education-spending-related issues. For each outcome, I show regression specifications with no controls, individual controls, local controls, and individual and local controls. 95% confidence intervals are shown. Replication materials additional results 2 document full regression results.

Figure 2 reports results from OLS regressions with member perceptions of representation in politics as the outcome and an interaction between local union culture and political ideology, plus varying individual- and local-level controls (see replication materials additional results 2 for full regression output). I focus on new member orientations given the consistency of findings from Figure 1. The left panel of Figure 2 shows that across specifications, conservative members feel better represented by ISEA’s advocacy on education spending in locals with regular orientations, by about 0.20 units on the 1–5 scale, but these differences are not significant at conventional levels of significance. In the right panel of Figure 2, I find stronger and more consistent results. Across specifications, I find that conservative members feel better represented by the ISEA on non-education issues in locals with regular orientations, by an average of about 0.50 units on the 1–5 scale.⁹ Figure 2 thus indicates that conservative members feel better represented by their union in locals with strong participatory cultures, particularly on non-educational issues that might be more likely to be subject to cross-pressures from conservative members’ political identities.

UNION CULTURE AND MEMBER POLITICAL ENGAGEMENT IN THE 2023 NEA SURVEY

In addition to the 2018–19 matched member and leader surveys, the February 2023 NEA survey permits me to examine the relationship between union practices and members’ attitudes and preferences toward union political activities. It also provides an opportunity to replicate my initial survey findings in a post-COVID-19 setting. Because I did not design this survey, there are not the same items on this survey as on my original member and leader surveys. There is, however, one item that captures the regularity of communication between unions and members that can foster norms of accountability, participation, and representation.

The question asks: “How recently has someone from your union or association asked you about what you want the union to do for its members?” I anticipate that this question will help capture the degree to which local leaders are regularly communicating with educators and creating norms of representing educators’ views in union activities and priorities (see especially Newman and Skocpol 2023, chapter 4). This question has the important additional virtue of permitting me to test members’ own perceptions of local union culture and norms, asking about members’ experiences directly rather than relying on leaders’ reports of relevant activities (such as newsletters or orientations). Responses include “in the last month,” “during this school year, but not in the last month,” “in the last two years, but not this year,” “longer than two years ago,” and “never.” I dichotomize this variable at the median response, grouping together “never” and “longer than two years ago” into one category and the remaining responses into the other.

I examine five outcomes related to individual union members’ willingness to engage in various political activities on behalf of the union as well as their perceptions of union political mobilization and the union as a whole. These measures complement and expand on the earlier analysis by examining a broader range of union members’ political activities and attitudes.

Willingness to engage in union political actions (Willingness to Take Anti-Voucher Legislation Action): First, I examine how willing members are to respond to a union call to political action. The survey described recent legislation passed by the Iowa legislature expanding the use of school vouchers and asked the following question: “If ISEA were to ask educators to take action to oppose this new law, how likely would you be to take each of the following actions?” Respondents were shown seven possible actions—writing a state legislator; encouraging friends, family, or neighbors to take action; attending a rally; encouraging colleagues to take action; signing petitions; campaigning for anti-voucher candidates; and donating to the ISEA PAC—and for each action could rate their likelihood on a 1–4 scale, ranging from not at all likely to very likely. An advantage of this outcome is that permits me to look at a range of political activities and, in addition, conservative members’ willingness to engage in advocacy against a traditionally conservative cause (i.e., expanding school vouchers). I average respondents’ likelihood across all seven actions. The average response was 2.9.

Perception of union legislative mobilization (How Well Does ISEA Represent You in State Legislature): Second, I examine union members’ assessments of how well they feel the union represents public education in the Iowa legislature with a question that asks how well, on a 1–4 scale ranging from not well at all to very well, members feel ISEA acts as “a powerful voice for public education in the state legislature.” The average assessment was 2.8.

Prioritization of union political mobilization (How Much Should ISEA Prioritize Politics): Third, I examine whether union members think the union should engage in political activities to shape state policy, an item that taps into the norms around union activities. The item asked members to rate how much of a priority a variety of activities should be for the ISEA; I use the following item: “Provide a voice for members’ interests in the state government.” Members could respond on a 1–4 scale, with higher values indicating a higher priority. The average response was 3.0.

Union involvement in politics as a reason to continue membership (Importance of Politics to Your Membership): Fourth, I capture an additional measure of activities that members tie to their membership in the union, asking members how convincing different descriptions of the ISEA are to continuing respondents’ membership on a 1–4 scale ranging from not at all convincing to very convincing. This item, like the previous one, helps tap into the norms around union membership related to political participation. I use the following item: “Politicians and non-educators make too many of the decisions about education. Through our union, we get a seat at the table to have input on the decisions that affect our jobs, our schools, and our classrooms.” The average response was 3.0.

General perception of the ISEA (General ISEA Performance): Last, I examine respondents’ general assessment of their representation by the ISEA with the following item: “In general, how would you rate the job Iowa State Education Association is doing for its membership?”, to which respondents could respond on a 1–4 scale ranging from “poor” to “excellent.” The average assessment was 2.9.

I estimate regressions as detailed in Equation 2:

(2) $$ {Y}_{il}={\displaystyle \begin{array}{l}{\beta}_0+{\beta}_1{Culture}_i+{\beta}_2{Ideology}_i+{\beta}_3{Culture}_i\\ {}\times {Ideology}_i+{\beta}_4{D}_i+{\beta}_5{LU}_l+{\varepsilon}_{il}\end{array}} $$

As before, $ {Y}_{il} $ is the outcome of interest, with i indexing individuals and l indexing union locals. Culture is the measure of local union participatory culture and norms measured at the individual level (how frequently members talk about their priorities for the union with union leaders), Ideology is a set of dummy variables for member political ideology, and I include an interaction between the measure of local union culture and member ideology dummies. As before, the liberal dummy is the excluded category. I include the following individual-level controls (D) in some models: age (six bins), gender, occupation (six bins), and length of tenure in their occupation (four bins). Supplementary Material S4 summarizes these variables. Across all regression models, I cluster standard errors by union local and apply NEA’s survey weights. Because the measure of local union culture in this analysis is measured at the individual level, I also include in some specifications local-specific fixed effects (LU), which help account for local-specific confounders. To more easily compare results, I standardized all outcomes to range from 0 to 1.

Replication materials additional results 3 document full regression output, and I summarize the main substantive results in Figure 3. As before, I plot the key quantity of interest: the difference in outcomes for conservative members reporting stronger versus weaker local union culture. Positive values indicate that conservative members indicate a stronger attachment to the union and more support for political activities in locals with a stronger local union culture, compared to their conservative counterparts in locals with a weaker union culture. Across all five outcomes, I find that conservative members who described stronger local union culture reported (1) being more likely to take political actions to oppose a Republican priority in the state legislature to expand school vouchers;¹⁰ (2) feeling that the ISEA represents them well in the state legislature;¹¹ (3) feeling that the ISEA should prioritize political activities;¹² (4) feeling that the ISEA’s political activities were an important reason members should stay in the union;¹³ and (5) that the ISEA overall was doing a good job representing them.¹⁴

Figure 3. Difference in Union Attitudes and Actions for Conservative Members, by Local Union Culture

Note: The figure plots the predicted difference in conservative ISEA members’ views of their union in locals with and without a strong union culture (i.e., more frequent conversations between union members and leaders about their priorities for the union). Positive values indicate that conservative members are more likely to feel more favorable toward the union or to take political action in the presence of a stronger local union culture. For each outcome, I show regression specifications with no controls and with individual controls and local union fixed effects. 95% confidence intervals are shown. Replication materials additional results 3 document full regression results. Outcomes standardized to run from 0 to 1.

These are meaningful differences. Consider two important outcomes. Using the estimates from the model with controls and local fixed effects, I find conservative members describing a stronger union culture reported a likelihood of taking actions to oppose voucher legislation 38% higher than conservative members describing a weaker union culture (p<0.001). In a similar vein, in the model with controls and local fixed effects, conservative members reporting a strong union culture gave an 84% higher rating to politics as a reason for membership compared to conservative members reporting a weaker union culture (p<0.001).

ADDRESSING SELECTION CONCERNS

One concern with my results is that as a long-time right-to-work state, Iowan educators can voluntarily join or leave the union at any time—and therefore conservative educators who join the union are fundamentally different from non-members. For this to be a concern, we would need to think that there were factors apart from unions’ participatory cultures that attracted conservatives into particular local unions and not others—net of the factors I control for above. The use of extensive individual and local-level controls, especially lagged PAC contributions and fixed effects for local unions in some specifications, helps assuage concerns that there are other factors than participatory culture driving my results within particular unions and accounting for individuals’ past decisions about political engagement.

Another strategy to address selection concerns is to leverage the fact that the 2023 NEA survey interviewed both members and non-members, including members who canceled their memberships. Examining these departed members’ perceptions of the union and willingness to support political action, I find that they are unsurprisingly much more negative about the ISEA’s performance and have much lower support for union political activities. But as I document in replication materials additional results 3, conservative educators who canceled their membership who still reported regular, participatory communications with the union reported stronger perceptions of the union and support for union political action in the state legislature than educators who did not. Ultimately, no single analysis can definitely address concerns about selection, but together the different analytical choices can help assuage this concern. Nevertheless, there may still be unobserved factors correlated with participatory culture that may explain these findings, and this is an important qualification to my findings.

MECHANISMS FOR UNION CULTURE AND POLITICAL ENGAGEMENT IN LEADER INTERVIEWS

Complementing the quantitative analysis, I drew on interviews with eight local leaders, chosen to represent locals similar except for their union culture, so as to examine the mechanisms connecting local union culture to conservative member political engagement. This represents a model-testing small-N analysis (similar to, but distinct from, Lieberman 2005). Table 1 summarizes the most important characteristics of the local presidents who participated in the interviews by each of the pairs interviewed.

Table 1. Locals Interviewed

These interviews help contextualize the quantitative results in several ways. First, they show how high-participatory-culture locals use a range of strategies to build norms around political engagement for their members, including orientations, newsletters, and modeling behavior themselves, but also many other strategies that go beyond these as well. Second, the interviews reveal how leaders representing similarly conservative areas took different approaches to dealing with the challenge of engaging their conservative members. While some leaders steered clear of politics, others developed strategies for building norms of participation from the start of members’ exposure to the union. The interviews thus underscore the importance of leaders’ own decisions to establish specific practices to foster a more participatory culture. And last, the interviews provide suggestive evidence for one important mechanism of local union culture around political participation: how union leaders denationalize discussions of politics to focus on locally relevant issues, rather than hot-button national issues like abortion or support for then President Donald Trump. In this way, the federated structure of the ISEA allowed locals to tailor their appeals and discussions of politics in ways sensitive and responsive to their membership.

I first summarize how leaders described participatory culture and norms in their local unions, and then describe how choices around participatory culture mattered for member political engagement.

How local leaders build participatory culture: I first consider concrete examples of what locals are doing in high participatory locals, like local H, which I contrast with local G. Even before I asked local H’s leader about their new member practices, as part of the background the leader was providing they emphasized that they “do a lot of things for our new members especially” to help them understand what the local is all about and why it is important not only to join the association but to stay involved throughout the year in political and non-political activities alike.

The membership process starts when the union gets the list of “names of all the new teachers” and then they “write letters [to each of them],” using it as an opportunity to “welcome them to the district” and encouraging them to tell the union “if they need any help or have any questions.” Importantly, the idea behind this personal outreach is not to push hard on new hires to join the union, but rather to welcome them and to get them to start thinking about the union as a potential resource—and community with distinct norms. Next, the union holds a breakfast with the superintendent for the new hires. That’s another chance for the union to “introduce ourselves, our role, and why we we’d like to have you as members.”

The breakfast with the superintendent is only the first step of new member recruitment and orientation, and the week after the breakfast, the union organizes a separate lunch that is free for any new teacher and subsidized for existing teachers as a means of encouraging high participation from the membership. In the leader’s experience, they tend to get good attendance from their existing members, who understand how important it is to show new hires just how much existing members are involved in the local and take participation seriously.

The discussion at the lunch focuses on the importance of being a member—and especially an active member—in the union and all the different ways that the union can help teachers. The union invites their representative from UniServ (union staff assigned to work with local unions) to attend the lunch to explain what the state and national association do to help locals through their unit representative, including help with grievances, contract negotiation, and politics. Notably, the union also invites state representatives “that are more education friendly” to the lunch to help teachers understand why it is so important for the union to work with legislators to advance education priorities in the state legislature. That has the beneficial consequence of both deepening ties to the state legislature and also helping members to understand how politics and the legislative process are directly relevant to their practice as educators on local, educational issues—and not just on hot-button and nationalized social issues.

Apart from new member meetings, the local holds monthly meetings and communicates with members online through a regular newsletter and a Facebook page. The local also organizes periodic social events. Both the social events and monthly meetings tend to be well attended, and the local’s leader recognized that there was a big difference between their local’s culture and other locals they and their members had worked in previously. “We’ve heard from other people who have moved here from other districts and we hear that we do a lot more [to engage members].” “We do a lot more because we need to,” the local’s leader summed up, emphasizing that they felt it was essential to maintain a high participation culture for their success in negotiating strong contracts and in ensuring a positive environment for education overall in the state.

Local D also had a high participatory culture, though not as high as local H (I contrast local D with local C). Like high-participatory local H, local D hosts regular new member meetings at the start of the school year. There is a lunch for new hires and existing members to convey the importance of joining the union, like in local H, as well as a subsequent “in service day” in which the union has negotiated time during the school day to meet with new hires, get to know them on a one-to-one basis, and further convey what the union can do for them. These new member recruitment and orientation events, like other regular meetings organized by the union, tend to be well attended, and the local’s leader is proud that this is a self-reinforcing cycle: because union membership and participation is so high, “if you’re not in the club [or active in the union], you’re on the outskirts…you feel left out,” evidence of the importance of social norms and pressure that change the solidary calculus members faced. As evidence of the high level of participation, not just membership, the leader stressed that they usually do not have problems getting members to serve on committees or in leadership positions on the union’s executive board: “everyone knows you should serve some term,” they emphasized, explaining that they “usually can find someone to step up.”

I contrast the high participatory culture in locals D and H with locals C and G. Local C demonstrated the challenges of communicating with new members, while local G demonstrates issues with communicating with existing members and convincing them to participate in the union. Unlike local D, local C does not have formal new member orientations, instead relying on informal interactions between the leader and potential new members. These conversations can be helpful for encouraging new membership, but do not tend to cover broader participatory activities in the union and, because they are ad hoc, do not do much to instill new norms. In the end, the leader connected this absence of deeper engagement with the challenges they face getting members to serve in union activities. That leader lamented that they “don’t have that many [people, especially] younger people to serve [in union activities]…although I invite them, it’s kind of the same people who are doing the same work [every year].”

Local C’s leader noted that they still retain a very high membership rate despite not having much of a new member orientation or participatory culture. But the high membership rate in the local is mostly a result of low turnover in the teacher population as a whole, not recruitment of new members. When asked about how the association engages its members, the leader acknowledged that it was “definitely an area where we need to try to improve.” It was telling, for instance, that at one point the leader realized they were receiving regular communications from the NEA and ISEA about happenings in state politics that “the rest of the members were not getting, so that was a surprise to me.” The leader is trying to do more to ensure those lines of communication are working—but there were no specific plans for new activities at the time of our interview.

How participatory culture matters for political engagement: Having established what participatory culture—and its absence—looks like across the matched pairs, we can now turn to exploring how these locals approached political engagement of their members, both in elections and in legislative debates, and how local union cultures may have constrained or enabled mobilization of members, especially conservative or Republican members.

The description of local H’s recruitment process and new member orientation already emphasized just how much politics entered into the local’s identity and self-presentation to potential new members from the start of those members’ relationship with the union. That made it easier for leaders to make the subsequent case to members about why they should get involved in elections and legislative debates. When there were elections, the leader emphasized, members “went door-knocking, some of us did parades, that kind of thing.” It helps as well that the leader has been involved in their community’s local Democratic party committee and thus could identify volunteer opportunities for their members and also introduce members to political candidates around elections.

Local H also made use of their trusted regular channels of communication with members, like their newsletter, social media, and emails, to share information about opportunities for staying abreast of legislative debates and key moments when members should weigh in with their elected officials. As the leader described to me, “we try to contact our legislators, cause we’re trying to get everyone to do that, we have a newsletter, an email that comes out that keeps them [updated], that tells them the bills are coming up, which ones we need to contact our legislators on; we tell them where the forums are.” In local H, as elsewhere in Iowa, it is often difficult to get conservative individuals who are committed to Republican positions on a number of issues to support the union and participate in politics. Nevertheless, the leader reported that they still managed to attract a number of Republicans and conservatives as members—and even get them participating in political activities—by emphasizing how important it is to support pro-education policies and candidates and making connections between what happened to teachers in the classroom and local and state policy decisions.

That strategy was also echoed by local B (matched with local A), whose leader underscored how, when their local made endorsements of political candidates or legislation, they try to “get [teachers] to vote their job, that’s the key phrase, and to pick candidates [and legislation] that are going to support public schools and students.” The process of developing that message started early on in members’ careers. Every year during their regular orientation, the local’s president met with members “one-on-one” and mentioned the importance of participating in politics, including contributing to the PAC, to improve schools and education in the state. That leader also emphasized to members how rigorous the endorsement process was and how it was not based on party. Instead, it is a thorough vetting where an elected committee of members, including Republicans, looks into the background of each potential candidate and how they align with the union’s positions on schools and education. “Knowing where their money goes, and there is a rigorous process that goes into selection not just on political affiliation really helps,” local B’s leader explained.

Local D’s leader (matched with local C) similarly began our conversation about member participation in politics by stressing that such recruitment is hard in Iowa, given that they represent a population that can be “pretty conservative, and that is not where ISEA leans with its candidates [that the ISEA endorses].” (Notably local D was the most conservative of the eight locals according to the 2018 survey.) As a result, the leader explained, “we run into a lot of problems when we want to talk about political things because so many of our members are conservative.” But although “it’s hard,” the local still manages to share information about politics and ISEA-endorsed candidates and get its members participating regularly, sharing and promoting many opportunities to do so from its UniServ representatives. One especially interesting example involves townhall meetings that the local attends with their state legislators. The local’s leader will encourage member participation alongside representatives from their school board and school administration. Local D contrasts with the experience in its pair from local C, which has a similarly conservative population but has a difficult time engaging its members in politics and by its own leader’s account does not do much to reach out to new or existing members through orientations or regular communications.

Moving to a local with a very low participatory culture—local A, matched with local B—we see an even greater reluctance to engage members in politics because of the fear of backlash from their conservative members. As the local leader explained, in 2018, the ISEA “wanted us to [do more political engagement around the midterm elections],” but they ended up “not doing much, in part because of recertification [of collective bargaining status]” but also because of the “strong Republican contingent” in their membership. As the leader explained, “that’s partly why I didn’t push the election stuff, and I didn’t want to turn off any of our Republican members…who might feel that we’re going against them.” Local A does not do much in legislative debates, either, apart from sending the occasional member to help lobby in the statehouse once a year.

The leader of another very-low-participatory-culture local—local C, matched with local D—relayed similar concerns about the partisan division of their members. Although the local could get good member participation at the school board level—they boasted of high mobilization and turnout in a recent race—the leader explained that “It’s a little touchier when it comes to [politics at the] state level…and the big picture at the national level…due to the political environment today.” As a result, there are not many efforts in their local to engage members in politics. But the leader is hopeful that it may be changing as more and more conservative members of the union and the broader community see the consequences of Republican state control for teachers and schools.

UNIONS AND CIVIC ENGAGEMENT IN AN ERA OF POLARIZATION

Together, the quantitative and qualitative evidence are consistent with the idea that civic organizations vary in their capacity to cross partisan and ideological lines and motivate the participation of individuals who might disagree with some of an organization’s political positions. Looking across local teacher unions in Iowa, I found that locals that fostered a greater participatory culture were able to encourage a stronger sense of political representation and activity among cross-pressured conservative and Republican teachers. I find that conservative teachers are less likely to support union-led political activities, to feel represented by the union in politics, and to feel that their union should be involved in politics as compared to liberal educators. But conservatives and Republicans within local unions that fostered stronger participatory norms, including through regular member orientations and discussions with their members about union priorities, were more likely to be involved in politics, to feel represented by the union in politics, and to feel that their union should be involved in politics compared to their conservative and Republican counterparts in locals without such norms.

There are several important limitations to my analysis. First, while this study points to the role of local union culture in member political participation, it is less focused on the adoption of union culture or changes in local culture. This is a valuable question for future research, though my interviews and past research on local union revitalization suggest several starting points—for instance, crises or scandals, transitions in leadership, and supportive state and national federations with relevant expertise and resources (Voss and Sherman 2000).

Relatedly, this paper has focused on a subset of strategies unions can deploy to appeal to cross-pressured members by fostering norms of participation, such as regular member orientations and conversations with rank-and-file members. But these strategies are ultimately means of establishing a participatory culture within local unions, not ends in themselves. There may well be other ways for local unions to establish the same norms through different organizational mechanisms. While the interviews pointed to a number of other practices, such as holding regular townhalls with elected officials, making locally rooted appeals, and stressing the internally democratic and accountable nature of unions, more work is needed to identify these mechanisms.

Another open question is how my findings would travel to other contexts, including other unions, civic associations, or even business settings (such as corporate PACs; see especially Li 2018). Given the makeup of the ISEA and the state’s laws at the time of the study, I anticipate that the results will be most relevant for other unions whose affiliates are strongly associated with one political party or set of ideological positions that conflict with the views of local communities where unions operate. According to the 2016 American National Election Study, the ISEA’s partisan composition mirrors that of the labor movement as a whole, suggesting broader applicability: 55% of union members identify as Democrats, 15% as Independents, and 30% as Republicans.

These findings also carry implications for our understanding of political behavior and attitudes in the context of civic organizations and institutions. This analysis underscores the importance of workplaces as sites for politically cross-cutting interactions. Despite mounting polarization, American workers still generally choose where to work on the basis of non-political considerations and therefore encounter coworkers or managers whose views are not aligned with their own (Hertel-Fernandez 2018; Mondak and Mutz 2001; Mutz and Mondak 2006). In turn, this paper illustrates how unions may be uniquely equipped to take advantage of the cross-cutting nature of the workplace to organize individuals across political lines. This study thus invites a closer examination of the workplace as a site of politics and political involvement (Hertel-Fernandez 2018; Li 2018).

A second broader implication of this study is the importance of federation and decentralization to civic associations seeking to recruit and represent a diverse membership. As Skocpol, Ganz, and Munson (2000) have argued, federated civic organizations emerged to parallel the political opportunity structures presented by the American government—and organizations that built a linked presence at the local, state, and federal levels have tended to be more politically influential than those that did not (Skocpol 2003). This paper reinforces the importance of federation not just for achieving political influence in government but also for accommodating diversity in membership (see also Newman and Skocpol 2023), especially in an era of increasing polarization and nationalization of politics (Baldassarri 2011; Hopkins 2018; Mason 2018). As a federated organization, NEA local and state associations can and do adopt messages and positions that differ from the national organization.¹⁵ As the interviews underscored, ISEA locals that were more successful at engaging conservative members did precisely this—talking about politics not in terms of national issues but instead stressing the connection between state and local politics and school conditions in their own districts. More research on federated organizations ought to study how they manage the tensions introduced by a diverse membership and between different layers of an organization taking different and potentially contradictory stands, as well as how federated organizations’ strategies may have changed in response to polarization and nationalization of politics.

Lastly, these results have a bearing on political representation in an era of rising economic inequality. Politicians tend to be more responsive to the interests of lower- and middle-income Americans where unions are stronger and more politically active (e.g., Becher and Stegmueller 2021; Hacker and Pierson 2010). Yet unions are under increasing strain in the face of economic and political pressures (Hertel-Fernandez 2019; Rosenfeld 2014). This study has suggested that there are specific organizational changes unions can implement to bolster political power even in the face of a challenging political climate. In turn, this implies that even in an era of union decline, labor organizations like the ISEA can still provide voice and political representation to their members and the broader working class (Rosenfeld 2014).