What Drives the News Coverage of US Social Movements? Free

Abstract

What drives the news coverage of social movements in the professional news media? We address this question by elaborating an institutional mediation model arguing that the news values, routines, and characteristics of the news media induce them to pay attention to movements depending on their characteristics and the political contexts in which they engage. The news-making characteristics of movements include their disruptive capacities and organizational strength, and the political contexts include a partisan regime in power, benefitting from national policies, and congressional investigations. To appraise these arguments, we analyze approximately 1 million news articles mentioning 29 social movements over the twentieth century, published in four national newspapers. We use negative binomial regression analyses and separate time-series analyses of the labor movement to assess the model’s robustness across different movements, time periods, and news sources. In each analysis, the results support the hypotheses based on the institutional mediation model. More generally, we argue that the influence of social movements on institutions depends on the structure and operating procedures of those institutions. This insight has implications for future studies of the influence of movements on major social institutions.

What drives the news coverage of social movements in the professional news media? The news of movements is a cultural consequence of them and marks their impact on the public sphere (Gamson and Wolfsfeld 1993; review in Amenta and Polletta 2019). Social movements seek to change modes of thinking, cultural codes, and public discourse and can do so through the news (Earl 2004). News coverage can also promote other key goals of movements—providing public attention to their issues, increasing their legitimacy and support, and influencing the political and policy agendas (Lipsky 1968; Baumgartner and Jones 1993; Koopmans 2004; Vliegenthart, Oegema, and Klandermans 2005; Andrews and Caren 2010; see reviews in Amenta et al. 2017; Caren, Andrews, and Lu 2020). However, news coverage can also discredit movements and their organizations and sometimes leads to their decline. That was the case for organizations ranging from the German-American Alliance to Students for a Democratic Society to the Black Panther Party (Gitlin 1980; Davenport 2010; Seguin 2016; Amenta and Caren 2022).

The news attention to social movements poses some puzzles for theoretical claims about the impacts of movements. This is especially true for the influences of political contexts on the news attention to movements (Schudson 2002). Political opportunity theory (Meyer and Minkoff 2004) and political mediation models (Giugni 2007) expect movements to benefit when their allies gain political power, but partisan regimes often provoke great news attention to their movement opponents. Conservative movement actors gained attention during Franklin Roosevelt’s 1930s Democratic administration, as did the Tea Party under Democrat Barack Obama; politically left anti war advocates gained attention during the 1980s Republican administration of Ronald Reagan as did the Resistance during Donald Trump’s administration. Also, a movement sometimes becomes highly newsworthy when its organizations and leaders are being called before Congress, as was true for the Communist party and the Teamsters in the 1950s. Finally, research shows that movements gain extensive news play when their issues first appear on the issue-attention cycle (Downs 1972). However, movements often become more newsworthy after the passage of beneficial legislation. That was the case for the US anti-alcohol and nativist movements in the 1920s, after the passage of Prohibition and anti-immigration legislation, for the labor and the old-age pension movements in the 1930s after the passage of the National Labor Relations Act and the Social Security Act, and for the Black rights movement in the 1960s, after the Civil Rights and Voting Rights Acts.

This paper seeks to answer this question and solve the puzzles by examining movements from the point of view of professional news organizations (Fishman 1980; Bennett 2007; Schudson 2011). Theoretically, we argue that specific internal movement characteristics and external political contexts interact with journalistic practices to propel movements into the news. The mechanisms of influence work in some ways that are similar to those for political outcomes, especially regarding the effects of internal movement characteristics (see review in Amenta et al. 2010). Resource mobilization (McCarthy and Zald 1977) and infrastructural resources theory (Andrews 2004) would expect organizational strength lead to influence for movements, and disruptive collective action is also sometimes argued to drive favorable political outcomes (Piven and Cloward 1977; cf. Amenta et al. 2010). Because organizational strength and disruptive capacities play into the operating procedures of professional news organizations and align with news values, we also expect them to drive news attention to movements. However, the political contextual conditions promoting political influence are expected to work differently for the news. We argue that having any coherent partisan regime in power will drive the news coverage of movements, whether the regime is friendly or hostile to the movement. These regimes heighten interest in social movements and their constituents. Moreover, we argue that having legislation passed in favor of a movement’s constituency will increase, not decrease, news attention for a movement, as this will make its organizations more legitimate as the sources of news. Finally, congressional investigations will also inject movements into the news, even though investigations are typically launched by their political opponents and typically damage movement organizations.

To appraise these arguments, we examine all articles mentioning national US social movement organizations in 29 movements across the twentieth century in four national newspapers. This research goes beyond standard analyses of the influence of movements, which usually relies on case studies of movement actors and limited news sources. We examine news coverage in the New York Times, which is typically analyzed by scholars. However, this source often focuses on regional movement collective action, and so, we also examine the Washington Post, Los Angeles Times, and Wall Street Journal. We also examine all major social movements across the entire twentieth century. We analyze the news attention to movements with negative binomial regressions and subject these analyses to a series of robustness tests concerning time periods, news sources, and movements to appraise the generalizability of the arguments. This sort of appraisal is rare for the literature on the consequences of social movements, which has been based on case studies (for robustness analyses of the news coverage of interest groups, see Binderkrantz, Bonafont, and Halpin 2017). We also include time-series analyses of the news coverage of the labor movement, which is often neglected in studies of social movements yet received the most news attention of any movement (Amenta and Caren 2022). To foreshadow the results, the hypothesized determinants have significant and expected effects on the news coverage of movements in almost all specifications. We conclude by discussing the implications of these arguments and findings for the influence of movements on other institutions.

Motivations for Research, an Elaborated Institutional Mediation Model, and Five Hypotheses

The news media’s attention to social movement organizations is important. It influences their growth (Vliegenthart et al. 2005), cultural and discursive change (Ferree, Gamson, Gerhards, and Rucht 2002; Amenta and Polletta 2019), and the identification of social problems and the development of public policy (Walgrave et al. 2008). Movement organizations provide critical resources to seek social change (McCarthy and Zald 1977), construct political identities and interests (Skocpol 1992), survive hard times (Staggenborg 1988), and spur civic engagement (Sampson, McAdam, MacIndoe, and Weffer-Elizondo 2005). The attention of the mass news media matters to these organizations (Ferree et al. 2002; Koopmans 2004), can increase their legitimacy (Berry 1999) and support (Vliegenthart et al. 2005; Banerjee 2013), and influence public opinion and the policy agenda (Lipsky 1968; Baumgartner and Jones 1993; Walgrave et al. 2008). However, mainstream news treatment can also discredit and harm movement actors and their campaigns (Davenport 2010; Seguin 2016; Isaac et al. 2022).

Although most arguments in the social movement literature apply widely across movements, little scholarship compares many movements and organizations over long periods of time or across many news sources (reviews in Earl et al. 2004; Amenta et al. 2017; Caren et al. 2020). Most US scholarship on movements focuses on the period after the Second World War and progressive movements, such as the Black civil rights and the second-wave women’s rights movements (Mac Sheoin 2016; Amenta and Caren 2022). Here we examine the news attention across the entire century and examine 29 movements, including movements of the political right. Most movement scholarship does not include the labor movement, and here, we include it in analyses and analyze it separately. We also go beyond the scholarly focus on the New York Times by analyzing three other national news organizations: the Washington Post, Los Angeles Times, and Wall Street Journal. We can also ascertain whether the determinants of movement news are peculiar to specific types of movements, time periods, or news sources. Moreover, most studies focus on movements when they are most active. Here, we compare across many movements over long stretches of time, including when they made little impression in the public sphere as well as the more publicized moments.

To advance hypotheses about the conditions under which movements will be highly newsworthy, we elaborate an institutional mediation model (see also Amenta et al. 2019) that starts from the literature on the social organization of the news (Fishman 1980; Bennett 2007; Schudson 2011). This literature highlights the influence of the institutional characteristics of the professional news media on their operating procedures and focuses on their news routines and news values. At the turn of the twentieth century, US newspapers won independence from parties and elites through mass subscriptions and extensive advertising revenues, were shielded from state interference through judicial decisions, and became more professionalized and nationalized in the twentieth century in tandem with the professionalization and nationalization of government (Schudson 1978; Starr 2004). US news organizations have conformed to a “liberal” model—dominated by for-profit, increasingly professionalized enterprises, retaining editorial pages with partisan slants, but focused on providing objective and balanced news (Hallin and Mancini 2004). The main news organizations became increasingly professionalized across the century, with a major leap forward at midcentury (Schudson 2011; Pressman 2018). Professional news organizations work from common standards, news values, and routines regarding what constitutes “news” (Gans 1979; Tuchman 1980). Professional journalists have been called a “herd of independent minds” in these shared understandings and procedures (Rosen 2020).

We also follow the political mediation model (Giugni 2007; King 2008) in arguing that movement influence over news institutions derives from both the characteristics of movements and the political contexts. Like political outcomes, the news coverage of challengers is external to movement actors, as news organizations decide when and whether to cover them. However, because news institutions work differently from political ones, the movement characteristics and actions and political contexts can differ from those posited by that model. As we indicate in more detail below, news organizations can view movements as newsworthy in terms of being widespread phenomena, through their concern with conflict and social order, or regarding their central focus on politics and policy, including their interest in political scandals. We propose two characteristics of movements and three political contexts that we expect will influence the news coverage of challengers, five hypotheses altogether.

Disruption, violence, and threats to social order are closely attended to by news media (Kovach and Rosenstiel 2007; Mencher 2008; see also McLeod 2007), and accordingly, we argue that a characteristic of social movements that will advance their news coverage is disruption. Disruption frequently backfires for movements seeking broader political influence (review in Amenta et al. 2010), and some studies find that protest-oriented movement organizations are not as well covered in the news as ones that rely on insider tactics with greater resources (Andrews and Caren 2010; Elliott et al. 2016). However, more disruptive and violent protests are more likely to be covered (McCarthy et al. 1996; Oliver and Myers 1999; Earl et al. 2004; Gillion 2020). In addition, disruption by protest-oriented movement actors can provide occasions for news coverage for more established, non-disruptive organizations (Gamson and Wolfsfeld 1993), leading to greater attention at the movement level. These points lead to our first hypothesis about the internal influences of movements on their news coverage. Each hypothesis below is expected to hold good net of other potential influences on coverage.

H1: A Social movement’s disruptive capacities will have a positive and significant influence on its news coverage.

Journalists view widespread social phenomena as newsworthy (Gans 1979), and for that reason, we expect movements with many organizations and members to be covered more frequently. This idea also fits resource mobilization theory (McCarthy and Zald 1977) and the infrastructural resources perspective (Andrews 2004) on the political impact of movements. Some studies also support this idea, as protest events are more likely to be covered in newspapers if a large organization is involved (Oliver and Maney 2000; Earl et al. 2004). Some movements have many organizations and members, such as the labor movement’s many unions or the environmental movement’s organizational effervescence since 1970, and across US history, many federated organizations with large mass membership had high public profiles (Skocpol 2003). This brings us to our second hypothesis about internal influences on the news coverage of movements.

H2: A social movement’s organizational strength will have a positive and significant influence on its news coverage.

News organizations are centrally concerned with politics and follow institutional political officials and policymaking processes as a matter of routine (Gans 1979; Fishman 1980; Tuchman 1980). Major changes in US policymaking, especially those related to social movements, often happen in “big bangs,” (Skocpol 1992) when partisan regimes take charge of the government (see review in Amenta et al. 2010). We expect that these partisan regimes will provoke both movement mobilization and their coverage, given the potential stakes involved by the rise to power of these regimes that can dramatically change policy. Although partisan regimes in the same ideological direction as movements will often aid their political efforts (Meyer and Minkoff 2004), we argue that the rise to power of unified partisan regimes, of the left or right, will boost the news coverage of all movements by spurring their mobilization. A unified Democratic regime in the 1930s spurred both the labor movement and conservative organizations, as did one in the 1960s for the civil rights and nativist movements. The unified Democratic regime under Barack Obama (2009–2010) spurred the rightist Tea Party and attention to it, similarly with the anti-war movement of the early 2000s when Republicans dominated Washington (Heaney and Rojas 2015). Our first hypothesis about the influence of political contexts on movement coverage is as follows:

H3: Partisan Political regimes, whether left or right, will have a positive and significant effect on social movements’ news coverage.

A second and, we argue, longer-term political effect on the news coverage regards the passage of policies connected to the constituencies of social movements. Policymaking receives a high profile in newspaper coverage, as it often greatly impacts peoples’ lives and involves prominent elected officials, whom reporters follow in their beats (Gans 1979; Fishman 1980; Tuchman 1980; Oliver and Maney 2000; Bennett 2007). Although the discussions of news cycles often focus on news attention in the discussion of a new policy (Downs 1972) and scholars focus on the influence of movements on political agendas (see Amenta et al. 2010), we argue that once policies are enacted, they will strongly influence movements’ future news coverage for several reasons. Historical institutionalists (Pierson and Skocpol 2002) also argue that policies alter politics, and we argue that there are positive feedback effects for policies on the news coverage of movements. Policies not only often encourage the mobilization of groups identified in them by providing resources and signaling their group interests (Campbell 2003) but also can legitimize advocacy and movement organizations connected to the policy, with specific organizations often being treated as spokespersons for the group regarding the issues. Such was the case with the veterans, veterans’ policy, and the American Legion; the elderly, Social Security, Medicare, and AARP; and African Americans, civil rights, voting rights, and the NAACP. We argue that policy influences movements and their coverage (Baumgartner and Jones 1993; Berry 1999), as political contention surrounding policies may draw newspaper attention to these organizations. These points lead to a second hypothesis about political contextual influences on the news coverage of movements.

H4: Policies favoring a movement’s constituency will have positive and significant effects on the news coverage of related movements.

A third hypothesized effect of political contexts on the news coverage of social movements comes from official congressional investigations. Because news organizations focus extensively on both politics and scandal (Mencher 2008), we expect movements with organizations under official political scrutiny to gain news attention. Throughout US history, congressional inquiries have been launched on organizations across the political spectrum, including those in the nativist, old-age, labor, communist, civil rights, and anti-war movements, (Maher, Seguin, Zhang, and Davis 2020; Seguin, Maher, and Zhang 2021). Although we expect congressional investigations to increase news about movements, they seem likely to bring bad news that harms their targets. Also, we see them as providing a short-term effect on the news. When the investigation ends, the coverage should end. This leads to our third hypothesis about the influence of political contexts on the news coverage of movements and fifth hypothesis:

H5: Congressional investigations of social movement organizations will have a positive and significant effect on their movements’ news coverage.

Data, Methods, and Measures

To appraise these hypotheses and ascertain the determinants of coverage, we employ data on 29 US social movements in the twentieth century from the Political Organizations in the News (PONs) data set. PONs include every mention of 1,514 organizations in national newspapers in the twentieth century. These organizations include politically oriented ones with national goals. Searches generated about 415,000 articles in the New York Times, 292,000 in the Washington Post, 282,000 in the the Los Angeles Times, and 74,000 in the Wall Street Journal. Each organization was allocated to one of 29 substantive movements or three residual movements—progressive, other; conservative, other; and civil rights, other. This grouping is mutually exclusive and exhaustive. We analyze the total article mentions across the 29 movements over the 100 years of the twentieth century, with 2,024 movement-years to analyze (for further description, see the online supplementary material for Appendix).

The main thing we seek to explain is why movements were covered extensively when they were during the twentieth century. The main measure is news coverage. It is the number of articles that mention a movement organization in a movement, by year, across the four news outlets, leading to 2024 movement-year observations. We have six main causal measures that capture the hypothesized internal movement and external political influences on the news coverage of social movements.¹

We employ negative binomial regression models to estimate the effects of movement characteristics and political contexts on the news coverage of social movements. A generalized form of the Poisson model, negative binomial regression models are appropriate for analyzing count data with overdispersion, where the variance of the measure is greater than its mean (Allison 2009). Our news coverage measure exhibits this property, making negative binomial regression models suitable for analysis. Negative binomial regression models address overdispersion by introducing an additional parameter to Poisson regression models and estimate the coefficients using an unconditional maximum likelihood function.

Because news coverage of social movements is likely to cluster by movement and influenced by unobserved movement-specific processes, we specify “between-within” or “hybrid” negative binomial models (Allison and Waterman 2002; Greene 2005; Guimarães 2008; Allison 2009). This model specification allows us to adjust for confounding effects of within-movement covariates due to unmeasured between-movement covariates. First, movement-specific means are used to capture between-movement covariates. Then, we generate the within-movement covariates by centering all the time-variant covariates as deviations from movement-specific means (see also Baltagi 2008: 35–36). Finally, we estimate negative binomial regression models with both the deviations and the movement-specific means as explanatory variables. By assuming that movement-year article count, |${y}_{it}$|⁠, has a negative binomial distribution with expected value |${\lambda}_{it}$|⁠, the model can be written as

where (⁠|${\alpha}_i+{\varepsilon}_{it}$|⁠) is the composite error term.

The coefficients |$\boldsymbol{\beta}$| for the deviation variables can be interpreted as fixed-effects estimates, as the model controls for time-invariant covariates. The mean variables |$\boldsymbol{\gamma}$| control for the possible confounding correlations between time-varying variables and unobserved movement-specific heterogeneities (see also Bell and Jones 2015). The model with both the deviations and means as independent variables allows the interpretation of differences in the within-movement and between-movement estimates. This strategy also avoids issues with model convergence and improves computational efficiency, compared to fitting fixed- effects models with dummy measures using unconditional maximum likelihood estimation.

To address Hypotheses 1 and 2, we start with disruptiveness, which is a categorical measure that varies by movement and year and scores one if any organization in the movement was in the news and engaged in disruptive action such as large protests, strikes, boycotts, occupations, civil disobedience, and protests with violence or drawing the violent reaction of authorities, as reported in scholarly monographs, articles, organizational websites, news accounts, and other data sets (for details, see the online supplementary material for the Appendix). We have two alternative measures for organizational strength. The first, organizational presence, measures the number of organizations in a movement each year. Because early twentieth-century movements often included very large, federated membership organizations with state chapters with considerable autonomy (Skocpol 2003), we give movements additional credit for these chapters. Specifically, we count as separate organizations the state chapters of the federal organizations on Skocpol’s (2003: 26–28) list of the largest US voluntary membership organizations—each with at least one percent of the US population—when they were at this high level of membership. We also calculate a measure of organizations in the news, a subset of the previous measure, which counts organizations if they appeared in the news that year. In the analyses, we substitute this measure for the first organizational strength measure as a check on the results.

We also have two measures to address Hypothesis 3 regarding the influence of political contexts surrounding partisan politics. The first is partisan regime, a categorical measure that varies by year and identifies progressive and conservative regimes. A left-wing, progressive, or liberal regime is defined as having a liberal Democrat in the White House with a liberal majority in both houses of Congress. A right-wing or conservative regime is defined as having a conservative Republican in the White House and a conservative majority in both houses of Congress. A liberal Democratic president is one who scores 65 percent or higher on an averaged “social” and “economic” ideology score, and a conservative Republican president is one who scores 35 percent lower (Segal, Timpone, and Howard 2000). A liberal or left-wing Congress is one that has a majority in both Houses made up of northern Democratic and radical third-party representatives, plus 25 percent of southern Democrats. Similarly, a conservative Congress has majorities in both chambers of Republicans and right-wing third-party members, plus 40 percent of southern Democrats (Lewis et al. 2021). For the twentieth century, there are accordingly two periods of liberal dominance (1935 through 1938 and 1965 through 1966) and two periods of conservative dominance (1921 through 1930 and 1981 through 1982). Partisan regime is a categorical measure, and to check on it, we calculate a measure based on congressional ideology, using the D-W nominate “median representative ideology” score of each Congress (Lewis et al. 2021). Because the hypothesis focuses on more extreme partisan alignments, we employ the absolute value of this measure.

To address the fourth hypothesis about the impact of policy on news coverage for social movements, we calculate enacted and enforced policy, a time- and movement-varying ordered categorical variable. It represents the comprehensiveness of major policies, including court rulings, laws, and bureaucracies to enforce them, regarding a movement’s constituency and ranges from zero to five. The measure is based on monographs about the specific movements and related policies, agencies administering policies, and the Policy Agendas (Policy Agendas Project 2020) (see Appendix for details on its construction). The African American rights movement, for instance, received a score of zero from 1900 to 1941, when the Fair Employment Practices Act passed, moved to two in 1954 for the Brown v. Board of Education decision. It added points for the Civil Rights Act (1964), the Voting Rights Act (1965), and the Civil Rights Act of 1968 (also known as the Fair Housing Act). After the 1978 Bakke decision, the score was reduced to four.

To address Hypothesis 5, the measure of congressional investigations is the number of days (logged) a movement had an organization under investigation by Congress (Maher et al. 2020; Seguin et al. 2021). Frequently, social movement organizations were called on the carpet by the House Un-American Activities Committee (HUAC), which was formed in the 1930s and which investigated the Communist party for more than a decade in the middle of the century. However, other organizations from a wide range of movements were also extensively investigated and by congressional committees other than HUAC. In the first half of the century, these included the German-American Alliance, Anti-Saloon League, United Auto Workers, and Townsend Plan. In the second half of the century, Congress targeted the International Brotherhood of Teamsters, Ku Klux Klan, Black Panther Party, and Students for a Democratic Society. This measure is logged.

We also examine a series of control measures that might influence the news coverage of social movements. Two concern the output of news organizations, as the coverage of movement action is often related to available news holes (Oliver and Maney 2000). We include the yearly counts of all articles published and political articles, those mentioning either of the two major parties; each is logged. The resource mobilization theory (McCarthy and Zald 1977) suggests that movement activity is spurred by increased disposable income and possibly so is the news of movements, and so, we include real gross domestic product in billions of dollars (logged). (Johnston and Williamson 2021). Other scholars find movement actors are spurred by economic grievances (e.g., Caren, Gaby, and Herrold 2017), and to address that, we include the unemployment rate (Lebergott 1957; U.S. Bureau of Labor Statistics 2019) (descriptive statistics appear in table 1).

One key advantage of our data set with its coverage across movements and over time is that we can examine gaps in social movement research. We can appraise the hypotheses by examining the coverage of each of the 29 movements across the century in the four news sources but can also determine whether the results are similar, and the hypotheses are supported for the portions of the data set concerning the movements, news outlets, and time periods most commonly studied by social movement scholars. Notably, we can compare results regarding the highly studied New York Times results versus the less examined other national newspapers, the results for the more widely examined progressive movements with those of the right, and results from the well-studied period beginning in 1960 with results across the century.

Results

The initial multivariate negative binomial regressions, which examine whether the hypothesized measures influenced the news coverage of social movements, provide support for the hypotheses. We start by examining the influence of the control measures only (see table 2, Model 1). Each measure has a coefficient in the expected direction, and each has a significant influence on the measure of coverage (reported coefficients are for the centered measures, which indicate the predicted change in the logged newspaper article counts within a movement over years, for a one-unit change in each predictor. All tests are one-tailed, given that the hypotheses and controls are directional). Models 2 and 3 each add five measures connected to the hypotheses. In these models, we separate the measures of the partisan regime and congressional ideology because they are different measures of a similar concept. As expected, each of the six coefficients is positive and significant, and likelihood ratio (LR) tests indicate that these models have a significantly better goodness-of-fit with the data than the control-based model. Some examples indicate the size of the effects. In Model 2, the exponential slope for disruptive capacities (⁠|${e}^{.808}=2.243$|⁠) suggests that having such capacities leads a movement to gain 124 percent more coverage in the news. The exponential slope for enforced policy (⁠|${e}^{.290}=1.336$|⁠) indicates that, in a year, with a one-step higher level in enforced policy, a movement tends to receive 33.6 percent more news coverage. The results are substantially the same with the second organizational measure (see Models 4 and 5).

Next, we analyze separately different news sources and different periods of time. We first examine the news of movements in the New York Times, which is typically analyzed by scholars, as it is the newspaper of record. Then, we compare those results with analyses of the other national news sources. In the analyses of the New York Times alone, each coefficient for the six measures remains positive and significant, although the coefficient for congressional ideology is significant only at the 0.10 level (see table 3). The results for the other three news sources are remarkably consistent. In the regression analyses including them, each of the six coefficients is positive and significant, and the fit of each model overall is like its counterpart for the New York Times. Scholarship has focused mainly on social movements that had made their greatest marks since the 1950s. Also, during that period, the Los Angeles Times and Washington Post became more nationally oriented and professionalized, modeling themselves on the New York Times (see Schudson 2011; Pressman 2018). For those reasons, results from this more recent period might be expected to diverge from the results from the earlier period. However, they do not (see table 3). The partial exceptions come from the measures of political partisanship. The coefficients for the measure of congressional ideology are not significant, and the partisan regime is significant only in the later period. It is possible that the results are due to reduced variance in the measures.

Next, we turn to the analyses of different types of movements. Most US social movement studies are on progressive movements (Amenta and Caren 2022), with many key theories about the causes and consequences of social movements based on the analyses of the Black civil rights movement and the second-wave women’s movement (e.g., Banaszak 1996; McAdam 1999; Andrews 2004; Soule and King 2006). For that reason, we start by analyzing progressive movements separately. This conceptualization excludes the four conservative or right-wing movements (namely, the anti-abortion, Christian right, gun rights, and nativist/white supremacist movements), as well as some others that are more difficult to locate on the right/left scale, including the anti-alcohol and veterans’ movements, or are more extreme, for a total of 21 movements and 1,458 movement-years of coverage. The hypotheses are largely supported by the results of analyzing progressive movements alone (see table 4) (analyzing the non-right movements separately produces similar results—not shown but see the online supplementary material for Appendix). Each coefficient for the measures related to the hypotheses is positive and significant, except the coefficient for partisan regimes, which is positive but insignificant. An important stream of social movement research, based on the Dynamics of Contention Action data set (McAdam et al. n.d.), excludes the labor movement from analyses, as its collective action is often targeted at businesses through strikes. Also, labor received the most news attention of any movement in the century while scoring very high in organizations and scoring one in every year for disruptive capacities. Again, however, the results excluding labor are not greatly different and support the hypotheses (see table 4).

Scholars argue that conservative movements or movements on the political right have different determinants—and often easier ones—of mobilization than left movements (McVeigh 2009; Blee and Creasap 2010) and may also have different paths to influence (Martin 2013), including in the news. Movements of the right also often view the professional news media as politically biased against their causes and constituents (Rohlinger 2015) and frequently seek policy gains while flying under its radar (Hertel-Fernandez 2019). To address the possibly different determinants of news influence for these movements, we analyze separately four movements from the right: the anti-abortion, Christian right, gun rights, and nativist/white supremacist movements. The results are more tentative, based as they are on a smaller number of cases, but they are largely in line with those for all movements (see table 4). In the model with partisan regime, each of the coefficients attached to the hypothesized measures is positive and significant, although investigation is significant at only the 0.10 level. In the model with congressional ideology, however, only three of the five coefficients are positive and significant, and the coefficient for congressional ideology is negative. Although many of the determinants appear to work similarly, it is worth exploring further the possibility of different determinants of news for movements of the right.

Finally, we examine the labor movement by itself. It received the most news attention of any movement, with its coverage ranging from under 600 articles in 1909 to almost 12,000 in 1937. Here, we also provide movement-related causal measures that apply specifically to labor, providing a closer connection between concepts and measures than is possible in analyses across all movements. The disruption measure is based on strikes, and we rely on workdays lost, a measure of strike “volume,” which ranges from around 60 million workdays in 1959 and 1919 to about 2 million, in 1999 (US Bureau of the Census 1949; U.S. Bureau of Labor Statistics 2019). For years where there are data only on the number of workers involved (1900–1905 and 1916–1926), we generate scores based on when data were complete (US Bureau of the Census 1949). For 1906 through 1915, during which official strike data are missing, we estimate the workdays lost from the University of Washington (2020) IWW History Project. We also replace the measure of organizational strength with one based on union members. For the years 1900–1929, we calculated union density by dividing union membership by employed workers (US Census Bureau 1949) and otherwise used the union density measure in Mayer (2004, see the online supplementary material for Appendix A, table A1). The other measures remain the same (see table 1).

Our modeling technique differs here. We employ time-series analysis. We do not use a negative binomial specification because the dependent measure and some of the independent measures are non-stationary and require differencing to avoid biases, which alters their distribution. The autocorrelation function in ARIMA indicated that union density and some control measures exhibit first-order autoregression, and so, we specify the model with differenced measures for them. Differencing corrects non-stationarity, minimizes serial correlation in the regression residuals, stabilizes the mean of the time series, and reduces trends in the data. Equation 2 presents the time series model we fit. |$\Delta{y}_t$| denotes the first-differenced measure of news coverage. |${\boldsymbol{x}}_{\boldsymbol{t}}$| denotes covariates that are stationary, and |${\boldsymbol{\eta}}_{\boldsymbol{t}}\mathbf{\Delta }{\boldsymbol{\gamma}}_{\boldsymbol{t}}$| denotes those that are non-stationary and thus first-differenced. We also control for the differenced |$y$| at time |$t-1$|⁠, denoted by |$\Delta{y}_{t-1}$|⁠, to account for lag effects.

In the models reported below, we did not include control measures, however, because each coefficient was insignificant, because of the loss of degrees of freedom, and because the results are similar when they are included (not shown, though available on request).

The results for the news coverage of the labor movement alone also largely support the hypotheses, though with some important caveats (see table 5). The coefficient for strike volume is positive and statistically significant. One thousand additional workdays lost resulted in 38 more articles in the news, controlling for other influences. Union density also shows a statistically significant and positive effect on newspaper coverage. A one percentage point increase in the union density results in 210 additional newspaper articles mentioning the labor movement, with all other measures controlled for. The coefficient for partisan regime is also positive and statistically significant, generating 637 additional newspaper articles in those years. However, the coefficient for enforced policy is not statistically significant, and the coefficients for congressional ideology and investigations are negative. The results are similar when the cases for which strike activity is estimated based on the IWW project indicators are dropped from the analyses or when only post-1915 data are analyzed (see the Appendix, tables A2 and A3). In short, most of the relationships regarding the determinants of news coverage across all movements hold good for the labor movement.

As with the wider analyses, we also examined the periods before and after 1960 to ascertain possible changes in influences over time (see table 5). The results are largely similar, though with some key differences. The influence of strikes on coverage is strong in the first period but not in the second, whereas the coefficients for union density are similar in each period, though stronger in the later one. The coefficient for partisan regime is positive in each period, though it is significant only in the first one. That result is like the results for all movements and is perhaps due to the relative lack of partisan regimes in the second period. The results are similar when the cases for which strike activity is estimated based on the IWW project indicators are dropped from the analyses or when only post-1915 data are analyzed (see the online supplementary material for Appendix, tables A2 and A3). The results regarding strikes and unionization support an earlier finding (Amenta et al. 2009) that news attention may be more dependent on disruptive action in earlier periods and on organization in later ones, at least for movements that were successfully organized. It also seems possible that the smaller coefficient for strike activity in the later period is due to strikes being smaller and less frequent in the later period.

Discussion and Conclusion

Gaining attention in the news is important for social movements for many reasons. On the positive side, news attention can help movements publicize social problems relevant to them, add their issues to political agendas, improve the public perception of their constituents, and boost support for their organizations. However, gaining unwanted and unfavorable attention can sometimes harm movement actors, their causes, and followers and lead to the decline of movements. We advanced hypotheses from an institutional mediation model built from insights from the literatures on the sociology of the media and social movements to help answer key questions about news attention to movements. We argued that the newsgathering routines and values of professional news organizations will lead them to cover movements depending on internal movement characteristics and political contexts that play into these routines and values. News organizations focus on conflict and widespread phenomena, and so, disruptive capacities and organizational strength were expected to drive news attention. The political conditions that play into news values and routines and influence the coverage of movements include partisan contexts that can make the major changes demanded by movements politically plausible, policy changes that aid movements, and congressional investigations of organizations.

We tested these hypotheses with a variety of regression analyses of the coverage of all national US social movement organizations over a century in four nationally oriented newspapers. We find that the institutional mediation model is supported robustly across a series of different specifications. In negative binomial regression analyses, the measures of organization strength, disruptive capacity, partisan regimes, policy gains, and congressional investigations had significant and positive influences on movement attention, the net of the influence of measures that might also influence this attention. These results often vary from findings in the more extensive literature on the political influence of social movements (see Amenta, Andrews, and Caren 2019). Although organizational presence is typically thought to aid the influence of movements in politics, reviews of research show that many of these other determinants of news coverage do not aid challengers’ political efforts. Disruption is expected to be helpful only sometimes in politics, winning new policy changes often signals the end of policy influence, and having a regime in power opposed to the movement tends to harm movements in their bids to influence political outcomes—whereas each increases news coverage. Although it has not been analyzed, being under congressional investigation will very likely also harm the political efforts of movement actors, despite resulting in increased news attention. In short, the reasons for movement influence over institutions is likely to vary across them (Rojas 2006; Giugni 2007; King 2008).

Because the Political Organizations in the News data set ranges across different news sources, types of movements, and time periods, we could ascertain how the results compared to the more limited types of analyses typical in the literature on social movements. These analyses have focused mainly on progressive movements in the second half of the twentieth century, with data coming from the New York Times. The results mainly hold up when we analyze the subsets typically found in these analyses. The hypotheses are supported when analyzing progressive movements or conservative ones, the New York Times or the other three news sources, the latter part of the century or the entire century, and without the labor movement or with the labor movement alone. There were some exceptions, however. Only in some cases did the determinants strongly influence the coverage of conservative or right-wing movements, suggesting that they may have additional or alternative drivers to their news attention. In the time-series analyses of the labor movement, moreover, political contextual measures, including policies in favor of organized workers and investigations, also proved not to advance news coverage significantly. In contrast, both strike activity and unionization were strong predictors. It seems possible that labor’s news attention depends or depended more on internal conditions and less on institutional political ones because its main target has been organized business. These results also support previous studies that suggest that movements of the right and the labor movement may be valuably analyzed separately.

These analyses and findings raise many questions for additional research. Just as is the case with influence of social movements in politics or over businesses (Giugni 2007; McAdam and Boudet 2012; Bartley and Child 2014; Dixon, Martin, and Nau 2016), gaining extensive news coverage for movements most likely requires the combinations of favorable internal movement conditions and political and news contexts. It is worth thinking further about which combinations will drive high attention to movements. Moreover, it is worth addressing the influence of disruption further and other aggregated aspects of movements examined here. Although disruptive action increases news attention to movements, the individual disruptive organizations may not have received the bulk of the coverage. Research on the environmental (Andrews and Caren 2010) and LGBTQ movements (Elliott et al. 2016) indicates that protest-oriented organizations receive less coverage than better-resourced and institutional organizations. Whether organizations’ disruption drives their own coverage or provides news opportunities for moderate organizations (Gamson and Wolfsfeld 1993), or both, is a subject for further research. Moreover, the arguments may apply best to highly professionalized and nationally oriented news organizations, as well as to countries like the United States in having a liberal form of news media (Hallin and Mancini 2004).

Also, it seems likely that not all news is good news for social movements, in terms of news attention helping them gain supporters; improving their image; or influencing in their favor cultural, political, and other institutional outcomes. Research has found that news coverage tends to increase support for movement organizations (Vliegenthart et al. 2005), but that effect likely depends on the reasons behind their news, the characteristics of the organizations, and how they are treated in their news coverage. It seems likely that some forms of disruptive action produce news that is unhelpful to movement actors. The “protest paradigm” suggests that protest will usually be covered in news articles that focus on social order instead of the claims of movement actors and in ways that disparage them (McLeod 2007). Also, news coverage has often reduced the effectiveness of strikes taken by labor organizations (Isaac et al. 2022). In addition, publicity from investigations has been followed with setbacks for movement organizations, including the German–American Alliance, Anti-Saloon League, Townsend Plan, International Brotherhood of Teamsters, Progressive Party, Communist party, Ku Klux Klan, and Black Panther Party (Amenta and Caren 2022). Further research should address whether some of the drivers of news attention for social movements and movement organizations mainly provide negative treatments. Research should also address whether negative coverage harms movement organizations and their bids for influence in the policy process and over other institutional actors.

Further research will also be needed to ascertain whether these results hold up in the current news media ecosystem, which has been transformed over the last two decades with the rise of right-wing disinformation outlets, notably Fox News, 24-hour cable news channels like CNN, and the ubiquity of the internet and social media, accompanying the decline of print news organizations (Pew Research Center 2015). For all these changes, the national news organizations remain central institutions of newsgathering, retain great legitimacy, and have become relatively more important among professional news organizations with the decline of local and regional papers. The New York Times and Washington Post still set the agenda for television network news, and their articles are amplified by aggregating websites and social media (Pew Research Center 2015; Gottfried and Shearer 2016). News organizations’ coverage has influenced recent European political agendas (Vliegenthart et al. 2016), the Tea Party’s mobilization (Banerjee 2013), and the discursive impact of Occupy Wall Street (Gaby and Caren 2016), yet today’s news ecosystem provides many more opportunities for movements of the right, which benefit from attention from disinformation media that sometimes induces coverage in the professional news media (Freelon, Marwick, and Kriess 2020; Amenta and Caren 2022).

Our arguments and findings suggest that the theories of the impact of social movements should take as their starting point addressing how those institutions work and what has been shown to influence them. Scholars should then theorize and analyze how movements might intervene in these causal processes. Here, we found that making sense of how movement actors are treated in the news meant taking the organization, routines, and news values of news organizations as the starting point. From there, we considered how social movements’ characteristics and actions, along with the political contexts in which they engaged, fit with media organizations’ routines, values, and approaches to deciding what is news. Similar institutional-centered thinking would be valuable for those studying other institutional targets of movements, such as corporations, universities, or religious organizations.

About the Authors

Weijun Yuan is a Ph.D. candidate in sociology at the University of California, Irvine. Her research interests include social movements, networks, and organizations. She studies the extent to which social movement organizations receive extensive, positive, and substantive news attention. In addition, her ongoing work focuses on the dynamics through which activist groups form coalitions and build consensus.

Neal Caren is an associate professor at the University of North Carolina at Chapel Hill. He received his Ph.D. in sociology from New York University. His research interests include social movements and the media. He uses quantitative methods to analyze large-scale datasets, including text and social media data, to better understand the dynamics of social and political change. He is the editor of the interdisciplinary social movements journal, Mobilization.

Edwin Amenta is a professor of sociology (and political science) at the University of California, Irvine. He is the coauthor (with Neal Caren) of Rough Draft of History: A Century of US Social Movements in the News (Princeton, 2022). He is the author of When Movements Matter: The Townsend Plan and the Rise of Social Security and Bold Relief: Institutional Politics and the Origins of Modern American Social Policy, both published by Princeton University Press.

Funding

This research was supported in part by the National Science Foundation (SES-1657872).

Data Availability

The data and code underlying this article are available at https://osf.io/xn5zf/.

Endnotes

References