Abstract
Although declining newspaper circulation has important political consequences, less is known about how this decline affects local elections. We address this gap, pairing evidence of legislative ballot rolloff with data on news circulation. Results suggest that newspapers play a role in whether voters leave a portion of their ballot blank. This relationship is particularly acute in noncompetitive races, suggesting that the absence of newspapers is most important when other democratic institutions are less robust.
Total newspaper circulation in the United States dropped every year between 1987 and 2018 (Pew Research Center, 2018). In 2018, only 16% of the public reported getting their news from print newspapers, lower than online news, social media, radio and television (Pew Research Center, 2018)—although some news they consumed online surely was produced by newspapers. “Algorithm reliance” has often supplanted editorial decision-making in the types of stories readers are able to view (Abernathy, 2020).
The picture is especially grim for small-market newspapers. Between 2004 and 2019, nearly 2,100 of the country’s 8,900 local newspapers—almost all weeklies—closed their doors and weekday circulation among daily newspapers fell by more than 54 million readers (Abernathy, 2020). The decades-long overall consolidation of the newspaper industry has continued, even while many newspaper chains have come and/or gone. These changes are alarming to both journalism practitioners and scholars, raising serious questions about how content is to be delivered, equity of access to information and the extent to which the news is robustly “place-based” (Usher, 2019).
These changes also have downstream consequences on Americans’ political attitudes and behaviors (Darr et al., 2018). Newspapers have the potential to both inform voters and improve political participation rates (e.g., Delli Carpini et al., 1994; Eveland & Scheufele, 2000). This makes sense. Newspapers provide information, and informed voters are more likely to participate in elections (Lassen, 2005). However, there are three challenges in connecting the dots from declining circulation to voter participation rates. First, while the news-turnout connection is intuitive, the empirical literature is mixed on whether declining circulation is leading to lower turnout (Gerber et al., 2009). Second, insofar as newspapers are being supplanted with a growing online and national news presence, it is unclear how the trend toward election “nationalization” and the increasing online character of the news influences local elections specifically (Hopkins, 2018). Finally, a cursory look at presidential election returns reveals that, if anything, the era of declining local paper circulation coincides with an uptick in voter turnout. (For example, see McDonald [2019], http://www.electproject.org/national-1789-present) How does one square the intuitive story of news-driven turnout decline with increasing turnout in presidential contests?
We address these puzzles, making the case that the decision to vote is distinct from the decision to participate in electoral contests all the way down the ballot, and that this latter decision is most clearly connected to local newspaper decline. The phenomena of turning out to vote but not completing the ballot is known as “ballot rolloff” (Wattenberg et al., 2000). Ballot rolloff, we contend, is uniquely connected to the local news media environment. Voters face unique informational costs when evaluating local candidates (Mondak, 1995). Local news provides voters with information on local races, and these contests are still, to an extent, independent of national politics (Abramowitz & Webster, 2016; Hopkins, 2018). By comparing districts with low and high rolloff rates, we can draw inferences about the extent to which news availability contributes to these differences. We can also examine the extent to which the role of local news media is moderated by other factors. In particular, we investigate the role of competitive electoral campaigns, which also can supply voters with information about local politics (Lipsitz, 2011).
We proceed by situating the rolloff/news connection in the broader literature on political participation. We hypothesize that rolloff will be highest in communities with limited news availability. We test this proposition with a novel dataset, combining data on state legislative rolloff with news circulation data. The evidence is consistent with the hypothesis, suggesting that local news is a critical informational component in a healthy participatory culture. The evidence also speaks to broader debates about the changing role of “place” in journalism.
Local News, Information and Local Political Participation
Connecting Local News to Voter Turnout
Scholarship indicates that newspapers play a role in increasing electoral participation (McLeod et al., 1999; Schulhofer-Wohl & Garrido, 2009; Shaker, 2014). For example, Gentzkow et al. (2011) used a large-scale difference-in-differences study to track the effect of newspaper entry and exit on voting patterns, finding a relationship between their presence and turnout in presidential and congressional elections. Similarly, Hayes and Lawless (2015) analyzed data from the Cooperative Congressional Elections Study (CCES), finding that individuals who are exposed to more media report are more likely to vote. In a follow-up study, Hayes and Lawless (2018) used a panel design to track self-reported political knowledge and participation as a function of coverage of congressional races by local papers. Results suggested that declines in local news corresponded to declines in both political knowledge and engagement. This is a powerful demonstration. However, the authors measured participation indirectly by asking respondents about “intention to vote” in the congressional race to avoid conflating top-of-ticket turnout with down-ballot participation. To preview our approach, treating ballot rolloff as a separate measure of electoral engagement allows us to directly examine the relationship between local news and local participation.
Information is the primary explanation for the newspaper/participation connection (Larcinese, 2007; Lassen, 2005). Arnold (2004) found that newspaper coverage and readership have a strong relationship with knowledge of candidates. Similarly, Peterson (2021) found that individuals in districts better-aligned with local newspaper markets were better at identifying the names of their representatives and placing them on an ideological scale. Eveland and Scheufele (2000) used data from the American National Election Study to demonstrate that heavy use of newspapers and television helped to closed information gap between well-educated and poorly educated people (see also de Vreese & Boomgaarden, 2006; Ksiazek et al., 2010; Prior, 2007).
Critically, the decision to participate depends on specific information about a particular race or candidate—being more informed generally is not universally predictive of turnout. Mondak’s (1995) natural experiment in Pittsburgh and Cleveland provides one example. Using an 8-month newspaper strike in Pittsburgh as a source of exogenous variation, Mondak found a large drop in self-reported knowledge of local political campaigns relative to citizens in Cleveland. Knowledge of national and international politics remained unchanged, indicating the importance of local print media in local politics. Likewise, Hopkins (2018) argued that place of residence (particularly living in state capitals) and news media content are associated with higher levels of local knowledge. Respondents who relied primarily on state news media sources were more likely to list the names of in-state representatives on a naming task, suggesting an association between the news media one consumes and the type of information one possesses about politics (see also Delli Carpini et al., 1994). Of course, differences in news circulation could be endogenous to the actual interest levels of voters. Snyder and Stromberg (2010) addressed this issue by examining political knowledge in areas with exogenous differences in news coverage of members of Congress. Citizens who have more access to coverage of their member of Congress are better able to recall details about their member than those who have less access.
As Lipsitz (2011) has pointed out, newspapers are not the only source of information available to voters—In electorally competitive environments, campaigns can also play a role in informing the public. Because competitiveness can also influence voters’ informational wherewithal, it is important that any study of voter participation take stock of the extent to which campaigns are subsidizing the costs of information. However, while other scholars have connected electoral competition to ballot rolloff (Hall, 2007; Wattenburg et al., 2000), this research does not test for the possibility that news media might also function as an important informational resource in down-ballot voting.
Complicating the News–Turnout Connection
Although the literature reviewed above suggests an intuitive link between local newspapers, informed voters and turnout, there are reasons to suspect that the role of newspapers is more nuanced. First, studies have found inconsistent evidence of a stimulative effect. Gerber et al.’s (2009) field experiment found that providing Virginia voters with a newspaper subscription did not increase turnout in either the 2005 or 2006 elections. Eveland and Scheufele (2000) found that newspapers are associated with increases in both political knowledge and overall political participation, but failed to find a significant effect of newspapers on voting. Cook’s (2019) study of a local election involving school consolidation concluded, “While the paper did effectively reach and inform community members, it was not a primary source of inspiration for civic or political action” (p. 84).
A more pressing critique is that local newspaper decline has not occurred in a vacuum. Elections are increasingly taking on a national character. Scholars have argued that national political dynamics (i.e., support of or opposition to the president) are important to understanding turnout (Carson & Hitefield, 2018; Hopkins, 2018). To the extent that this is the case, interest in the presidential election might suffice to drive turnout, and this engagement in national politics could be engendered by cable news, not necessarily small-town newspapers (Hopkins, 2018).
Although all types of races have become nationalized (such that voters might utilize a consistent party cue to choose candidates down the ballot), nationalization is less prominent in state legislative contests (Abramowitz & Webster, 2016). This pattern is reflected in split-ticket voting. Voters who prefer one party at the national level are open to voting for the other at the local and state levels. (While nationalization is clearly a more important factor in voting at the federal versus state and local level, we do not wish to overstate the case. See Rogers [2016] and Zingher & Richman [2019].)
Indeed, Darr et al. (2018) found evidence of less ticket-splitting in areas with local paper closures, suggesting that local papers themselves play an important role in countering the trend toward nationalization. (Darr et al. [2018] do not find evidence of less rolloff. However, this study examines Senate contests, not races further down-ballot. As we argue shortly, rolloff is a more pressing issue in low-information elections, so the null finding here is not especially surprising.)
This implicates a role for information in local elections. If voters in state legislative contests are open to ticket-splitting given particular local circumstances, it is plausible that voters might also abstain from participation in the absence of such information. Thus, we would also expect to find higher rates of ballot rolloff in state legislative contests in places with low local news circulation.
The participatory consequences of nationalization are nuanced and depend on the characteristics of the race and local media market. Hopkins (2018) finds evidence of declining mayoral turnout relative to presidential turnout, greater interest in federal races relative to state races and more internet searches for presidents than governors. However, news media in state capitals contain more coverage of legislative affairs. Living in state capital media markets is accordingly associated with a higher rate of gubernatorial turnout relative to presidential turnout. This suggests that news media exposure that is decidedly de-nationalized can improve knowledge of and participation in local politics—even as this local coverage grows increasingly rare (see also George & Waldfogel, 2006; Mondak, 1995). Importantly, to isolate the association between state/local news media coverage and turnout, Hopkins (2018) is primarily looking at elections that were not concurrent with the presidential contest. Previewing our own argument, the phenomenon of ballot rolloff allows us to extend Hopkins’s findings by examining participation differences in contests on the same ballot.
A related concern is that the decline of print news media is occurring alongside the rise of digital news. A growing body of research suggests that online news can promote political mobilization (Bakker & de Vreese, 2011; Howard et al., 2011; Valenzuela, 2013). For example, Towner (2013) found that attention to online sources of information increased the likelihood of participation in the 2012 U.S. presidential election. Likewise, a panel study in the Netherlands suggested that social media news consumption increased internal efficacy and voter participation (Moeller et al., 2014). Although these studies suggest an important participatory role for digital news, this view is not universally held (Boulianne, 2015), and may vary by country and/or by time. Moreover, there is reason to suspect that online journalism tends to neglect information about local politics (Hayes & Lawless, 2018; Hindman, 2011). For example, the filtering algorithms used by platforms like Google tend to elevate news produced by national outlets, leading to diminished local content (Usher, 2019).
Thus, even if digital journalism has a stimulative effect, these effects may not permeate local politics. If ballot rolloff is persistently high in races where print news media circulation is low, it would suggest that digital media are not adequately “backfilling” informational deficits in local political coverage. This has broader implications for both newsrooms and journalism scholars. Researchers have suggested that digital journalism can perform important democratic functions (Jangdal, 2020), although the content of digital media is ultimately responsive to a different set of underlying incentives than print (Peters & Witschge, 2015). Ballot rolloff is an ideal way to test for potential democratic ripple effects in a changing news media landscape.
Rolloff and the Informational Environment
The previous section argued that a move toward nationalization of electoral politics and online journalism both have the potential to increase turnout, yet neither are likely to educate voters about local candidates (Hayes & Lawless, 2018; Hindman, 2011; Hopkins, 2018). For this reason, we suggest the absence of local newspapers will have its largest effect on ballot rolloff in local races. Aggregate trends are consistent with this claim. National turnout numbers have been inconsistent in the era of declining local news, but have generally trended upward. Turnout hit a local low of 51.7% in 1996 and a local peak of 66.7% in 2020. (For example, see McDonald [2019], http://www.electproject.org/national-1789-present and http://www.electproject.org/2020g)
In contrast, trends in rolloff mirror national patterns in newspaper circulation decline. Miller, M. (personal communication, January 27, 2020). precinct-level analysis puts Presidential-U.S. House race rolloff at 3.41% in 1996. House rolloff steadily climbed to 5.02% in 2008 (before dipping again in 2012). Although not definitive, these aggregate numbers are consistent with the notion that local newspaper circulation has an appreciable impact down the ballot.
Although local news and rolloff has not been studied, we know that other forms of information are important to down-ballot participation. (Information is not the only contributor to ballot rolloff. There is indeed a fatigue component to rolloff, and ballot formatting and clarity have roles to play in determining whether or not voters make a choice in each race; for example, Darcy & Schneider, 1989; Nichols & Strizek, 1995).
Wattenberg et al. (2000) concluded, “voters skip items on the ballot not because they are lacking in education or members of minority groups, but rather because they do not have enough information upon which to base a vote” (p. 248). Bullock and Dunn (1995) examined a municipal election in Atlanta, finding that the salience of the contest was a major factor in rolloff. Hall (2007) examined state supreme court elections, concluding that partisan labels (an informational cue) and competition (which provides information) can reduce rolloff. Streb et al. (2009) examined rolloff in low-information appellate court elections. In these contests, people make use of the cues they have available—partisanship, incumbency, gender and ethnicity. Voters’ reliance on these heuristics suggests that the main impediment to down-ballot participation involves informational deficits.
Hypotheses
We know that information is a precursor to participation, and that local newspapers are a critical information source for voters. Although other forms of political information (i.e., national news, social media) provide coverage of national and high-profile state races, they are not a suitable replacement for local newspapers. Accordingly, we hypothesize that
Furthermore, research indicates that competitive electoral environments are an important source of information for voters (Lipsitz, 2011). Numerous studies have found that competitiveness decreases ballot rolloff (Hall, 2007; Streb et al., 2009; Wattenberg et al., 2000). For this reason, it is likely that local newspapers will be missed less in competitive electoral environments. Specifically,
All this suggests an important, yet circumscribed, role for local newspapers. However, top-of-ticket competitiveness and campaign spending are capable of driving voters to the polls (Geys, 2006). Simply put, voters in competitive high-spending states will encounter a great deal of information about gubernatorial and Senate races and this alone is enough to increase turnout. Under these conditions, voters will likely be “engaged with and knowledgeable about national politics to the exclusion of state or local politics” (Hopkins, 2018, p. 3). Turnout, primarily driven by top-of-ticket races, will not be sensitive to deficits in local news. Accordingly, we suspect that
Data and Method
To examine the relationship between rolloff and newspapers, we merge data on local newspaper circulation with results from state legislative elections. Data on newspaper circulation come from the University of North Carolina’s Center for Innovation and Sustainability in Local Media’s Database of Newspapers (Abernathy, 2018). This dataset includes 8,000 newspapers, coded for location, total circulation and frequency of distribution (daily or weekly). To match the newspapers to their respective legislative districts, we aggregated the newspaper data at the level of municipality. Because these municipalities do not typically perfectly align with the upper chamber of state legislative districts, we relied on the University of Missouri’s Geographic Correspondence Engine (Geocorr) to make sure the circulation of each newspaper was correctly apportioned to the proper legislative districts. Geocorr is a tool that overlays one unit of analysis (like a city) onto another unit of analysis (like a legislative district), according to the amount of “overlap” in the two populations. In this case, we use Geocorr to create a list of Census-designated places (cities, towns and other municipalities) by the proportion of those places located in each upper legislative district in the country. When a municipality was not an official Census-designated place, as was the case for a small number of unincorporated towns and neighborhoods of mid-sized to large cities, we used publicly available maps of district lines to approximate the districts in which those places fit best. Newspaper locations were matched with places on the Geocorr list. In this way, we were able to allocate the appropriate local news circulation volume in every legislative district.
Data for ballot rolloff are taken from two sources. The first is a dataset of statehouse legislative results compiled by Carl Klarner (2018). The second was compiled by the news/commentary site Daily Kos and consists of 2016 presidential results aggregated at the statehouse district level. In cases where states have multiple members who are elected from the same district (e.g., Vermont), the total number of votes cast are divided by the number of seats being voted on. (We ran several models to address concerns about state-level abnormalities in election administration, including removing states with multi-members [under the assumption that these complex environments would augment the role of political information], and examining states that had a straight ticket voting option in 2016 [given concerns that this would increase nationalization effects in local races]. Each of these alternative samples yields the same substantive conclusion as the main models presented in the “Results” section.)
Following conventional practice, the dependent variable (rolloff) is measured as the proportion of votes cast in races for upper houses of state legislatures to votes cast on the same ballot for U.S. President in 2016 (Hall, 2007). By examining rolloff instead of raw percent turnout, we aim to effectively control for exogenous differences in motivation to vote in the presidential race that could be partly driving local turnout numbers. If one community has interest in the presidential election and another does not, then those two communities are expected to turnout at different levels, influencing baseline presidential and legislative participation. Using rolloff as a dependent variable captures the effects of the main explanatory variables of interest, independent of these “top-of-ticket” forces. This approach also has the benefit of sidestepping issues such as inconsistencies between voting-eligible and voting-age population, which can lead to inappropriate inferences about participation (Holbrook & Heidbreder, 2010).
We use a variety of independent variables to control for potentially confounding effects. We control for median age in the district, given connections between age and both turnout and print-media consumption (Simon, 1996). We control for the unemployment rate, the percentage of college education voters in a district and the median income of the district (in 1,000s of USD) to account for the relationship between socioeconomic status and political participation (Verba et al., 1995). We also control for the percent of the population that is White, given that race is related to both turnout and news media consumption patterns (Oberholzer-Gee & Waldfogel, 2005). These data are all taken directly from the U.S. Census Bureau’s American Community Survey (ACS) 2017 five-year estimates for 2016 state legislative districts. The unemployment, bachelor’s degree and race variables are rescaled to range from 0 to 100. We also include population density in 1,000s of people per square mile. This variable was computed by using the ACS 2017 five-year estimates for population divided by the total land area of the legislative district. Finally, borrowing from previous work that uses margin-of-victory to approximate electoral competitiveness (Streb et al., 2009; Wattenberg et al., 2000), we operationalize race competitiveness by taking the votes received by the runner-up and dividing that number by the number of votes received by the winner. (Descriptive statistics for all variables are reproduced in the supplemental online appendix.)
Results
We begin by examining the rolloff rates across legislative races. As Figure 1 illustrates, rolloff varies considerably by district. In all, 59% of all districts had ratios of between .9 and 1, meaning between 90% and 100% of people who marked a vote in the presidential race also marked one in the state house race (and between 0% and 10% did not). These rolloff rates are electorally consequential. In Minnesota, for example, where only about 4% of presidential voters did not cast a ballot (the sixth best performance in the nation), rolloff resulted in about 110,000 lost votes statewide, or an average of about 1,650 votes for each of the state’s 67 upper house districts. Eight legislative districts in the state were decided by a smaller margin.
Legislative to Presidential Vote Ratio
Next, we turn to
As we have no a priori expectations about whether per capita total, weekly or daily circulation is most closely related to rolloff, we model each of these key independent variables separately (Table 1). Table 1 also includes controls for race competitiveness, unemployment, education levels, median income, median age, population density and percent White. The dependent variable in all models is the ratio of votes cast in state legislative races over votes cast in the 2016 presidential race. A larger value on the dependent variable indicates that state legislative turnout more closely resembles presidential turnout. A positive coefficient therefore indicates that a variable is associated with decreased rolloff (higher participation down-ballot). Results are all unstandardized regression coefficients (standard errors in parentheses).
Explaining Rolloff With Circulation and Demographic Controls
p < 0.05; **p < 0.01; ***p < 0.001.
As Table 1 indicates, the coefficients for circulation are all positive and significant. This indicates that for each of the three types of circulation, an increase in the number of papers per person is associated with a decrease in the percent of people in a district who failed to vote down-ballot. That is, ballot rolloff decreased in districts with higher per capita circulation—a finding that is consistent with
Next, we fit a series of models to test how the competitiveness of a legislative election moderates the relationship between per capita circulation and rolloff. The results of the models, once again for total, weekly and daily per capita circulation, are reported in Table 2.
Explaining Rolloff With Circulation-Competitiveness Interactions
The results of the interaction models are consistent with
Predicted Rolloff Ratios
The evidence suggests that the effects of local news deficits are uniquely felt in down-ballot races, making rolloff the appropriate outcome variable. However, given research on the relationship between news media consumption and turnout, one might object that these effects generalize to all types of turnout. We test this objection by regressing turnout percentage on local news circulation. (We measure turnout as presidential votes divided by the voting-age population in the district. Voting eligible population numbers are not available at the state legislative district level.) Results are presented in Table 3. First, consistent with
Explaining Turnout With Circulation and Demographic Controls
p < 0.05; **p < 0.01; ***p < 0.001.
Discussion
The evidence presented in this article supports all three hypotheses: Districts with lower local news circulation experience higher ballot rolloff. This relationship is most pronounced in uncompetitive state legislative districts, suggesting that local news is an especially valuable information source when campaign activity is less pronounced. These same effects were not observed for turnout. Although there are several plausible explanations for this null effect, the pattern is consistent with the argument that voters are not driven to the polls in large numbers by local politics. Elections are often viewed through a national lens, and voters—immersed in cable news and online sources in a nationalized electoral environment—show up to the polls to vote for the president and other high-profile races. For this reason, it is not surprising that the presence or absence of a local newspaper has a negligible relationship with turnout. In 2016, if anything was going to grab the interest of low-propensity voters, it was the drama of Trump versus Clinton as depicted by national news outlets. But if the presidential campaign brought voters to the dance, it was local news that kept them there. Local papers have a tremendous impact on what happens after voters have registered their presidential preference.
These results invoke practical questions about local journalism and the changing media landscape. At one level, in places where news organizations are present and robust, they appear to be meeting the democratic needs of local communities. This is quite remarkable, especially given that newsrooms have suffered substantial cuts in recent years (Abernathy, 2020). This suggests that local newspapers are doing more with less, and are efficient in how they deliver critical information to voters, and/or cut political and government coverage last when cuts happen.
Our analysis leads to skepticism that technological changes like the move to social media and “user-generated content” can perform these informational/democratic function as well as traditional print news media. As Peters and Witschge (2015) have argued, with move to digital journalism, the focus may shift from citizen “participation through news” to “participation in news.” News outlets increasingly rely on the individualized experiences of users, often neglecting the role of local journalism in building a collective “public ethos” (Peters & Witschge, 2015, p. 31). Although the data in this article cannot directly evaluate user-generated content, the results are broadly consistent with Peters and Witschge’s (2015) main claim. If social media and other types of user-generated content were effectively backfilling news local deficits, we would not expect to find the relationships we observe in this article. Consider, for example, the significant interaction between electoral competition and news circulation. Long-shot candidates in districts without a local newspaper are incentivized to generate their own content, and with a multitude of free social media platforms at their disposal, even poorly funded candidates can mount a serious social media campaign that could theoretically inform voters of their candidacy and reduce rolloff in informational vacuums. However, while theoretically plausible, this is not what is occurring in practice. In uncompetitive districts without a local news presence, voters simply leave a large portion of their ballot blank, suggesting that user-generated (and even candidate-generated) media are not enough to overcome informational deficits.
We suspect that much of print’s unique informational role has to do with reporting and editorial practices that are simply not present in the digital realm. When it comes to campaigns, local reporters dutifully work a beat and conduct interviews that speak to the concerns of their community, and editors make determinations about what is and is not newsworthy. As Abernathy (2020) points out, digital platforms often rely on algorithms to determine newsworthiness. These two products are qualitatively different, and traditional print news media are better situated to produce “context-specific” and “authoritative” place-based knowledge (Usher, 2019). Journalism lacking sufficient command over “place-knowledge” could ultimately erode public trust in media, with “concerning consequences for the American public” (Usher, 2019, p. 49). Our research affirms this claim, suggesting that local journalism is uniquely equipped to meet basic democratic demands.
We would be remiss not to mention three limitations of this study. First, the matching procedure was imperfect. The lines of state legislative districts do not perfectly match the reach of local newspapers, and not all papers in the University of North Carolina database are located in Census-designated places. Although this is clearly a limitation, it is unlikely that this contributed to an overestimation of rolloff effects because there is no reason to suspect that circulation was over/under assigned to high/low rolloff communities. If this was the case, it should have also produced a significant coefficient in the turnout models, which it did not. A second limitation is that these data are not longitudinal—Any causal inference we draw is circumstantial. Future work would benefit from a panel design, examining the same districts as newspapers exit (and occasionally enter) the market. It is possible that high rolloff rates were caused by other underlying variables such as social capital or community-mindedness, and that this lurking third variable could be responsible for both high rolloff and low news circulation. Third, the literature review uncovers evidence that digital journalism might have an informational and participatory function. As we speculate above, there is reason to be skeptical of the claim that digital journalism is compensating for the loss of print, but unfortunately, we do not have data that will allow us to examine markets where subscription-based digital journalism or social media are playing a substantial informational role. This is an important direction for future research.
While it is important to acknowledge these limitations, it is also important to highlight the evidence documenting a relationship at the intersection of news media research and democratic well-being. Concerns about robust democracy align with concerns about a declining local press. In an electoral environment increasingly saturated by national affairs, voters are increasingly likely to think federally, but fail to act locally. Efforts aimed at reinvigorating the local press, our research suggests, will also spillover into efforts to reinvigorate local democracy.
Supplemental Material
sj-docx-1-nrj-10.1177_07395329211014968 – Supplemental material for Newspapers and political participation: The relationship between ballot rolloff and local newspaper circulation
Supplemental material, sj-docx-1-nrj-10.1177_07395329211014968 for Newspapers and political participation: The relationship between ballot rolloff and local newspaper circulation by Christopher Chapp and Peter Aehl in Newspaper Research Journal
Footnotes
Supplementary Materials
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
