A Multilevel Examination of Local Newspaper Credibility

Abstract

This study adopts a multilevel framework to test whether perceived credibility of local newspapers is explained by individual- and community-level factors. Data from a community survey in the United States show that structural pluralism is negatively related to local newspaper credibility. Data also reveal that conservative ideology, newspaper use, social trust, and political trust are significantly related to local newspaper credibility.

Keywords

media credibility structural pluralism social trust political trust political ideology

The motivation to study credibility is strong and multifold for local newspaper organizations. A lack of credibility weakens their audience base. Consumers are unlikely to read newspapers that they think are not credible (Kiousis, 2001). A decline in readership can also lower advertisers’ interests and weaken the financial base of newspapers (Kirchhoff, 2011). In fact, as data indicate a loss of public confidence in newspapers (Saad, 2016), the industry has experienced declines in circulation and advertising revenues (Barthel, 2016). Furthermore, a decline in credibility undermines the role of newspapers in informing citizens of public affairs. Research indicates that local newspapers are a key to learning about local issues and events and engaging in community activities (e.g., J. M. McLeod, Scheufele, & Moy, 1999).

Extant literature on media credibility identifies several key factors that systematically account for variations in audiences’ perceptions of credibility, including personal involvement, such as one’s group membership, issue stand, partisanship, and political ideology (e.g., Gunther, 1992; T.-T. Lee, 2010), audiences’ use of a particular medium (e.g., Johnson & Kaye, 2014; Kiousis, 2001), and social and political trust (e.g., Gunther, 1992; T.-T. Lee, 2005, 2010) in addition to demographic backgrounds (e.g., Johnson & Kaye, 2014; Robinson & Kohut, 1988). Recent work has revealed that these factors continue to play an important role in shaping audiences’ credibility perceptions of newer forms of media, such as online news media (Johnson & Kaye, 2010), social media (Johnson & Kaye, 2014, 2016; Kim, 2015), and citizen journalism (Carr, Barnidge, Lee, & Tsang, 2014).

Despite a rich body of research in this area, there are still avenues for theoretical extensions, and we have identified two such avenues. First, previous studies have not fully explored how a set of factors identified in the literature explains variations in local newspaper credibility across different communities. Prior research has focused primarily on the credibility of newspapers as a whole medium (Johnson & Kaye, 1998; Robinson & Kohut, 1988), with a limited focus on how people evaluate the credibility of local newspapers serving their respective communities. Second, although recent work has suggested the multilevel nature of credibility (Tsfati & Ariely, 2014; Yamamoto, Lee, & Ran, 2016), contextual factors have not been fully examined in this line of research. It is important to address this issue because people’s perceptions of reality are shaped by not only their personal characteristics but also social environments they interact with in daily living (Sampson & Wilson, 1995). It is therefore plausible that how audiences judge the credibility of local newspapers in their respective communities varies by their place of residence. In addition, contextual influences on audiences’ credibility perceptions have received limited attention in the United States. Whether extant findings are extended to a U.S. community context remains to be seen.

To address these limitations in the existing literature, the present study integrates individual- and community-level analyses to account for variations in the credibility of local newspapers, with a specific referent to local newspapers that serve respondents’ communities. To this end, we analyze data from a representative survey of individuals in a U.S. southeastern state. Based on the existing literature on this subject (e.g., Johnson & Kaye, 2014; T.-T. Lee, 2005, 2010; Yamamoto et al., 2016), we examine theoretically motivated individual- and community-level factors, including newspaper use, political ideology, partisanship, social and political trust, structural pluralism, and political heterogeneity. Findings will extend the literature by pointing to the importance of considering multiple levels of influence to fully understand why audiences differently perceive a news organization’s credibility.

Audience Perceptions of Credibility

In general terms, credibility is defined as “a perceptual variable rather than as an objective measure of the quality of some information or source of information” (Flanagin & Metzger, 2007, p. 321). Although some work treats credibility in terms of believability (Bucy, 2003), research in this domain typically approaches credibility as a multidimensional construct. The measurement of credibility has been varied, but credibility has been commonly examined in terms of accuracy, lack of bias, fairness, trustworthiness, and depth of information (e.g., Flanagin & Metzger, 2007; Johnson & Kaye, 2016). Hence, credibility lies in a message recipient’s perception of how, for example, accurate or fair a specific message or source of the message is. It is not a property that lies in a specific message or source of the message in and of itself that is judged based on observable criteria apart from an audience’s perception (Flanagin & Metzger, 2007).

Credibility has deep roots in journalism and mass communication research with three domains: message, source, and media credibility (e.g., Kiousis, 2001; Metzger, Flanagin, Eyal, Lemus, & McCann, 2003; Sundar, 1999). Message credibility, when applied to news, concerns the characteristics of news, or the degree to which news stories are considered credible. Sources include a broad range of message senders, such as news anchors (e.g., Anderson Cooper), news organizations (e.g., New York Times), and Websites (e.g., Yahoo!). Media credibility involves the qualities of a medium that distributes messages such as television and newspapers.

It should be noted that the boundaries among the three domains of credibility are not as clear as suggested above and likely overlap. For example, when people encounter a news story that they think is biased, that judgment can affect the way they evaluate the credibility of a news organization that produced the story. Similarly, if people think a news organization is politically biased, that perception can shape how they rate the credibility of individual stories produced by that organization. With this complexity in mind, the present study of local newspaper credibility likely falls within the domain of source credibility. Although local newspapers can be considered one type of media, we focus on a local newspaper that operates in an audience’s community. When one judges the credibility of a local newspaper in his or her community, he or she thinks about a specific newspaper, instead of the local newspaper as a whole. In this context, it is also important to point out that the literature in this area commonly uses the term, source, to refer to a specific communicator (Kiousis, 2001; Metzger et al., 2003; Sundar, 1999), instead of individuals who journalists interview to acquire information in order to write a story and those who are cited in stories.

Correlates of Credibility: Individual-Level Factors

Existing literature indicates that audiences’ credibility perceptions are shaped by their personal characteristics, including media use, political predispositions, and social and political trust (e.g., Johnson & Kaye, 2016; T.-T. Lee, 2010). First, the literature consistently shows that audiences who use a particular source or medium tend to judge that source or medium as credible (e.g., Johnson & Kaye, 2014, 2016). This linkage is consistent with the selective exposure thesis that people tend to favor and trust information sources that support their preexisting beliefs (Stroud, 2008). Based on this literature, we propose the following hypothesis.

H1: Newspaper use will be positively associated with local newspaper credibility.

Second, the literature indicates that political predispositions play an important role in audiences’ credibility judgments. One proposition in this area, derived from the hostile media effect, states that audiences with strong issue involvement perceive balanced news coverage as hostile to their own stands (e.g., Gunther, 1992). For example, T.-T. Lee (2005, 2010) showed that conservative ideology and Republican party affiliation were negatively related to media trust. T.-T. Lee (2005) argued that conventional journalistic norms require reporters to present competing viewpoints, and thus when news involves controversial issues such as gun control and climate change, it necessarily highlights conflict between established systems versus ideas that challenge the status quo. As conservatives and Republicans tend to embrace traditional values and practices, they may perceive such patterns of news reporting as biased in favor of a liberal agenda. Based on this reasoning, the present study tests the following hypotheses.

H2: Conservative ideology will be negatively associated with local newspaper credibility.

H3: Republican party affiliation will be negatively associated with local newspaper credibility.

Third, the literature shows that social and political trust can affect people’s perceived credibility of media. Social trust is defined as the belief that others will fulfill socially shared expectations in mutually beneficial ways (Putnam, 2000). Research shows that trusting people tend to trust others in general, participate in social and civic activities, and show confidence in public institutions (Putnam, 2000; Uslaner, 2002). Social trust would be expected to affect the way audiences evaluate the media, as trusting individuals tend to transfer their trusting stance toward an object of evaluation (Uslaner, 2002). Research also suggests that trusting people are more tolerant of alternative ideas, groups, and lifestyles (F. L. F. Lee, 2014). Therefore, they may be more tolerant and understanding of news reporting that may not be entirely fair and balanced. Prior research has indeed shown that personal trust is positively related to trust in the media (T.-T. Lee, 2005).

Political trust refers to one’s confidence in the political system (Pinkleton & Austin, 2001). Citizens high in political trust tend to trust the political system and believe that political actors serve the needs of the public, whereas those low in political trust feel that politicians care only about special interests (Pinkleton & Austin, 2001). Citizens who distrust politics also tend to distrust the media to report issues fairly and accurately (Cappella, 2002; Gunther, 1992). In particular, as politics is among, if not the, most salient news subjects audiences interact with on a daily basis, political news may come to mind when they think about how credible the media are, which may prompt the transfer of their preexisting political (dis)trust to the evaluation of media (T.-T. Lee, 2005). In fact, research has shown that political trust is linked with higher credibility ratings (Johnson & Kaye, 2014; T.-T. Lee, 2005, 2010). Therefore, we formulate the following hypotheses.

H4: Social trust will be positively associated with local newspaper credibility.

H5: Political trust will be positively associated with local newspaper credibility.

Correlates of Credibility: Community-Level Factors

As can be suggested from the above review, existing research has focused mainly on personal characteristics of individuals to account for variations in their credibility perceptions. Yet there are a few exceptions. A handful of studies have turned to contextual explanations to assess whether macro-level structural characteristics influence audiences’ credibility judgments. For example, Tsfati and Ariely (2014), using data from the World Values Survey, showed that levels of media trust were related to countries’ levels of post-materialistic value orientation. Analyzing data from the Japanese General Social Surveys, Yamamoto et al. (2016) showed that structural pluralism and political heterogeneity had a negative link with media trust.

Building on Yamamoto et al.’s (2016) work, we consider structural pluralism and political heterogeneity as community-level explanations for Americans’ perceived credibility of local newspapers. Structural pluralism refers to “the degree of differentiation in the social system along institutional and specialized interest groups lines, in a way that determines the potential sources of organized social power” (Tichenor, Donohue, & Olien, 1980, p. 16). The degree of differentiation in the social system involves the division of labor and heterogeneity. The division of labor grows in response to increased population and subsequent competition for limited resources for survival (Durkheim, 1893/1997). Seeking efficient means for survival, social actors divide up labor and become functionally specialized and interdependent on each other, which facilitates demarcation and common interests along the lines of specialized roles and functions (Durkheim, 1893/1997). The potential sources of organized social power involve degrees (de)centralization of power structure (Hindman, Littlefield, Preston, & Neumann, 1999). These dimensions are interrelated. In larger, more structurally pluralistic communities, power is dispersed among a wider range of organized groups, because such communities contain a large and heterogeneous population base necessary to form various organized groups with influence over decision-making processes (Fischer, 1975; Wilson, 1986).

The concept of structural pluralism was originally formulated to understand distinct patterns of news reporting and editorial decision making in different community settings (e.g., Donohue, Tichenor, & Olien, 1973; Tichenor et al., 1980). This tradition of research assumes that news production cannot be fully reduced to personal characteristics of individual journalists and, instead, constrained and/or enabled by community structural conditions in which individual journalists work (Donohue et al., 1973; Hindman, 1996; Tichenor et al., 1980). An accumulated body of work indicates that newspapers in more structurally pluralistic communities cover social conflict more often than those in less structurally pluralistic communities (e.g., D. P. Demers, 1996; Tichenor et al., 1980). Social conflict is a common feature in more structurally pluralistic communities, as a greater diversity of organized groups compete to achieve their goals in decision making. In contrast, social conflict is viewed as disruptive to community cohesion in less structurally pluralistic communities, and hence it is commonly settled through consensus among select elites in a centralized power structure (e.g., D. P. Demers, 1996; Tichenor et al., 1980). Differential coverage of social conflict is reflective of such community characteristics.

Why might structural pluralism affect audiences’ credibility perceptions? Two explanations are plausible. First, people in more structurally pluralistic areas are more likely to encounter, experience, and learn a greater diversity of social groups and their corresponding cultures, lifestyles, and institutions than those in less structurally pluralistic areas. Although research points to people’s tendency to interact with like-minded others (McPherson, Smith-Lovin, & Cook, 2001), pluralistic community settings at least present opportunities for exposure to, and potentially understanding of, different ways of life. However, research has shown the skewed, inaccurate portrayals of various social groups in mass media, with dominant elites, institutions, and values preferentially treated over alternative ones that challenge the status quo (e.g., Donohue, Tichenor, & Olien, 1995; Hallin, 1986; Larson, 2006; D. M. McLeod & Hertog, 1999). Thus, residents of structurally pluralistic communities may find discrepancies between what they experience in the course of their daily living and media portrayals of the reality they know. Such discrepancies likely make them question the media’s ability to properly cover their communities. In this sense, people use community settings as a cognitive landscape to understand the larger society (Sampson & Wilson, 1995), including the performance of a news media organization.

Alternatively, news media in more structurally pluralistic communities tend to adopt more critical, investigative reporting, and such divisive news may potentially generate mixed responses from audiences from various spectrums of opinions. In fact, research has uncovered that newspapers in more structurally pluralistic communities take a more critical stance toward established, mainstream groups such as governmental agencies and corporations, as they reflect to some degree the diversity of the communities they serve (D. P. Demers, 1996). Moreover, D. K. Demers (1998) conducted a survey of journalists at U.S. newspapers and a survey of mayors and police chiefs in cities served by those newspapers. He revealed that mayors and police chiefs in more structurally pluralistic communities perceived local dailies in their communities were not fair and did not tell the whole story in terms of coverage of their office or department.

Both scenarios are more likely to be observed in more structurally pluralistic communities characterized by a greater concentration of sociocultural diversity. Thus, we formulate the following hypothesis:

H6: Structural pluralism will be negatively associated with local newspaper credibility.

The second community-level factor we examine is political heterogeneity, which is defined as the extent to which a range of political groups and support are evenly present in a certain community (Scheufele, Hardy, Brossard, Waismel-Manor, & Nisbet, 2006). Political heterogeneity indicates the potential for exposure to, and learning of, different political views (Scheufele et al., 2006), with members of politically heterogeneous communities more likely to encounter a greater diversity of political viewpoints and differences than those of politically homogeneous communities.

Scholars have examined the utility of political heterogeneity in a political discussion context. Political heterogeneity is politically important, as it fosters cognitive engagement by motivating one to learn about alternative viewpoints and reflect on the validity of his or her own issue stands (Scheufele et al., 2006). It also fosters seeking and careful processing of political information, because people who talk often about politics with individuals of different political backgrounds tend to anticipate encountering different views in future discussions and thus need to prepare for disagreements in discussions (Scheufele et al., 2006). Existing evidence indicates that political discussion heterogeneity leads to positive political outcomes, such as exposure to different political viewpoints (Diehl, Weeks, & Gil de Zúñiga, 2015), hard news use (Scheufele, Nisbet, Brossard, & Nisbet, 2004), and political participation (J. M. McLeod et al., 1999; Scheufele et al., 2004).

Political heterogeneity in a larger community environment plays a similar role in fostering an informed and engaged citizenry. In particular, Scheufele et al. (2006) examined how community political heterogeneity affects individuals’ political engagement. Their study showed that independent of racial and religious diversity, political structural heterogeneity was positively related to political discussion heterogeneity at the individual level, which in turn was positively related to political participation. The finding suggests that political heterogeneity is a unique contextual factor that needs to be examined in its own right. Thus, structural pluralism and political heterogeneity are analyzed separately in the present study.

Based on the above literature, it is plausible that political heterogeneity influences audiences’ perceptions of local newspaper credibility. Members of politically heterogeneous communities are more likely to encounter and learn a greater diversity of political viewpoints than those of politically homogenous communities. Such a political environment can foster an informed and sophisticated citizenry (e.g., Scheufele et al., 2006). Research indicates, however, that mainstream media coverage of politics is narrow in scope. It tends to focus only on a small range of issues, events, and candidates, and they are typically framed in a way that does not fully inform the public and, instead, serves the interests of news organizations such as making a profit (e.g., Aalberg, Strömbäck, & de Vreese, 2012; Dimitrova & Strömbäck, 2012; Dunaway & Lawrence, 2015). Relatedly, media coverage of political parties and candidates is highly imbalanced, with minor parties and candidates receiving little news attention (Margolis, Resnick, & Levy, 2003). Those of politically heterogeneous communities would likely notice discrepancies between such patterns of coverage and political diversity they observe and experience in daily living and question the capacity of local media in their communities to adequately cover public affairs. Based on this reasoning, the following hypothesis is formulated:

H7: Political heterogeneity will be negatively associated with local newspaper credibility.

Method

Data Collection

Data in this study came from a statewide mail-out survey of residents in a U.S. southeastern state and U.S. Census archives. The survey used a probabilistic, representative sampling technique. Initially, 4,000 survey questionnaires were mailed between March 6 and 10, 2009, with 3,666 follow-up post cards mailed on March 19, 2009. The same questionnaires were mailed to 3,123 non-respondents between May 6 and 8, 2009. The survey was closed on June 23, 2009, with 1,154 completed responses returned. Out of the 4,000 residents, 184 were not eligible due to inaccurate address or no longer residing at the address. Therefore, the survey yielded a response rate of 30.2% based on 3,816 eligible residents and 1,154 completed responses. Using the 2000 U.S. Census data, the sample was compared with several population parameters of the state where the survey was conducted. On the whole, the sample was somewhat skewed toward male, older, and more highly educated individuals (see Appendix).

At the community level, county-level data were collected using the 2000 U.S. Census. Respondents were asked to indicate their county of residence. The county is an adequate unit of analysis in which political, economic, and sociocultural structure and influence develop. Therefore, the county represents meaningful variations in structural features and serves as a site in which a robust theory can be tested (Jeffres et al., 2012). After removing respondents who did not provide their counties of residence, the sample was reduced to 1,137 distributed across 114 counties out of 120. County-level data were added to an individual-level data file from the community survey, using the respondents’ counties of residence.

Measures

Local newspaper credibility

Two items were used to measure local newspaper credibility. Respondents were asked to indicate the extent to which they agreed or disagreed with the following two statements on a 5-point Likert-type scale (1 = strongly disagree; 5 = strongly agree): “My local newspaper provides unbiased information on issues affecting my community” and “My local newspaper provides accurate information on issues affecting my community.” It should be noted that both items used a specific referent to newspapers in respondents’ respective localities. The two items were summed to form an additive scale (M = 5.57, SD = 2.20, r = .87, p < .001). Although this measure does not fully cover all dimensions of credibility noted above, it still captures two important aspects of journalistic practices, a lack of bias and accuracy. Thus, it is reasonable to assume that the variable has at least adequate face validity.

Newspaper use

Newspaper use was measured by two items. Respondents were asked how often they read (1 = never; 5 = frequently), and how much attention they paid to (1 = not at all; 5 = close attention), newspaper stories about local politics and community issues. The two items were combined to form an additive scale (M = 7.24, SD = 2.20, r = .68, p < .001).

Political ideology

Two items were employed to measure political ideology. Respondents were asked to indicate, on a 7-point Likert-type scale (1 = very liberal; 7 = very conservative), their stands on economic issues and social issues, respectively. The items were then combined to form an additive scale (M = 9.36, SD = 3.01, r = .80, p < .001). This measure has been found to be a valid indicator of political ideology (Pratto, Sidanius, Stallworth, & Malle, 1994). This measure differentiates among respondents on multiple dimensions of ideological orientation and thus helps extend prior research on media credibility that has shown the relationship between political ideology and media credibility using a common single item asking respondents how conservative or liberal they are on a scale (e.g., Gunther, 1992; T.-T. Lee, 2005).

Party affiliation

Party affiliation was measured by asking respondents to indicate which of the following they considered themselves to be: Republican, Democrat, Independent, or others. Responses were recoded with the high value as Republicans (M = .35, SD = .48).

Social trust

Social trust was measured by a single item. Respondents were asked to indicate, on a 5-point Likert-type scale (1 = strongly disagree; 5 = strongly agree), the extent to which they agreed or disagreed with the following statement: “I trust others in my community to do what’s best for the community” (M = 2.91, SD = .97).

Political trust

Two 5-point Likert-type scale items were used to measure political trust. Respondents were asked to rate the extent to which they agreed or disagreed with the following statements: “I trust public officials to make the best decisions for my community” and “I trust our local government to do the right thing” (Cho, Shah, Nah, & Brossard, 2009). The two items were combined to form an additive scale (M = 5.11, SD = 2.01, r = .75, p < .001).

Structural pluralism

A structural pluralism index tapped two structural dimensions (e.g., D. P. Demers, 1996; Hindman, 1996; Hindman et al., 1999; Nah & Armstrong, 2011). A structural pluralism dimension contained four indicators, including county population, percentage of education with bachelor’s degree or higher, per capita income, and percentage of occupations excluding agriculture, fishery, and forestry. An ethnic pluralism dimension was captured by percentage of minorities excluding White. The five indicators were first standardized and combined to form an additive index (M = 0.00, SD = 3.67, α = .79). Data were obtained from the 2000 U.S. Census.

Political heterogeneity

Blau’s (1977) diversity index was used to calculate political heterogeneity. The index is expressed as (1 – Σ $p_{i}^{2}$ ), where p represents the proportion that belongs to a given category and i is the number of different categories. A score of zero reflects perfect homogeneity and a score of one indicates maximum heterogeneity with members evenly spread among an infinitesimal number of categories. Based on previous research (Scheufele et al., 2006), a political heterogeneity index was based on the proportion of county votes each party (i.e., Democrat, Republican, Independent, and other) received in the 2008 U.S. presidential election (M = .47, SD = .09). The data were obtained from the U.S. Elections Archive (Leip, 2008). The political heterogeneity measure was moderately correlated with the structural pluralism index (r = .43, p < .001).

Statistical controls

To investigate the unique effects of the variables of theoretical interest, several variables found as correlates of American consumers’ perceived credibility of media were included as statistical controls (e.g., Johnson & Kaye, 2014; T.-T. Lee, 2005, 2010). Sex was measured with females as the high value (M = .34, SD = .48). Age was measured in an open-ended format (M = 56.61, SD = 13.94). Education was measured by a 7-point scale item based on the last school respondents had completed (M = 4.51 [about some college, no degree/associate’s degree], SD = 1.65). Race was measured with Whites as the high value and all others as the low value (M = .95, SD = .22). Television news use was measured by two items. Respondents were asked how often they watched (1 = never; 5 = frequently), and how much attention they paid to (1 = not at all; 5 = close attention), television news stories about local politics and community issues (M = 7.60, SD = 2.02, r = .64, p < .001). Online news use was measured by asking respondents, on a 5-point scale (1 = never; 5 = frequently), how often they read news stories about local politics community issues using the Internet (M = 2.49, SD = 1.37).

Analytic Strategy

The hypotheses were tested using multilevel modeling based on the SPSS MIXED procedure (Peugh & Enders, 2005), because respondents were nested within the counties of residence. One key consideration in implementing multilevel modeling is the sample size at each level for stable coefficient estimates. Although there are a few sample size guidelines in relevant literature (e.g., Mass & Hox, 2005), implementing such guidelines can be challenging with a small number of individuals or even a single observation dispersed across a larger number of groups. Such conditions could produce biased variance component estimates and thus erroneous contextual inferences. Simulation studies have reported that an increase in the proportion of singletons leads to biased variance component estimates (Clarke & Wheaton, 2007). Estimated variance components become more stable with an increase in average group sample size and the number of groups and a decrease in the proportion of singletons (Clarke & Wheaton, 2007). Similarly, Clarke (2008) showed that although variance component estimates were biased when groups were represented by two observations, they became more stable with five cases or more per group. Based on the literature, the present analysis focused on the counties with five or more observations. This selection criterion yielded a final analytical sample of 846 respondents spread across 65 counties, with the average number of respondents per county of 13.¹ The descriptive statistics of the variables reported above are based on this analytical sample.

To assess between-county variation in local newspaper credibility, an unconditional means model was first estimated. This model generated two variance component estimates—between- and within-group variances. The estimates yielded an intraclass correlation coefficient calculated of .0557, indicating that 5.57% of the total variance was explained by the county level while the remaining variability was attributable to the person level. Individual- and county-level explanatory variables were then introduced to account for the variability of local newspaper credibility. The model can be expressed as follows:

{(local newspaper credibility)}_{i j} = β_{0 j} + \sum_{(q = 1 - 11)} β_{q} X_{q i j} + r_{i j},

where β_0j is the intercept; X_qij is the value of covariate q for respondent i in county j; β_q is the regression coefficient for that covariate; and r_ij is the residual.

The between-county model can be expressed as:

β_{0 j} = γ_{00} + γ_{01} ({structural pluralism}_{j}) + γ_{02} ({political heterogeneity}_{j}) + u_{0 j},

where $γ_{00}$ is the grand mean; $γ_{01}$ and $γ_{02}$ are the regression slopes associated with structural pluralism and political heterogeneity, respectively; and u_0j is the county-level residuals. In the absence of theoretical reasoning, all within-county slopes were constrained to be constant across counties. The model was estimated with restricted maximum likelihood estimation. All predictors were grand-mean centered prior to analysis.

Results

Table 1 presents a multilevel model whereby the credibility of local newspapers is regressed on 11 individual- and 2 community-level predictors. H1 predicted that newspaper use would be positively associated with local newspaper credibility. This hypothesis was supported. As shown in Table 1, newspaper use had a positive relationship with local newspaper credibility (t-ratio = 3.58).

Table 1.

A Multilevel Model of Local Newspaper Credibility.

	Local newspaper credibility
	Coefficient^a	SE	t-ratio
Intercept	5.58	0.09	65.66***
Person level^b
Age	−0.01	0.01	−1.32
Sex (1 = female)	0.08	0.14	0.57
Education	−0.10	0.04	−2.24*
Race (1 = White)	−0.56	0.31	−1.81^†
Political ideology	−0.06	0.03	−2.59*
Party affiliation (1 = Republican)	−.03	0.15	−.17
Social trust	0.19	0.09	2.15*
Political trust	0.38	0.04	8.74***
Newspaper use	0.14	0.04	3.58***
Television news use	0.05	0.04	1.17
Online news use	−0.08	0.05	−1.62
Community level^b
Structural pluralism	−0.06	0.02	−3.21*
Political heterogeneity	−0.69	1.07	−.64
		Variance components
Within county		3.54
Between county		0.065

Entries are unstandardized coefficients.

Person-level n is 846. County-level n is 65.

†

p < .10. *p < .05. **p < .01. ***p < .001.

H2 stated that conservative ideology would be negatively associated with local newspaper credibility. This hypothesis was supported. Political ideology was negatively associated with local newspaper credibility (t-ratio = −2.59), with more conservatively oriented respondents reporting lower levels of local newspaper credibility than more liberally oriented respondents. In contrast, H3, which stated that Republican party affiliation would be negatively associated with local newspaper credibility, was not supported. The measure of party affiliation was not significantly related to local newspaper credibility (t-ratio = −.17).

H4 expected that social trust would be positively associated with local newspaper credibility. This hypothesis was supported. As shown in Table 1, social trust had a significant positive link with local newspaper credibility (t-ratio = 2.15). H5, which predicted that political trust would be positively associated with local newspaper credibility, was also supported. Political trust was related to increases in local newspaper credibility (t-ratio = 8.74).

Turning to the contextual effects, H6 predicted that structural pluralism would be negatively associated with local newspaper credibility. This hypothesis received empirical support. With 11 individual-level predictors and political heterogeneity simultaneously taken into account, structural pluralism had a negative link with local newspaper credibility (t-ratio = −3.21). In other words, respondents in larger, more structurally pluralistic communities tended to think that local newspapers in their communities were not credible than those in smaller, less structurally pluralistic communities. In contrast, political heterogeneity was not significantly related to local newspaper credibility (t-ratio = −.64). Therefore, H7 was not supported.

Summary and Discussion

Our individual-level analysis was confirmatory in nature. Consistent with the literature (e.g., Johnson & Kaye, 2014, 2016; Kiousis, 2001), newspaper use was associated with increases in local newspaper credibility, supporting the proposition that individuals who use a certain source tend to rate that source as credible. Next, consistent with prior work (Johnson & Kaye, 2014; T.-T. Lee, 2005, 2010), conservative ideology was negatively related to local newspaper credibility. As part of journalistic norms, news presents alternative ideas that may challenge established values. People with strong conservative stands, who tend to embrace the status quo, may find such patterns of news coverage as the consistent presence of a liberal agenda and bias, which may impact their credibility judgments. Unlike past studies (T.-T. Lee, 2005, 2010), however, partisanship was not related to perceived credibility of local newspapers, which might be due to how partisanship was measured in our study (i.e., Republican, Democrat, Independent). Past studies have used a finer-grained measure of partisanship assessing degrees of party identification (e.g., strong, moderate, weak Democrat). Finally, the relationships between social and political trust, respectively, and media credibility reported in prior studies were found to be extended to local newspapers (Johnson & Kaye, 2014; T.-T. Lee, 2005, 2010).

A unique conclusion can be derived from our community-level analysis. Our data showed that structural pluralism was negatively related to audiences’ credibility judgments of local newspapers in their communities. Based on the structural pluralism model (e.g., D. P. Demers, 1996; Hindman, 1996), two explanations are plausible. First, pluralistic community settings act as cognitive landscapes in which people form perceptions of reality (Sampson & Wilson, 1995). Members of more structurally pluralistic areas are more likely to observe sociocultural diversity than those of less structurally pluralistic areas. People may perceive discrepancies between what they encounter and experience in day-to-day living and skewed media portrayals of social issues, events, and groups, possibly leading to a critical evaluation of the media’s capacity to report the diversity and corresponding viewpoints present in pluralistic areas. Alternatively, news media in more structurally pluralistic communities tend to criticize and challenge conventional institutions and value systems than those in less structurally pluralistic communities (D. K. Demers, 1998; D. P. Demers, 1996). Such potentially divisive coverage may generate a wider range of reactions, including negative ones, from audiences from different spectrums of opinions.

Our study did not investigate which theoretical mechanism plays a greater role in accounting for the observed association. To extend this finding, future work could conduct a survey of people from various community structural settings, directly measure their perceptions of sociocultural diversity in their communities and how critical they perceive local media are of traditional institutions and value systems, and assess whether these perceptions are differentially related to their credibility judgments of local media. Also, it would be critical to content-analyze news items published by local media in more or less structurally pluralistic communities in order to examine the extent to which local media reflect sociocultural diversity in their communities. Doing so would help assess whether discrepancies between people’s perceived sociocultural diversity in their communities and patterns of media coverage indeed influence local media credibility.

Conversely, the finding indicates that members of less structurally pluralistic communities reported higher levels of credibility for newspapers in their communities. As implied by the individual-level predictors noted above, there seems to be a pattern underlying this result. For example, those with lower levels of educational attainment and higher levels of social trust tend to make up small, homogenous communities (Pew Research Center Social & Demographic Trends, 2007; U.S. Department of Agriculture, 2017). Moreover, independent of personal characteristics of residents, research shows that trusting attitudes and traditional ways of life tend to be more common in those communities (Lichter & Brown, 2011). Thus, structurally homogeneous community settings might foster more trusting attitudes toward established institutions such as news media.

In contrast, political heterogeneity was not significantly related to local newspaper credibility. This finding differs from prior research conducted in Japan (Yamamoto et al., 2016). It is possible that this contextual effect is unique to a specific society. It also is plausible that local newspapers in the U.S. adequately represent local political heterogeneity. The null finding also might be due to how political heterogeneity was measured. In the present study, it was measured based on the proportion of county votes each party received in the 2008 U.S. presidential election. As a majority of voters in the state where the survey was conducted preferred either Republican or Democratic candidates (about 95%; Leip, 2008), and as mainstream news media tend to focus on the two major parties (e.g., Margolis et al., 2003), a discrepancy between local political heterogeneity and media representation of politics might not have been substantial to the extent to which it would affect audiences’ evaluation of local newspaper performance.

What do the findings that structural pluralism was predictive of local newspaper credibility while political heterogeneity was not indicate? It is plausible that in a U.S. community context, community structure complexity derived from the presence and activity of various groups, organizations, and institutions matters more in audiences’ credibility judgments than a local public sphere as indicated by the presence of varied political views in the community. That is, the way U.S. audiences evaluate newspapers in their localities is shaped by community structural factors, such as socioeconomic status, economic development, and racial diversity, that are perhaps easily noticeable in the course of their daily living relative to political viewpoints and differences. Regardless, the finding aligns with the notion that credibility judgments are a multilevel phenomenon which cannot be entirely reduced to the personal characteristics of individuals and, instead, need a consideration of larger social structural factors that envelop individuals (Tsfati & Ariely, 2014; Yamamoto et al., 2016).

Despite these insights, the current study is still limited in a few important respects. First, the measurement of local newspaper credibility needs improvements. Although our measure tapped a lack of bias and accuracy, other dimensions of journalistic credibility such as depth of information were not captured (e.g., Johnson & Kaye, 2014). To replicate the present results, future studies should use an established scale of credibility. Second, the data were based on a survey of adults living in a U.S. southeastern state, and therefore the generalizability of the findings is necessarily limited. To address this limitation, the current findings need to be tested in more politically and civically diverse community settings. Third, the sample overrepresented male, older, and more highly educated individuals compared with the population of the state, which may have introduced nonresponse bias. Thus, caution should be taken in interpreting the findings.

Fourth, the data analyzed in this study were collected in 2009, which may reduce the applicability of the findings to today’s media landscape. In recent years, newspapers have seen increased competition from other media sources and online news Websites. The way audiences receive news also has been changing. Audiences no longer read newspapers exclusively in print or on their Websites. It is increasingly common for audiences to incidentally encounter and read newspaper articles through Web portals (e.g., Yahoo!), emails, and social network sites. In such incidental exposure, audiences may read only a linked or shared article (Barthel, 2016) and some audiences may judge the credibility of a newspaper based on such limited experience. To ensure that the findings are applicable to today’s news environment, future research should replicate the results with more recent data. Fifth, although the measure of structural pluralism was based on prior work, it can be refined by incorporating a direct measure of leadership structure (Armstrong, 2006). Finally, it would be useful to increase within-group sample size for more stable parameter estimates.

Nevertheless, the advantages of the current design tapping a probability-based community survey outweigh the above limitations. With a specific referent to newspapers serving respondents’ communities, we found that those in more structurally pluralistic places were less likely to perceive local newspapers in their communities as credible than those in less structurally pluralistic places. Our data also showed that newspaper use, political ideology, social trust, and political trust helped explain why some audiences more favorably assess their local newspapers than others, regardless of where they lived. Overall, the results indicate that local newspaper credibility is a multilevel phenomenon that warrants a consideration of both audiences’ personal characteristics and their community contexts.

Footnotes

Appendix

Sample Demographics

	Sample	2000 census
Male	64.1%	48.9%
Female	35.9%	51.1%
Age (median)	51	35.9
Race
American Indian and Alaska native	0.17%	0.21%
Asian	0.17%	0.74%
Black or African American	2.95%	7.32%
Native Hawaiian and other Pacific Islander	0.17%	0.04%
White	94.8%	90.1%
Others (e.g., some other, multiple races)	1.73%	1.61%
Education
Less than 9th grade	3.67%	11.7%
Some high school, no diploma	6.99%	14.2%
High school graduate or GED	26.11%	33.6%
Some college, no degree	25.41%	18.5%
Associate’s degree	8.30%	4.9%
Bachelor’s degree	15.02%	10.3%
Graduate or professional degree	14.50%	6.9%

Note. Entries in the sample column are based on the full sample (N = 1,154).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Masahiro Yamamoto (PhD, Washington State University, 2012) is assistant professor in the Department of Communication at the University at Albany, State University of New York. His research interests include communication in community contexts, civic and political participation, and social media. His work has appeared in such journals as Communication Research, Journalism & Mass Communication Quarterly, Mass Communication & Society, and New Media & Society.

Seungahn Nah (PhD, University of Wisconsin-Madison, 2006) is associate dean for Graduate Affairs and Research in the School of Journalism and Communication at the University of Oregon. His research centers on the interrelationships among communication, community, and democracy with special emphasis on the roles of digital communication technologies in community and democratic processes and outcomes. His work has appeared in such journals as Journal of Communication, Journal of Computer-Mediated Communication, New Media & Society, Mass Communication & Society, Journal of Broadcasting & Electronic Media, and Journalism: Theory, Practice and Criticism among others.

References

Aalberg

Strömbäck

de Vreese

C. H.

(2012). The framing of politics as strategy and game: A review of concepts, opertionalizations and key findings. Journalism, 13, 162-178. doi:10.1177/1464884911427799

Armstrong

C. L.

(2006). Revisiting structural pluralism: A two-dimensional conception of community power. Mass Communication and Society, 9, 287-300.

Barthel

(2016, June). Newspapers: Fact sheet. Pew Research Center Journalism & Media. Retrieved from http://www.journalism.org/2016/06/15/newspapers-fact-sheet/

Blau

P. M.

(1977). Inequality and heterogeneity. New York, NY: Free Press.

Bucy

E. P.

(2003). Media credibility reconsidered: Synergy effects between on air and online news. Journalism & Mass Communication Quarterly, 80, 247-264.

Cappella

J. N.

(2002). Cynicism and social trust in the new media environment. Journal of Communication, 52, 229-241. doi:10.1111/j.1460-2466.2002.tb02541.x

Carr

D. J.

Barnidge

Lee

B. G.

Tsang

S. L.

(2014). Cynics and skeptics: Evaluating the credibility of mainstream and citizen journalism. Journalism & Mass Communication Quarterly, 91, 452-470. doi:10.1177/1077699014538828

Cho

Shah

D. V.

Nah

Brossard

(2009). “Split screens” and “spin rooms”: Debate modality, post-debate coverage, and the new videomalaise. Journal of Broadcasting & Electronic Media, 53, 242-261. doi:10.1080/08838150902907827

Clarke

(2008). When can group level clustering be ignored? Multilevel models versus single-level models with sparse data. Journal of Epidemiology & Community Health, 62, 752-758. doi:10.1136/jech.2007.060798

10.

Clarke

Wheaton

(2007). Addressing data sparseness in contextual population research: Using cluster analysis to create synthetic neighborhoods. Sociological Methods & Research, 35, 311-351. doi:10.1177/0049124106292362

11.

Demers

D. K.

(1998). Structural pluralism, corporate newspaper structure, and news source perceptions: Another test of the editorial vigor hypothesis. Journalism & Mass Communication Quarterly, 75, 572-592. doi:10.1177/107769909807500311

12.

Demers

D. P.

(1996). The menace of the corporate newspaper: Fact or fiction. Ames: Iowa State University Press.

13.

Diehl

Weeks

B. E.

Gil de Zúñiga

(2015). Political persuasion on social media: Tracing direct and indirect effects of news use and social interaction. New Media & Society, 18, 1875-1895. doi:10.1177/1461444815616224

14.

Dimitrova

D. V.

Strömbäck

(2012). Election news in Sweden and the United States: A comparative study of sources and media frames. Journalism, 13, 604-619.

15.

Donohue

G. A.

Tichenor

P. J.

Olien

C. N.

(1973). Mass media functions, knowledge and social control. Journalism Quarterly, 50, 652-659.

16.

Donohue

G. A.

Tichenor

P. J.

Olien

C. N.

(1995). A guard dog perspective on the role of media. Journal of Communication, 45, 115-132. doi:10.1111/j.1460-2466.1995.tb00732.x

17.

Dunaway

Lawrence

R. G.

(2015). What predicts the game frame? Media ownership, electoral context, and campaign news. Political Communication, 32, 43-60. doi:10.1080/10584609.2014.880975

18.

Durkheim

(1997). The division of labor in society ( Halls

W. D.

, Trans.). New York, NY: Free Press (Original work published 1893).

19.

Fischer

C. S.

(1975). Toward a subcultural theory of urbanism. American Journal of Sociology, 80, 1319-1341. doi:10.1086/225993

20.

Flanagin

A. J.

Metzger

M. J.

(2007). The role of site features, user attributes, and information verification behaviors on the perceived credibility of Web-based information. New Media & Society, 9, 319-342.

21.

Gunther

A. C.

(1992). Biased press or biased public? Public Opinion Quarterly, 56, 147-167. doi:10.1086/269308

22.

Hallin

D. C.

(1986). The “uncensored war”: The media and Vietnam. New York: Oxford University Press.

23.

Hindman

D. B.

(1996). Community newspapers, community structural pluralism, and local conflict with nonlocal groups. Journalism & Mass Communication Quarterly, 73, 708-721. doi:10.1177/107769909607300315

24.

Hindman

D. B.

Littlefield

Preston

Neumann

(1999). Structural pluralism, ethnic pluralism, and community newspapers. Journalism & Mass Communication Quarterly, 76, 250-263. doi:10.1177/107769909907600205

25.

Jeffres

Horowitz

Bracken

C. C.

Jian

Neuendorf

K. A.

Yoon

(2012). Structural pluralism and the community context: How and when does the environment matter? Mass Communication and Society, 14, 787-815. doi:10.1080/15205436.2010.515371

26.

Johnson

T. J.

Kaye

B. K.

(1998). Cruising is believing: Comparing Internet and traditional sources on media credibility measures. Journalism & Mass Communication Quarterly, 75, 325-340.

27.

Johnson

T. J.

Kaye

B. K.

(2010). Still cruising and believing? An analysis of online credibility across three presidential campaigns. American Behavioral Scientists, 54, 57-77. doi:10.1177/0002764210376311

28.

Johnson

T. J.

Kaye

B. K.

(2014). Credibility of social network sites for political information among politically interested Internet users. Journal of Computer-mediated Communication, 19, 957-974. doi:10.1111/jcc4.12084

29.

Johnson

T. J.

Kaye

B. K.

(2016). Some like it lots: The influence of interactivity and reliance on credibility. Computers in Human Behavior, 61, 136-145. doi:10.1016/j.chb.2016.03.012

30.

Kim

(2015). Partisans and controversial news online: Comparing perceptions of bias and credibility in news content from blogs and mainstream media. Mass Communication and Society, 18, 17-36. doi:10.1080/15205436.2013.877486

31.

Kiousis

(2001). Public trust or mistrust? Perceptions of media credibility in the information age. Mass Communication and Society, 4, 381-403. doi:10.1207/S15327825MCS0404_4

32.

Kirchhoff

S. M.

(2011). The U.S. newspaper industry in transition. Journal of Current Issues in Media and Telecommunications, 2, 27-51.

33.

Larson

S. G.

(2006). Media and minorities: The politics of race in news and entertainment. Lanham, MD: Rowman & Littlefield.

34.

Lee

F. L. F.

(2014). “Tolerated one way but not the other”: Levels and determinants of social and political tolerance in Hong Kong. Social Indicators Research, 118, 711-727. doi:10.1007/s11205-013-0433-5

35.

Lee

T.-T.

(2005). The liberal media myth revisited: An examination of factors influencing media bias perception. Journal of Broadcasting & Electronic Media, 49, 43-64. doi:10.1207/s15506878jobem4901_4

36.

Lee

T.-T.

(2010). Why they don’t trust the media: An examination of factors predicting trust. American Behavioral Scientist, 54, 8-21. doi:10.1177/0002764210376308

37.

Leip

(2008). US president general—County level vote data (Dave Leip’s atlas of U.S. presidential elections). Available from http://uselectionatlas.org/

38.

Lichter

D. T.

Brown

D. L.

(2011). Rural American in an urban society: Changing spatial and social boundaries. Annual Review of Sociology, 37, 565-592. doi:10.1146/annurev-soc-081309-150208

39.

Margolis

Resnick

Levy

(2003). Major parties dominate, minor parties struggle: US elections and the Internet. In Gibson

Nixon

Ward

(Eds.), Net gain? Political parties and the impact of new information communication technologies (pp. 53-69). London, England: Routledge.

40.

Mass

C. J. M.

Hox

J. J.

(2005). Sufficient sample sizes for multilevel modeling. Methodology: European Journal of Research Methods for the Behavioral and Social Sciences, 1, 86-92. doi:10.1027/1614-2241.1.3.86

41.

McLeod

D. M.

Hertog

J. K.

(1999). Social control, social change and the mass media’s role in the regulation of protest groups. In Demers

Viswanath

(Eds.), Mass media, social control, and social change: A macrosocial perspective (pp. 305-330). Ames: Iowa State University Press.

42.

McLeod

J. M.

Scheufele

D. A.

Moy

(1999). Community, communication, and participation: The role of mass media and interpersonal discussion in local political participation. Political Communication, 16, 315-336. doi:10.1080/105846099198659

43.

McPherson

Smith-Lovin

Cook

J. M.

(2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444. https://doi.org/10.1146/annurev.soc.27.1.415

44.

Metzger

M. J.

Flanagin

A. J.

Eyal

Lemus

McCann

(2003). Credibility in the 21st century: Integrating perspectives on source, message, and media credibility in the contemporary media environment. Communication Yearbook, 27, 293-335.

45.

Nah

Armstrong

(2011). Structural pluralism in journalism and media studies: A concept explication and theory construction. Mass Communication and Society, 14, 857-878. doi:10.1080/15205436.2011.615446

46.

Peugh

J. L.

Enders

C. K.

(2005). Using the SPSS MIXED procedure to fit cross-sectional and longitudinal multilevel models. Educational and Psychological Measurement, 65, 717-741. doi:10.1177/0013164405278558

47.

Pew Research Center Social & Demographic Trends. (2007, February 22). Americans and social trust: Who, where and why. Retrieved from http://www.pewsocialtrends.org/2007/02/22/americans-and-social-trust-who-where-and-why/

48.

Pinkleton

B. E.

Austin

E. W.

(2001). Individual motivations, perceived media importance, and political disaffection. Political Communication, 18, 321-334. doi:10.1080/10584600152400365

49.

Pratto

Sidanius

Stallworth

L. M.

Malle

B. F.

(1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67, 741-763. doi:10.1037/0022-3514.67.4.741

50.

Putnam

R. D.

(2000). Bowling alone: The collapse and revival of American community. New York, NY: Simon & Shuster.

51.

Robinson

M. J.

Kohut

(1988). Believability and the press. Public Opinion Quarterly, 52, 174-189. doi:10.1086/269093

52.

Saad

(2016, June). Americans’ confidence in newspapers at new low. Gallup. Retrieved from http://www.gallup.com/poll/192665/americans-confidence-newspapers-new-low.aspx

53.

Sampson

R. J.

Wilson

W. J.

(1995). Toward a theory of race, crime, and urban inequality. In Hagan

Peterson

R. D.

(Eds.), Crime and inequality (pp. 37-56). Stanford, CA: Stanford University Press.

54.

Scheufele

D. A.

Hardy

B. W.

Brossard

Waismel-Manor

I. S.

Nisbet

(2006). Democracy based on difference: Examining the links between structural heterogeneity, heterogeneity of discussion networks, and democratic citizenship. Journal of Communication, 56, 728-753. doi:10.1111/j.1460-2466.2006.00317.x

55.

Scheufele

D. A.

Nisbet

M. C.

Brossard

Nisbet

E. C.

(2004). Social structure and citizenship: Examining the impacts of social setting, network heterogeneity, and interpersonal variables on political participation. Political Communication, 21, 315-338. doi:10.1080/10584600490481389

56.

Stroud

N. J.

(2008). Media use and political predispositions: Revisiting the concept of selective exposure. Political Behavior, 30, 341-366. doi:10.1007/s11109-007-9050-9

57.

Sundar

S. S.

(1999). Exploring receivers’ criteria for perception of print and online news. Journalism & Mass Communication Quarterly, 76, 373-386. doi:10.1177/107769909907600213

58.

Tichenor

P. J.

Donohue

G. A.

Olien

C. N.

(1980). Community conflict & the press. Beverly Hills, CA: SAGE.

59.

Tsfati

Ariely

(2014). Individual and contextual correlates of trust in media across 44 countries. Communication Research. doi:10.1177/0093650213485972

60.

U.S. Department of Agriculture. (2017). Rural education. Retrieved from https://www.ers.usda.gov/topics/rural-economy-population/employment-education/rural-education/

61.

Uslaner

E. M.

(2002). The moral foundations of trust. New York, NY: Cambridge University Press.

62.

Wilson

T. C.

(1986). Community population size and social heterogeneity: An empirical test. American Journal of Sociology, 91, 1154-1169.

63.

Yamamoto

Lee

T.-T.

Ran

(2016). Media trust in a community context: A multilevel analysis of individual- and prefecture-level sources of media trust in Japan. Communication Research, 43, 131-154. doi:10.1177/0093650214565894