Abstract
Declining political involvement of adolescents in western society has caused widespread concerns about the health of democracy in the future. This study investigates the role of the media in the formation of political attitudes and political mobilization of adolescents. Based on a secondary data analysis of the European Social Survey (N = 5657), the influence of exposure to news and entertainment content on political trust, signing petitions and consumer politics is assessed in a multi-level regression analysis. Additionally, the impact of the political and educational system on political attitude formation and civic engagement of adolescents is investigated. The results show a higher level of engagement in countries with a well-functioning democracy. At the individual level, news media exposure is positively related to engagement in consumer politics, whereas exposure to entertainment is negatively related to mobilization.
The media, most notably television and the Internet, have a considerable impact on the lives and political development of adolescents. A teenager today spends on average eight hours a day using media (Strasburger et al., 2009). By doing so, he or she potentially learns about current events, becomes acquainted with the actors of the national and international political arena, and is introduced to expert evaluations (Graber, 1997). These are important ingredients of the socialization process towards becoming a political citizen. But this only happens if he or she chooses to do so in the abundance of choice in today’s media landscape (see Prior, 2007). It is the aim of this article to investigate how and to what degree media influence the political socialization of adolescents.
Young adults, their media use, and the potential consequences of the latter on their political preferences and behaviour are frequently studied in the social sciences. However, most studies refrain from studying the phenomenon from a socialization perspective and differences in political attitudes between adolescents and adults are interpreted mostly as cohort effects (Mößner, 2006; Niedermayer, 2006; Patterson, 2007). In fact, the literature devoted to the influence of the media as an agent of political socialization is relatively small. Following the pioneering work by Chaffee and colleagues in the 1970s (see Chaffee and Yang [1990] for an overview of their work) the topic has virtually been absent from the scholarly agenda, with recent exceptions such as Shah et al. (2005). In the meantime the media landscape has changed profoundly and so has our understanding of civic engagement (Ward and de Vreese, 2010) and socialization processes. In contrast to socialization studies conducted in the 1970s, children and adolescents are now conceptualized as active participants in their own political upbringing. From this perspective media influence should not be regarded as an outside force. Media use should better be understood as a conscious choice of adolescents that has an impact on their political attitudes and behaviour (Buckingham, 1997).
Extant research linking media use of young adults and potential effects on political behaviour and attitudes is mostly based on single country studies, thereby ignoring the impact of contextual factors, even though these factors are highly relevant in explaining voting in general (Franklin et al., 1992) and political engagement in particular (Pacheco and Plutzer, 2008). Moreover, most extant studies rely on a relatively limited notion of media use by young people, not taking into account the abundance of choice that today’s media market offers. This is important since choice may lead people to turn away from news and prefer entertainment content (Prior, 2007). Finally, much of our current knowledge rests on a classic but limited conceptualization of civic engagement. Young persons may turn out less at the polls. But, as Norris (2003b) points out, a shift from a citizen-oriented engagement (voting, supporting a political party through voluntary work or donations) to cause-oriented activism, which includes consumer politics and legal as well as illegal demonstrations, could have taken place.
The study investigates the effect of media use on political attitudes and behaviour, and it extends previous research on the citizen-oriented repertoire of civic engagement (voting, party preference) to also include new, cause-oriented forms of participation and a general outlook on politics. The analysis presented in this study is based on data from the European Social Survey, a large dataset collected in 22 to 31 countries that includes measures of media use and political attitudes and behaviour, which provides the unique opportunity to study macro- and micro-level effects at the same time.
Conceptualizing political socialization
Political socialization is a concept that was widely studied and discussed in the 1960s and 1970s by, for example, Almond and Verba (1963) and Sigel and Hoskin (1981). Gimpel et al. (2003: 13) integrated the various conceptualizations and defined political socialization as ‘the process by which new generations are inducted into political culture, learning the knowledge, values, and attitudes that contribute to support of the political system’. Hence, political socialization has an explicit learning component – learning what an individual has to know, to feel and do, to become a citizen that is needed in a particular political system. In the case of most western countries the political system is democracy
Socialization is a complex and continuous process in which cognitions and behavioural patterns are actively learnt through agents, such as parents, school and peers. Though the most important period for socialization is childhood and early adulthood, we keep on negotiating our socialization throughout the course of life as the political world changes and we need to acquire new patterns of behaviour to deal with it (Sears and Levy, 2003).
A new political culture?
Gimpel et al.’s definition of political socialization refers to the term political culture as the object of political socialization, thus the collective attitude towards politics and the expression of those attitudes. However, the political culture of a system is not stable, but changes continuously as the societal parameters are transforming around it. The conceptualization of political culture in western democracies is regularly debated and reassessed. Buckingham (1997) proposes that we are at the dawn of a ‘new political culture’, in which traditional, institutionalized acts of expressions of citizenship (as voting, engagement in political parties) are replaced by involvement in short-term political alliances that are solely created to support a specific interest (e.g. prevent war in Iraq, protect the environment, rally against Wall Street capitalism), and dissolve after the issue is resolved or no longer of political relevance (Norris, 2003b).
This development is important with regard to political socialization research for two reasons: first, it means that the political involvement of an adolescent cannot only be assessed by his or her intention to turn out and vote in an election or proximity to a political party, but also, and maybe even more importantly, by the engagement of a young person in cause-oriented initiatives. Second, engaging in the new political culture can be done at any age and is not institutionally linked to the moment when adolescents legally become adults. Engagement in cause-oriented political activism is not limited by institutional obstacles, like a minimum voting age, that forecloses political involvement of teenagers. Hence, if one wants to research the civic engagement of teenagers before they are eligible to vote, it is a worthwhile considering the cause-oriented forms of political involvement, e.g. the conscious boycott of products for political reasons, or signing petitions, instead of the citizen-oriented repertoire of expressions of citizenship.
Attitudes towards politics
Political mobilization of adolescents is only one aspect of their introduction into the political culture of their country. According to Gimpel et al.’s definition of political socialization, the development of attitudes supporting the current political system is another relevant factor of the phenomenon. Research in political developmental psychology indicates that the process of development of those attitudes begins in a very abstract form with diffuse identification with political actors and institutions during preschool years. Van Deth et al. (2007) show, for example, that German children at the age of six, already have a general idea what a democracy is and are familiar with democratic values like equality before the law. Such perspectives dovetail with recent insights focusing on personality traits (Mondak et al., 2010) and genetics (Hatemi et al., 2010) as antecedents of political ideology and participation.
During adolescence these vague allegiances become more pronounced when children slowly develop a rational approach towards politics (Atkin and Gantz, 1978). Attitudes towards politics do not only crystallize and become more concrete during adolescence, they also become increasingly independent from their parents’ political attitudes and preferences (see Chaffee and Yang, 1990). Moreover, political preferences developed during adolescence remain mostly stable over the course of life (Sears and Levy, 2003). The object of these attitudes can be specific political figures, institutions, or policies, but it can also be the political system itself. In this article we investigate both: whether an individual reports proximity to a political party (as one of the main actors in a democracy), and the trust an adolescent has in the various institutions that constitute the political system.
Media and political socialization
The development of political attitudes as well as political behaviour is influenced by several agents, one of them being the media (see McLeod et al., 2010). According to political socialization theory, political attitudes are formed during childhood and adolescence and only to a small degree during adulthood (Huntemann and Morgan, 2001). Media effects research has also emphasized that investigating attitude formation processes at young ages is advantageous. It has been coherently shown that the prevalent effect of media exposure is to reinforce pre-existing attitudes. This is due to self-selection mechanisms: media users prefer selecting outlets that are likely to be consistent with their preferences to avoid psychological stress (Slater, 2007).
However, most empirical studies in the field of media effects research focus on mass media influence on the political attitudes of adults, who are already disposed to a developed set of political preferences. Contrary to this approach, this study investigates the influence of the media on adolescents.
Two causal mechanisms
To understand the impact of the media on political attitudes and behaviour, it is necessary to consider the causal processes that link media exposure and political attitude formation. Two theoretical approaches can help to understand these mechanisms: learning theory and media malaise. Whereas learning theory points to a positive impact of (news-) media consumption on political socialization, media malaise theorists claim that the media contribute to rising levels of political cynicism.
Learning theory
Media are the primary source of political information on current political events for adolescents (Graber, 1997). For adolescents and adults alike, it is virtually impossible to witness all relevant political events in person, let alone discuss their implications with all concerned parties to place them into an interpretative framework. Political information is therefore evidently second-hand information, and the media are the most accessible and commonly used source for it. In that light, it cannot surprise that Dalton (1996) and Inglehart (1990) both find a strong correlation between the availability of political information through mass media and the overall level of political knowledge on the macro level (see also Curran et al., 2009). They claim that the expanded availability of media, especially in the second half of the 20th century, has led to an increase in the spread of political knowledge.
Exposure to political information increases in turn the ability of individuals to engage in politics (Delli Carpini, 2000), as news provides information about current events that asks for participation and clarifies the processes in a democratic society (see also Shah et al., 2009). Accordingly, ceteris paribus, we propose the following hypothesis.
H1: The higher the exposure to news media, the higher the civic engagement.
Media malaise
Contrasting claims of learning theory, advocates of media malaise expect a negative influence of the mass media on political interest and engagement. The core notion of this approach is that media coverage is inherently negative about politics. This is due to higher news value of negative news items (Bennett et al., 2007; Galtung and Ruge, 1965). As a common sense journalistic doctrine puts it: ‘bad news is good news’. News items that reveal scandals in the political world, criticize the performance of those in power or alert to unfavourable consequences of particular decisions, appeal more to the interest of the audience than news stories that inform about the success of particular policies or the number of politicians that are scandal-free. 1 Positive information about politics is considered to be less relevant or interesting. This brings about a negative bias in the portrayal of politics in the news media (Kepplinger, 1998), which ultimately gives the audience the impression that politicians act irresponsibly most of the time and therefore cannot be trusted (Cappella and Jamieson, 1997).
Interestingly, media malaise theory has not been empirically substantiated by investigating the effects of exposure to news media, but for general media use (e.g. Putnam, 2000). Exposure to news media, on the other hand, has been shown to have a positive influence on political interest (Strömbäck and Shehata, 2010) and political involvement (Newton, 1999; Norris, 2003a). Holtz-Bacha (1990) stated explicitly that higher levels of political cynicism are not associated with exposure to political information in the mass media but entertainment media. The negative relationship of entertainment media use and political involvement can be explained by potential cultivating effects of mass media, first conceptualized by Gerbner et al. (1986). According to the cultivation approach, the media usually present a violent world in which danger is omnipresent. The audience, confronted with this biased image of reality, is left with a feeling of discomfort and insecurity, which results in a general alienation from the world, especially among those people who spend most of their leisure time using the media. The feeling of insecurity is paired with a feeling of inefficacy, apathy and distrust (Putnam, 2000) towards society, in particular the political system.
General alienation from the political world also leads to growing levels of political cynicism and a decline in civic engagement (Putnam, 2000). This means that exposure to entertainment media can have demobilizing effects. Accordingly, we can formulate the following hypothesis:
H2: The higher the exposure to entertainment media, the lower the level of political trust and civic engagement.
Influence of system-level variables
Sigel and Hoskin (1981) introduced the idea that the political system can affect processes of political socialization. They claim that the more smoothly a democratic system functions the less adolescents engage in politics. The rationale behind this idea is that adolescents – occupied with their education, social and romantic endeavours and important choices concerning their professional future – prefer to focus their attention on aspects of their lives other than their political environment. If they feel that there is little need for their political input it is a rational choice to save scarce time and other resources. In other words, an adolescent only becomes Aristotle’s famous political animal when his or her life is significantly blemished by his or her political environment.
The quality of democracy varies greatly among countries in Europe. Although the polity structure of all European countries is democratic, there are considerable differences in the degree to which they are able to achieve democratic goals: high levels of public health, education and wealth (Campbell, 2008):
H3: The higher the quality of democracy in the country in which adolescents live, the lower the level of civic engagement.
Methods
To understand the influence of the media on political socialization we conducted an analysis of the first three rounds of the European Social Survey (ESS), carried out in 2002, 2004, 2006, 2008 and 2010 respectively. The ESS is a general population survey carried out in 22 to 31 countries among a representative sample of approximately 2000 respondents per round and country. 2 The countries included in this analysis are Austria, Bulgaria, Belgium, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Netherlands, Norway, Poland, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, Ukraine and UK. We selected all respondents between the ages of 14 and 20 as the bases for our analysis. 3 Since we did not find significant differences in the dependent variables between the rounds, all data are combined for further analysis (N = 6935). We investigate two types of dependent variables in this study: political attitudes and civic engagement.
Dependent variables
Attitudes towards politics
Party affiliation: Proximity to a political party was assessed by asking the respondents whether they felt closer to a particular party compared to all other political parties.
Political cynicism: In this study political cynicism is operationalized as the opposite of political trust. Though these concepts are not equivalent, political trust has been used as a measure of media malaise in comparable studies (Avery, 2009) and is therefore chosen as an indicator in this study. Political trust is measured by a scale constructed of five items that individually measure the respondents’ trust in the national parliament, the legal system, the police, politicians and political parties (α: .869). All items were measured on an 11-point scale.
Civic engagement
Political involvement of adolescents is measured by two indicators of cause-oriented activism, as voting or other forms of citizen-oriented engagement is not relevant for the age group. The indicators are signing a petition in the past 12 months, and boycotting a product for political reasons in the past 12 months. 4
Independent variables
Media exposure
News exposure is operationalized as the number of hours a respondent spends on an average weekday reading newspapers, watching television news, or listening to political information on the radio. As each of those channels was assessed independently, they are summed to assess the cumulative amount of time spent with news media each day. Entertainment exposure is assessed in a comparable manner. The ESS does not include measures of entertainment exposure, but it has a measure of general media exposure. To calculate entertainment exposure, news exposure is subtracted from the number of hours spent with a particular medium in total, on an average weekday. Subsequently all of the individual entertainment scores are summed to assess the cumulative amount of time spent with entertainment media every day. This means that the measure for media entertainment also includes advertising or educational formats. Both measures are weakly correlated (r = –.20, p < .001). 5
Quality of democracy
Quality of democracy is assessed using the quality of democracy ranking (Campbell, 2008). In this index a number of dimensions of functioning democracies are combined (political freedom, economic performance, health system, gender equality, general political knowledge, environmental performance, corruption). The empirical bases for the index are reliable and established sources such as: the Freedom House Index of Political Freedom; World Bank; and Political Handbook of the World. We chose this measure as it is able to grasp differences in the functioning of mature democracies in Europe, that most other measures of democracy (Freedom House, Vanhanen’s Index of Democracy) cannot assess.
Control variables
To specify the model and avoid overestimation of the effects, a number of control variables were added that have been positively tested to be of empirical relevance in previous research (e.g. Esser and de Vreese, 2007; Gimpel et al., 2003; Norris, 2003b; Sigel and Hoskin, 1981). Political interest 6 was measured on a four-point scale as a response to the question: ‘How interested are you in politics?’; educational level was measured in years of education; 7 ideology of the respondents was assessed by a self-reported placement of the respondent on an 11-point left–right scale (10 = right). Finally age and gender were added as control variables.
Unfortunately, the setup of the ESS as a survey which is mostly directed at adults means that there are a number of limitations in the data. This particularly holds for the control variables. Important control variables like household income could not be included in the model as the vast majority of the sample of adolescents did not know the household income.
To provide a reference point for the results as called for by Roller et al. (2006), we compare the results concerning the political attitudes and behaviour of adolescents with the remainder of the sample.
The variables are described in the Appendix.
Results
Political attitudes and engagement change during the course of life. This is a result of both age and cohort effects. Figure 1 shows to what degree respondents of different ages are politically involved. As it becomes clear from Figure 1, the development varies between different modes of political involvement. The affiliation with a political party, for example, slowly increases with age, whereas participation in consumer politics peaks during adulthood. Taking a closer look at the specific development of adolescents, it becomes clear that adolescence is a crucial period for political socialization: the share of adolescents feeling close to a particular party steeply increases, whereas political trust declines considerably. At the same time news media are increasingly becoming part of the media menu.
Mean values of news exposure, attitudes towards politics and civic engagement by age.
Explanatory results
To assess the influence of the media on political attitudes and behaviour of adolescents multi-level regression analysis explaining each of the independent variables was applied. Logistic regression was applied in the analysis of political engagement. Political trust was analysed by applying linear MLE (maximum likelihood estimation) regression. To account for country differences a random intercept was added. To provide a reference point for the results, each model is also calculated for respondents older than 21. This way, we can assess whether a certain finding is particular for adolescents or coherent with a general trend.
Tables 1 and 2 present the findings of the analysis. The results reveal that, depending on the dimension of political socialization under investigation – behaviour or attitude – a different set of factors is significant for the explanation of its variance. This observation becomes particularly clear when looking at the influence of the two types of media exposure: exposure to news media and exposure to entertainment media. In most cases, the effects of news exposure and entertainment exposure do not coincide, but are complementary.
Fixed effects estimates for models of the predictors of political attitudes.
Note: Standard errors are in parentheses.
p < .05, ** p < .01.
Fixed effects estimates for models of the predictors of civic engagement.
Note: Standard errors are in parentheses.
p < .05, ** p < .01.
Media effects
Hypothesis 1 states a positive influence of news media exposure on civic engagement, whereas Hypothesis 2 postulates a negative influence of entertainment media exposure on both types of dependent variables: civic engagement and attitudes towards politics. The results indicate that the direction of the effects is indeed in line with the hypotheses in most cases, however not all of the effects are significant. Exposure to news only has a statistically significant effect on the likelihood to boycott a product for political reasons or sign a petition, but not for adolescents. These effects were only significant for older respondents, which means that Hypothesis 1 is not supported. The negative effect of other media exposure on political attitudes and behaviour (H2), on the other hand, can be confirmed for adolescents with regard to participating in politics as well as trusting politicians.
The overall tendency of the variables can also be deducted from a visual examination of the predicted scores depending on a low, high or medium level of media exposure (Figure 2).
Predicted scores of low, medium and high media exposure in a logistic multi-level regression. All other variables are set to mean.
When comparing the younger cohort to the age 34–40 cohort, it becomes obvious that the mobilizing influence of news media is higher in the older cohort. While we do not find significant effects of exposure to news on consumer politics or signing a petition in the younger cohort, they are evident for the older cohort. This observation could be substantiated using moderation analysis, in which we found significant interaction effects consisting of age and news exposure (β: sign petition: .001, p < .01, β: consumer politics: .001, p < .01). We also found a significant interaction effect of age and entertainment media use in the model explaining political trust (β: –.001, p < .01). This means that spending more time with entertainment media leads to higher levels of distrust among the older generation. In the case of proximity to a political party, the opposite was the case. Here, the youngest respondents show the strongest effects (β: .005, p < .01).
On the system level we find that citizens who live in countries with a lower quality of democracy generally show a lower level of political activism, which means that Hypothesis 3 cannot be supported. On the contrary, we find significance in the opposite direction in most cases, showing that stable democracies are more likely to bring about a politically active youth.
Discussion
In this article we addressed two questions concerning the relationship between media exposure and political socialization. The first question relates to the contradicting effects of entertainment exposure and news exposure. Whereas in previous research the latter has been proven to increase political knowledge and – in the long run – democratic participation of young adults, the first is presumed to contribute to media malaise, a sentiment of distrust in politics and politicians, nurtured by media products that present a violent, dangerous world.
Our results suggest that we can find both mobilizing and demobilizing effects of the media depending on the content an individual is exposed to. Use of entertainment content contributes to a feeling of distrust in politics and leads to lower levels of political engagement among adolescents, yet exposure to news media stimulates higher political participation, but only among those older than 21.
The second question we posed in this study concerns the impact on civic engagement. Our data suggest that there is a significant effect of quality of democracy on political attitudes and behaviour. However, the direction of the predicted effects was not in the expected direction. Based on the work of Sigel and Hoskin (1981), we expected young citizens to be less involved in mature democracies. Yet, our results indicate that in nations where democracy is not functioning at its best – and not the mature democracies – young citizens are left with little interest to engage. This may be because they feel that signing a petition or boycotting a product will not change anything as the political system is less responsive to the pressure of public opinion.
To test system-level influence we used data from the European Social Survey in which survey data from 22 to 31 European countries are gathered. Although the size of the dataset and quality of data collection in the European Social Survey imply that our results are of high reliability, the data were not collected directly for the purpose of this study which means that the validity of some of our measures should be discussed. First, the indicator of entertainment media use is rather crude. As it is an indirect measure calculated on general media use and news media use, it includes more than just entertainment media use, for example advertising or educative media content, and does not account for infotainment or other hybrid forms of political and entertainment content. Moreover, both indicators of media use are measures of exposure and not attentive media use as media malaise theory would suggest. Second, we could not include online political participation and media use. This means that this study does not account for new forms of political participation, for example through social network sites, which are especially relevant for the young generation. However, since online and offline participation are correlated (Bakker and de Vreese, 2011), explaining offline political participation is still worthwhile as it helps to identify sources of an active approach towards politics. Finally, measures of education and ideology are not assessed in an age-appropriate fashion with regard to adolescents. This means that our results should be interpreted with caution, acknowledging the limitations of the data.
To eventually fully identify the causal mechanism that relates media exposure and political socialization, research on the effects of the media on political socialization should move beyond a cross-sectional research design. Taking a longitudinal approach would allow us to address some interesting questions that are still open: How does the influence of the media develop over time? Do media mobilize adolescents or do mobilized adolescents select media that mobilize them even more? At what age do adolescents develop stable attitudes? This would also contribute to political development psychology with regard to media effects. To date, the timing of media effects during the life course is uncharted territory.
Our results showed that political trust declines steeply during adolescence. According to our analysis the growing lack of confidence is fuelled by use of non-informational media. Exposure to entertainment media content also has a significant negative effect on political participation. Adolescents and adults alike become less inclined to sign political petitions or boycott a product for political reasons as they spend more time using entertainment media. We interpret this finding in the light of media malaise theory implying that it is violent or negative media content that alienates its audience from the political world. It could also be argued that the relationship between entertainment media use and lack of participation is due to limitations in the time budget (Putnam, 2000). Time spent using media individually cannot be spent on political activities. However, two observations of our study cast doubt on this alternative explanation. First, the size of the negative effect remains the same throughout the course of life; though time spent with entertainment media decreases on average. Second, we find a significant negative influence of entertainment media use on political trust. This means that the attitude towards politics is also affected by entertainment media use.
When comparing media influence on political participation of adults and adolescents, we find that the effects are more pronounced in the older cohort. Adolescence appears to be the period in life during which media malaise and political learning through media begins, but the effects of the media grow stronger over the life course. This is interesting to note, because adolescence is also the period in life when exposure to media is highest. Hence, one would expect that the media effects are the strongest. Yet, the opposite is the case.
There are two possible explanations for this finding. First, the effects of media use unfold over time, implying that the cumulative amount of time spent with the media over the course of life, and not the frequency of news media use, encourages political participation. An alternative explanation is that media become a more important source of influence relative to other sources, as other agents of socialization (parents, school) become less influential during adulthood. Whichever explanation holds, both illustrate that political learning theory should be further developed to incorporate life cycle effects. It might very well be that the relationship between media use and political participation depends on age in a non-linear fashion.
The difference in the effect of news media between adolescents and adults can be interpreted as both, a cohort and an age effect. We suggest that cumulative media consumption during the life course affects political engagement. However, this does not mean that there will not be generational differences as well. In fact, literature in the field of political engagement research clearly suggests that one’s generation or cohort is crucial in determining political participation, for example if there was high political conflict during the first election experienced by a generation (Sears and Valentino, 1997). However, to disentangle age and cohort effects we have to wait for more waves of the European Social Survey.
By applying multi-level analysis we could show that political socialization is influenced by factors on the macro level and factors on the micro level. Quality of democracy turned out to be very influential in bringing about active citizens in Europe. The higher the level of public health, wealth or education, the more young citizens feel inspired to take on an active role in society. This means that in order to socialize a participating citizenry, one place to start is by providing a functioning political environment.
Footnotes
Appendix
Descriptives of the dependent and independent variables
| 14–20 |
21+ |
|||
|---|---|---|---|---|
| Mean | SD | Mean | SD | |
| Political trust | 4.95 | .02 | 4.66 | .007 |
| Party affiliation | .34 | .005 | .55 | .001 |
| Signed petition | .22 | .005 | .25 | .001 |
| Boycotted product | .11 | .003 | .17 | .001 |
| Age | 17.83 | .01 | 49.90 | .050 |
| Gender (1 = female) | .51 | .006 | .50 | .001 |
| Political interest | 2.15 | .009 | 2.56 | .002 |
| Education | 11.02 | .024 | 12.72 | .011 |
| Ideology | 5.05 | .02 | 5.144 | .006 |
| News exposure | 3.12 | .022 | 5.03 | .008 |
| Entertainment exp | 5.76 | .038 | 5.14 | .009 |
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.