Abstract
The present study investigated whether implicit and explicit attitudes predict news choice among Muslims with migration backgrounds living in Germany. We used both attitude constructs to predict a selection bias for news about the same event stemming from the host country (Germany) vs. from other countries. Using a survey (N = 1,107), we found that favorable implicit and explicit attitudes toward Germany increased a participant’s tendency to select German news. Each attitudinal construct predicted a unique variance in news choice. Using a subsample of Turkish citizens living in Germany who participated in the Turkish constitutional referendum 2017 (N = 241), we found that the attitude-based selection bias predicted their voting. We discuss implications for selective-exposure research and processes of integration.
Introduction
News-consumption behavior is determined by several selection decisions (Knobloch-Westerwick, 2015). News consumers demonstrate preference and avoidance patterns because it is impossible to allocate attention to all of the available news content (Zillmann and Bryant, 1985). Importantly for the present study, recent research emphasizes that news choice is determined by both automatic and deliberate mental processes: Empirical evidence is consistent with the notion that automatic affective evaluations (i. e., implicit attitudes) as well as overtly-expressed evaluations based on conscious reasoning (i. e., explicit attitudes) toward the media brand (Arendt, Northup, and Camaj, 2019) and the news content (Arendt, Steindl, and Kümpel, 2016) predict news choice.
The present study extends this line of selective-exposure research by testing the power of implicit and explicit attitudes toward Germany to predict news choice in a sample of Muslims with migration backgrounds living in Germany. As the primary contribution to the literature, we predicted news choice between news items about the same event stemming from the host country (Germany) versus from other countries (country of origin). We assumed that nation-related attitudes, in particular, can be conceptualized as automatically activated, affective “gut-level” responses that influence news choice in addition to more reasoned evaluations based on conscious thought.
As a supplement to the present study’s contribution to selective exposure theory, the it contributes to our knowledge of integration processes. Processes of Muslim integration have been widely discussed in recent years, and it has been emphasized that proper knowledge of the factors contributing to integration-related outcomes is important (Halm and Sauer, 2017). Media use is one of those factors. As Geißler and Weber-Menges (2013) put it, consumption of a host country’s media is “absolutely indispensable” (p. 34). This argument is based on the fact that individuals with migration backgrounds cannot realize their full potential in their host country without a proper knowledge of, broadly speaking, what is going on in the host country (see also Arnold and Schneider, 2007; Bonfadelli, Bucher, and Piga, 2007). As a notable supplement related to this aspect, we tested possible behavioral consequences of an attitude-based selection bias for the actual voting decisions of Turkish citizens living in Germany during the Turkish constitutional referendum in 2017.
In the present paper, we first provide a review of research on selective exposure with a special focus on the role of implicit and explicit attitudes. Afterwards, we relate media use to processes of integration. Based on this amalgamation, we develop two hypotheses which we tested by using data from a survey of Muslims with migration backgrounds living in Germany. Finally, we discuss our findings in terms of selective exposure theory and processes of integration.
Attitude-based selective exposure
According to Knobloch-Westerwick (2015), selective exposure can be defined as “any systematic bias in selected messages that diverge from the composition of accessible messages” (p. 3). Previous research has revealed that there are two fundamental, general motives responsible for news choice (Hart et al., 2009): First, individuals are motivated to expose themselves to accurate information. Accurate information helps them to adapt to and cope with the environment. Second, individuals are motivated to avoid emotional discomfort arising from exposure to dissonant information (Festinger, 1957). This second fundamental motive – often called confirmation bias – points to the fact that individuals tend to select information (e. g., political arguments) that is congruent with their attitudes. Importantly, a confirmation bias can have severe implications for democracy because listening to “the other side” is indispensable for democracy’s ideal of a marketplace of ideas (Knobloch-Westerwick, 2015). Research has accumulated supporting empirical evidence for attitude-based selective exposure across print (Noelle-Neumann, 1973) and broadcast news (Iyengar and Hahn, 2009). Recent research has contributed to this line of research by investigating related phenomena in modern digital communication environments, termed “filter bubbles” (Pariser, 2011) or “echo chambers” (Sunstein, 2009).
As a supplement to previous research that has relied almost exclusively on overtly-expressed evaluations based on more or less deliberate thought (i. e., explicit attitudes), recent research on selective exposure has hypothesized that automatically activated evaluations (i. e., implicit attitudes) can also predict selective exposure (Arendt et al., 2016, 2019; Galdi, Gawronski, Arcuri, and Friese, 2012). This assumption was guided by accumulating evidence from research on social cognition showing that both (more or less) controlled, verbalizable evaluative thoughts and impulsively activated, spontaneous “gut-level” reactions predict human decision-making (e. g., Gawronski and Bodenhausen, 2006; Greenwald et al., 2002; Olson and Fazio, 2009).
Applied to media (Arendt et al., 2016), recent theorizing assumes the following process: In a first step, exposure to news items (e. g., words, pictures) automatically (re-)activates their corresponding internal representations in the news-consumer’s memory. This process relies on a brain architecture that has developed and adapted throughout our evolutionary past (Buss, 2009), and is thus likely to be rooted in phylogenetically ancient mechanisms (Mahajan et al., 2011). Visuals (e. g., an angry or sad face) or text-based stimuli (e. g., affect-laden words such as “gun”) may automatically activate affective responses within a few hundred milliseconds (Lodge and Taber, 2013). Such automatic affective reactions are termed “implicit attitudes” (Gawronski and Bodenhausen, 2006). Previous research has shown that salient news-media brands (e. g., Fox News, see Arendt et al., 2019) as well as news content (e. g., words related to the European Union, see Arendt et al., 2016) can automatically activate affective responses, helping the human information-processing system to build a first “quick and dirty” evaluation of news items (see also Galdi et al., 2012).
In a second step, however, humans are able to assess automatically activated “gut responses” (i. e., automatically activated affective reactions) and consciously ruminate about the “deeper meaning” of a given news item. However, this assessment—often termed “reasoning”—depends on time, cognitive resources, introspective access, and the motivation to ruminate about a news item (see Gawronski and Bodenhausen, 2006). Importantly, even if automatic processing tends to favor specific news items—a person with negative implicit attitudes toward Fox News may impulsively tend to avoid Fox News news items—, reasoned thought can overrule this automatically activated behavioral tendency. For example, a liberal person may think about the assumption that exposure to opposing arguments is indispensable for democracy’s ideal of a marketplace of ideas: “I want to know the arguments made by conservatives—I will watch this Fox News news item.” Thus, there can be situations in which both attitudinal constructs point in the same direction, and other situations where automatic processing and reasoning imply different selections.
Previous research on selective exposure used a gate metaphor to theorize on the selection process based on implicit and explicit attitudes (Arendt et al., 2016). It has been assumed that one door of the gate is operated by implicit attitudes and the other is operated by explicit attitudes. The assumption is that both attitudinal constructs can open or close the gate. We will return to this theorizing on which we base the hypotheses after we have discussed the present study’s research in the societal context.
The relationship between media consumption and integration
Media use by individuals with migration backgrounds[1] has often been linked to processes of integration (Arnold and Schneider, 2007; Peeters and d’Haenens, 2005; Sauer 2010; Trebbe, 2007). Based on recent theorizing on selective exposure and media effects (Knobloch-Westerwick, 2015; Slater, 2007; Valkenburg and Peter, 2013), we conceptualize a reciprocal relationship between news consumption and integration, as characteristics of individuals (e. g., attitudinal predispositions) are predictors of selective exposure, and exposure to selected news content in turn elicits effects on important outcomes such as political attitudes or behavior. Thus, the reciprocal multi-causality of a media user’s motivation in terms of media use, selective media consumption, and media effects should be observable over time (Knobloch-Westerwick, 2015). Therefore, individuals with migration backgrounds may choose German news media because of attitudinal predispositions that they have developed over time, partly based on prior exposure to German (or other countries’) news media. This in turn may have consequences for important outcomes related to integration.
A general assumption is that exposure to a host country’s media provides information about a host country’s history, traditions, current issues, norms, values, and even humor (Geißler and Weber-Menges, 2013; Kim, 1988). In addition, the consumption of the host country’s media can serve as “orientation” and provide knowledge about opportunities for participation, for example, in politics, the labor market, and other social structures in the host society (Geißler and Weber-Menges, 2013). Especially the role of news media has been noted, leading to beneficial outcomes such as a high level of political knowledge (Chaffee, Nass, and Yang, 1990; Liu and Gastil, 2014). In fact, host country news-media consumption can be deemed desirable because the media transmit the host society’s information and values and therefore facilitate integration (Bonfadelli et al., 2007). Thus, the present study focuses on predictors of news choice in favor of the host country’s news items.
We focused on the question of whether or not news consumption and integration are linked together among Muslims living in Germany. Muslims make up around 6 % of Germany’s population, and this number is expected to rise in the future (Stichs, 2016). Even without further immigration, the number of Muslims in Germany is likely to increase because Muslims living in Europe are younger than the non-Muslim population and furthermore tend to have more children (Pew Research Center, 2017). Although Muslims in Germany come from different countries, individuals from Turkey represent a substantial segment of the Muslim population (Stichs, 2016).
Many Muslims in Germany speak their country-of-origin language (Haug, Müssig, and Stichs, 2009), enabling them to consume news from both the host country and/or their country of origin (Gerhard and Beisch, 2011). In fact, a recent survey of individuals with Turkish migration backgrounds found that 29 % of them only watched Turkish television, 48 % watched Turkish and German television, and 13 % only watched German television (Gerhard and Beisch, 2011). Yet, consuming country-of-origin media does not necessarily mean a lack of integration or even separation among Muslims living in Germany. Studies have shown that even well-integrated Muslims in Germany still use their country-of-origin media (Sauer, 2010; Trebbe and Weiß, 2007), which fits with the central idea of current conceptualizations of integration that assume that individuals with migration backgrounds can still maintain their home-country culture while adapting to a new one (Berry, 1997). Foreign media use only seems to be an issue when consumed exclusively or if the content has a disintegrative character (Geißler and Weber-Menges, 2013; Müller, 2009).
Predicting news choice
Previous research has revealed that factors such as education level, language ability, the media hierarchy in foreign countries (e. g., newspaper circulation is high in Germany, but newspaper consumption plays a less important role in many Southern European countries), accessibility to foreign media in the host country, and the length of time living in the host country all may contribute to news choice among individuals with migration backgrounds (see Bonfadelli and Moser, 2006). As a supplement to these factors and based on the theorizing presented above, we assumed that implicit and explicit attitudes would also influence news choice. Importantly, research investigating processes of integration have relied heavily on concepts such as national identity or attitudes toward the host country (Kunst, Tajamal, Sam, and Ulleberg, 2012; Sauer, 2010; Trebbe, 2007; Trebbe and Weiß, 2007). Based on this focus on nation-related concepts, we hypothesized that implicit and explicit attitudes toward Germany would predict news choice in favor of German news items (relative to news items from other countries) in Muslims with migration backgrounds living in Germany. This is the primary hypothesis of the present study.
Hypothesis 1: Favorable implicit and explicit attitudes toward Germany predict news choice in favor of German news items relative to foreign news items.
Behavioral consequences of an attitude-based selection bias
In Germany, the largest ethnic minority group is originally from a Muslim country: Turkey (Bundesamt für Migration und Flüchtlinge, 2017). Importantly, data collection for the study’s survey—as a part of a larger project on the role of the media in the life of Muslims living in Germany—was conducted in the aftermath of the 2017 constitutional referendum in Turkey. This enabled us to further test whether an attitude-based selection bias predicted the voting decision in the constitutional referendum among members of the largest ethnic minority group.
Constitutional referendum in Turkey
On April 16, 2017, Turkish citizens voted in a referendum on changes to the Turkish constitution. Some of the changes that were voted on included the abolition of the post of prime minister and a transfer of the executive power to the president, the transfer of some of the parliament’s key oversight to the presidency, and the empowerment of the president to appoint some high-level positions in the judiciary (OSCE, 2017). Opponents (who preferred a “No” vote, i. e., who were against the changes that had to be voted on) mostly feared a diminishment of democracy. Supporters (who preferred a “Yes” vote) mostly argued that the constitutional changes would lead to more stability and economic power via a centralized government.
The results showed support for constitutional change, though the results were narrow (Yüksek Seçim Kurulu [Supreme Board of Elections], 2017): In sum, 51.24 % voted “Yes” while 48.76 % voted “No”. Since Turkish citizens abroad were also allowed to vote, Turkish citizens living in Germany contributed to this narrow result: 1.43 million were eligible to vote, and around 660,000 made use of their vote. Interestingly, 63.07 % of the Turkish citizens living in Germany voted for “Yes”, a larger share than in Turkey, leading to a discussion as to why Turkish citizens living in Germany, in particular, voted strongly for “Yes”. This discussion was accompanied by a fear related to the consequences for Turkish (and sometimes even German) democracy. For example, one German social scientist even noted that a change would mean the “end of democracy” (Deutschlandfunk, 2017).
As already noted most Turkish citizens living in Germany can follow both German and Turkish news coverage (Gerhard and Beisch, 2011; Sauer, 2010). We assumed that Turkish and German news media might have reported very differently about the referendum. This working hypothesis was guided by previous research showing that news from different national contexts tends to differ when framing major news events, leading to different perspectives on politically relevant events (Nygren et al., 2016; Roman, Wanta, and Buniak, 2017). Research on the Eastern Ukraine military conflict can be used to illustrate this phenomenon. Nygren and colleagues (2016) noted that the “framing of the conflict [was] closely related to the political situation in each country” (p. 15). While Ukrainian media featured more articles supporting the Ukraine’s point of view, Russian media did the same for their country. Furthermore, the researchers found considerable differences in terms of the framing of pro-Russian individuals (Ukraine media: Separatists or Terrorists vs. Russian media: home guards, self-defenders). Roman and colleagues (2017) also found that “the coverage of the conflict was highly representative of the governments’ stance on the conflict” (p. 373).
A content analysis of Turkish television news revealed a reporting imbalance (OSCE, 2017). Analyses show that all five major television channels reported more heavily about (and supportive of) the “Yes” campaign. Only Fox TV and CNN Türk presented some voices critical of the “Yes” campaign, whereas Haber TRT1 and Show TV aired exclusively supportive views (OSCE, 2017). As Turkish television is present in almost every Turkish household in Germany (Gerhard and Beisch, 2011), it is also likely that Turkish citizens living in Germany consumed large numbers of those news reports. Even though content analyses of German news media regarding the referendum do not yet exist, we assumed that German news media did not support a “Yes” vote. This assumption was largely based on political developments that took place in the weeks before the referendum. German states blocked Turkish politicians in Germany from holding rallies that intended to promote the “Yes” vote. In turn, the Turkish President Recep Tayip Erdoğan, a leading supporter of the “Yes” vote, accused Germany of Nazi practices. This led to a wave of indignation among German politicians and the news media.
Consequences on voting
We expected an attitude-based selection bias to predict the voting decision: Previous studies have shown that exposure to news that is heavily supportive of a specific political standpoint can change news-consumers’ voting decisions (e. g., DellaVigna and Kaplan, 1997; Druckman and Parkin, 2005). Furthermore, selective-exposure research has shown that attitude-based selective exposure can have important consequences for political outcomes (Knobloch-Westerwick, 2015; Knobloch-Westerwick and Johnson, 2014). Based on these two discoveries, we assumed that Turkish citizens who thought about how to vote in the constitutional referendum may have built different voting decisions based on the news they cumulatively and consistently consumed. Thus, a selection bias in favor of German or Turkish news can be equated with exposure to different perspectives on the referendum (i. e., no or substantial support for a “Yes” vote). Accordingly, we developed the idea that favorable implicit and explicit attitudes toward Germany would predict news choice in favor of German news items among Turkish citizens living in Germany (see H1). Importantly, we additionally hypothesized that favoring German news would predict an increased tendency to vote “No” in the Turkish referendum. This idea can be represented in a mediator model: Implicit and explicit attitudes are conceptualized as the focal predictors (independent variables), news choice in favor of German news as the mediator variable, and voting decision (i. e., “Yes” or “No”) as the dependent variable. A visual depiction of the model will be provided in the results section.
Hypothesis 2: For Turkish citizens living in Germany, implicit and explicit attitudes toward Germany predict news choice, which in turn predicts their voting decision in the referendum.
Method
Muslims with migration backgrounds living in Germany participated in a web-based study that measured attitudes toward Germany, news choice, voting decision, and controls. Only those participants who self-identified as Muslims and had sufficient German-language skills were able to take part in the survey. We did not differentiate between asylum seekers, individuals who immigrated for other reasons, or individuals whose parents immigrated to Germany.
Participants
Recruiting process. Muslims living in Germany were recruited via several channels: Two commercial market research institutes recruited participants via their online access panels (n = 735 and n = 130). A third commercial market research institute collected the e-mail addresses of Muslims via telephone interviews (n = 8). Some participants were recruited via a non-commercial online access panel (n = 21). Moreover, three Muslim students recruited participants from their social environments (n = 155). Furthermore, we advertised the study on an online platform that connects individuals with Turkish migration backgrounds living in Germany (n = 11). Finally, we asked representatives of mosques and other Muslim communities in Germany to distribute the study to their members (n = 47). Due to these diverse channels, which were required to achieve a large sample size, we were unable to calculate the response rate.
Sample. A total of 1,107 individuals participated in the study. Of these participants, 65.4 % were female. The sample ranged in age between 18 and 78 (M = 28.56, SD = 11.48). A total of 28.5 % indicated having no high school diploma, 51.7 % indicated having a high school diploma, and 19.9 % indicated having a university degree. Approximately half of the sample (56.5 %) were German citizens, 53.0 % were Turkish citizens, and 10.3 % were citizens from other nations (dual citizenships were possible).
For the test of H2, a subsampleof Turkish citizens was used (N = 241). This sample consisted of (1) Turkish citizens living in Germany who (2) voted in the constitutional referendum and (3) told us their voting decision. This subsample ranged in age between 18 and 75 (M = 27.00, SD = 11.00) and 65.1 % were female. A total of 25.7 % indicated having no high school diploma, 63.5 % indicated having a high school diploma, and 10.8 % indicated having a university degree. Approximately one third of the sample (35.7 %) were German citizens.
Primary variables
Explicit attitudes. We asked about the participant’s “attitudes toward Germany”. We noted that we were interested in their “general feelings”. Participants could answer this question on a 7-point bipolar scale ranging from 1 to 7 using five items (“I think Germany is …”: good–bad, positive–negative, beneficial–harmful, fair–unfair, wise–foolish). Higher values on this measure indicate more positive attitudes (M = 5.09, SD = 1.18, α = 0.83).
Implicit attitudes. After measuring explicit attitudes, we measured implicit attitudes using the affect misattribution procedure (Payne, Cheng, Govorun, and Stewart, 2005). This procedure, relying on affective priming methodology, assessed automatic affective “gut-level” responses toward Germany. In each trial, participants were presented with a German or a foreign flag. The flag was presented for 80 ms and was assumed to prime the nation concept. We used flags from Muslim countries for the foreign category (e. g., Turkey, Saudi Arabia, and Tunisia). Immediately after the flag pictures disappeared from the screen, a Chinese character (i. e., ambiguous target symbol) was presented for 250 ms. After the presentation of the Chinese character, individuals were asked to rate the “visual pleasantness” of the Chinese character. Participants could rate the Chinese character as “unpleasant” or “pleasant”. According to the affective priming paradigm, individuals should evaluate a Chinese character more favorably when they have been primed with a flag that elicits a more positive automatic affective reaction (Payne et al., 2005). A total of 40 trials were undertaken (i. e., 20 trials with German flags and 20 trials with foreign flags). We calculated the relative frequencies of “pleasant” ratings for the German and foreign flag trials, and calculated a difference score. Higher values indicate more positive automatic affective reactions toward Germany (M = 0.10, SD = 0.32). As expected, there was a significant correlation between implicit and explicit attitudes, r(1105) = .28, p < .001.
News choice. As outlined above, relevant news items come from different news-media brands (e. g., ZDF or ATV) whose news organizations operate in different nations (e. g., Germany or Turkey) with potentially different perspectives on the same politically relevant event. We took the greatest of care in translating this observation into the task we used to assess a general tendency to favor German news items (relative to news items from other countries). In fact, we used six news-choice trials. In each trial, we offered one item from the host country and one item from another country. Both items covered the same news event, but differed in the source (e. g., Süddeutsche Zeitung vs. Sabah), the nation (Germany or Turkey), and the frame (e. g., “Big protests against Erdoğan in Istanbul”, “Gülen protests in Istanbul are smaller than noted by protest organizers”). We used headlines as news items because this content dimension is relevant in a number of news contexts including, but not limited to, a teaser on linear TV, as texts on media organizations’ websites, or as news-media organizations’ postings on social networking sites. We used newspaper and television sources that existed in reality. The news events were inspired by actual news coverage, but headlines were adapted for the present study. We used frames that reflected the political developments current at the time and thus considered external validity issues. In each trial, the German news item was coded as 1 and the foreign news item was coded as 0. We summed up all items. Higher values indicate a greater tendency to read German news items (M = 3.12, SD = 1.59, range = 0–6, αsplit-half = 0.60). A list of the headlines is provided in the Appendix. This measure was used for the test of H1.
Turkey related news choice. Two of the six news-choice trials included Turkish sources (Sabah, ATV). Consistent with the general news-choice measure, we summed up all responses. Higher scores indicate a greater tendency to choose German news items (M = 1.15, SD = 0.78, range = 0–2). This measure was used for the test of H2.
Voting decision. A total of 346 individuals with Turkish citizenship reported that they had participated in the Turkish constitutional referendum. Of these, 105 refused to tell us their voting decision. Of the remaining 241 individuals (= 100 %), n = 124 (= 51.5 %) voted for “Yes” [Turkish: “Evet”] and n = 117 (= 48.5 %) voted for “No” [Turkish: “Hayır”].
Secondary variables (controls)
Hostile media perception. We used a 7-point scale with four items to measure perceived negativity of Islam-related news coverage (e. g., “In German media, Muslims are represented as a threat to security”, “In German media, Muslims are represented only in terms of negative issues”) (M = 5.35, SD = 1.68, α = 0.94).
Political participation. We presented a list of 28 activities indicative of political participation. We asked individuals to tell us whether they had undertaken each activity during the previous two months (M = 4.24, SD = 3.99, αsplit-half = 0.87).
Political interest. A 5-point single item was used to measure political interest (“How much interest do you have in German politics?”) (M = 3.35, SD = 1.10).
Political knowledge. We used three knowledge questions using a multiple-choice answer format with four answer options (“Who votes for the German Chancellor?”, “What does the term ‘election secrecy’ mean?”, “What does the term ‘representative democracy’ mean?”). A minority (12.4 %) were unable to answer even a single question, 25.2 % correctly answered one question, 26.5 % correctly answered two questions, and 36.0 % correctly answered all three questions. We summed up all correct answers (M = 1.86, SD = 1.04).
German identity. Participants were asked to rate their agreement with four statements (e. g., “I strongly identify myself as a German”) (M = 4.32, SD = 1.74, α = 0.90).
Muslim identity. Participants were asked to rate their agreement with four statements (e. g., “I strongly identify myself as a Muslim”) (M = 5.24, SD = 2.05, α = 0.97). The wording of the items was the same as the wording for the German identity measure.
Perceived discrimination. We used a 7-point scale with six items to measure perceived discrimination of Muslims in Germany (e. g., “Muslim children are discriminated against by Germans”, “Many people in Germany avoid Muslims”) (M = 4.37, SD = 1.31, α = 0.82).
Democracy deficit. Participants were asked to rate a series of statements related to democracy and democratic values (e. g., “Parliaments such as the German Bundestag are useless”, “Women should be positioned below men”, “The country should be led by a strong man”) on a 7-point scale (M = 2.27, SD = 1.01, α = 0.73).
Statistical analysis
Consistent with previous research on the role of implicit and explicit attitudes as predictors for news choice (Arendt et al., 2016, 2019), we used hierarchical regression analysis for the test of H1. We put all controls (i. e., secondary variables and demographics) into the first step. Explicit attitudes were included in the second step and implicit attitudes were included in the third step. The change in R² of the second (third) step indicates whether explicit (implicit) attitudes show an incremental contribution in explaining variance. For the test of the mediation model (H2), we used PROCESS (Hayes, 2013). We predicted voting decision (outcome) by explicit and implicit attitudes (focal predictors). Turkey-related news choice was used as the mediator variable. We did not use control variables in this analysis due to the small sample size. The inclusion of control variables would have reduced the sample size even further.
Results
H1 assumed that implicit and explicit attitudes toward Germany predict news choice in favor of German news items. Correlational analysis revealed positive bivariate relationships between explicit attitudes and news choice, r(1105) = .31, p < .001, and implicit attitudes and news choice, r(1105) = .33, p < .001. As can be seen in Table 1, a hierarchical multiple regression model revealed that both implicit and explicit attitudes predicted news choice when controlling for several other variables. As indicated by the change in R² of the second step, explicit attitudes significantly added predictive value to the set of controls. Importantly, and as indicated by the change in R² of the third step, implicit attitudes explained additional variance. This analysis indicates that the more favorable the implicit and explicit attitudes toward Germany were, the more participants selected German news items. This supports H1.
On a descriptive level, the predictive power of implicit attitudes was stronger than that of explicit attitudes. For explorative purposes, we reran the regression model with z-standardized variables. This allowed us to calculate 95 % confidence intervals (CIs) for the regression coefficients of both attitude measures. Confidence intervals overlapped, implicit attitudes, B = 0.20, SE = 0.03, 95 % CI [0.14, 0.26], explicit attitudes, B = 0.14, SE = 0.04, 95 % CI [0.07, 0.21], indicating that implicit attitudes did not elicit a significantly stronger effect on news choice than explicit attitudes did.
Of the controls, age, hostile media perception, German identity, Muslim identity, and democracy deficit predicted news choice. An older age, a weaker hostile media perception, a stronger German identity, a weaker Muslim identity, and a weaker democracy deficit increased the selection of German news.
H2 predicted a mediator model for a subsample of Turkish citizens living in Germany. Implicit and explicit attitudes were hypothesized to exert an indirect effect on the voting decision via Turkey-related news choice (see Figure 1). Analysis supported this prediction. Importantly, participants who preferred selecting German news items tended to vote “No” in the referendum. Both indirect effects were significant. The full binary logistic regression model—predicting voting decision with implicit attitudes, explicit attitudes, and Turkey-related news choice—correctly classified 76.8 % of participants’ voting decisions, χ² (3) = 98.49, p < .001, Nagelkerke’s R² = .45. Taken together, the analysis supports H2. Furthermore, we reran the mediation model using the general news-choice measure instead of the Turkey-related news-choice measure. We obtained similar findings that can be obtained upon request.
Mediation analysis (N = 241 Turkish citizens living in Germany). All coefficients represent unstandardized regression coefficients. The number of bootstrap samples for bias-corrected confidence intervals = 1,000. Reading example: Positive implicit attitudes toward Germany increased the tendency to select German news items, which in turn increased the tendency to vote “No” in the referendum.
Hierarchical multiple regression analyses predicting German news choice.
News choice: Preference for German news | ||||
B | SE | β | p | |
Step 1: Controls ∆R² = .22, p <.001 German citizenship Age Gender Education Hostile media perception Political participation Political interest Political knowledge German identity Muslim identity Perceived discrimination Democracy deficit | .02 .01 .03 .09 –.11 .02 .04 –.12 .16 –.20 –.07 –.21 | .11 <.01 .10 .12 .03 .01 .05 .05 .03 .03 .04 .05 | .01 .06 .01 .02 –.11 .04 .03 –.08 .18 –.26 –.06 –.14 | .838 .062 .756 .443 .002 .170 .485 .025 <.001 <.001 .104 <.001 |
Step 2: Explicit attitudes ∆R² = .02, p <.001 German citizenship Age Gender Education Hostile media perception Political participation Political interest Political knowledge German identity Muslim identity Perceived discrimination Democracy deficit Explicit attitudes: Germany | .02 .01 .07 .08 –.10 .02 .03 –.11 .10 –.20 –.01 –.13 .25 | .10 <.01 .10 .12 .03 .01 .05 .05 .03 .03 .04 .06 .05 | .01 .08 .02 .02 –.10 .06 .02 –.07 .11 –.26 –.01 –.09 .19 | .826 .021 .500 .519 .003 .053 .589 .038 .001 <.001 .826 .016 <.001 |
Step 3: Implicit attitudes ∆R² = .03, p <.001 German citizenship Age Gender Education Hostile media perception Political participation Political interest Political knowledge German identity Muslim identity Perceived discrimination Democracy deficit Explicit attitudes: Germany Implicit attitudes: Germany | .03 .01 .08 .07 –.11 .02 .05 –.07 .08 –.17 <.01 –.15 .19 .99 | .10 <.01 .10 .12 .03 .01 .05 .05 .03 .03 .04 .05 .05 .15 | .01 .07 .02 .02 –.11 .06 .01 –.04 .08 –.22 <.01 –.09 .14 .20 | .735 .027 .431 .530 .001 .063 .873 .187 .013 <.001 .914 .006 <.001 <.001 |
Note: Full Model: F(14, 897) = 24.96, R² = .28, p < .001. Reading Example: The more positive a participant’s implicit attitudes toward Germany were, the more this person tended to select German news items (relative to foreign news items). Significant effects are indicated by bold numbers and letters in the full model (step 3). | ||||
Discussion
Most previous studies on attitude-based selective exposure have used explicit attitudes as a predictor variable for news choice. The underlying assumption is that overtly-expressed evaluations based on reasoning influence news choice. Recent research, however, has theorized that automatic affective reactions are also able to predict news choice. The present study contributes to this literature in three important ways: First, the present study additionally used implicit attitudes as a predictor variable and adds to the hitherto limited amount of supporting empirical evidence for both constructs’ predictive power. This supports recent theorizing on the role of implicit and explicit attitudes for selective exposure to news content (Arendt et al., 2016, 2019; Galdi et al., 2012). Second, we used a different news-choice measure with high societal relevance (i. e., news stemming from the host country vs. other countries). Third, and as a supplement to previous research that relied heavily on student samples, the present study provides supporting evidence using a non-student sample.
Results are consistent with a model theorizing the reciprocal multi-causality of media user’s predispositions, selective exposure, and corresponding effects (Knobloch-Westerwick, 2015; see also Slater, 2007, and Valkenburg and Peter, 2013): Evidence indicates that predispositions (implicit and explicit attitudes toward the host country) predict selective exposure to a host country’s news media which in turn elicits effects on societally relevant consequences.
An attitude-based selection bias in favor of German news was statistically related to the decision to vote “No” in the referendum. A “Yes” vote has been viewed as problematic for democracy, for example, as emphasized in a recent report by the Venice Commission (2017). Citizens who voted “Yes” (more or less) knowingly agreed to a political system in which executive power would be in a single person’s hands, while at the same time, parliamentary control of that power would be substantially weakened (Venice Commission, 2017). This can be interpreted as a step closer to authoritarian rule and one step further away from a democratic system (Venice Commission, 2017). A “No” vote, however, opposes such a system. Thus, a “No” vote can be deemed as more strongly in line with current German political values. Based on this perspective, an increased number of “No” votes may be interpreted as a beneficial outcome of an attitude-based selection bias (in favor of German news) on processes of integration in Germany.
However, we want to stress again that consuming country-of-origin media does not necessarily mean a lack of integration (Sauer, 2010; Trebbe and Weiß, 2007). Media consumption can become disintegrative when foreign media is consumed exclusively and/or when the content has a disintegrative character (Geißler and Weber-Menges, 2013; Müller, 2009). In fact, news use from other countries can be beneficial, especially when the host country’s media system only offers a very restricted “marketplace of ideas”. Under these circumstances, news consumption from other countries may even be able to enrich a news-consumer’s knowledge and contribute to a rich pool of diverse arguments. This individual may even be able to escape an echo chamber (Sunstein, 2009) or a filter bubble (Pariser, 2011) that has emerged within a given country by selecting news from other countries. This fact illustrates that the implications of host country news consumption have no deterministic, one-sided effect on processes of integration. It depends on the context.
As already noted, regular exposure to a host country’s news media has been deemed supportive for processes of integration (Peeters and d’Haenens, 2005; Sauer, 2010; Trebbe, 2007). The question emerges about what can be done, for example, by a government or NGOs to increase exposure? We want to emphasize two strategies that may supplement each other. First, individuals with exclusive country-of-origin news consumption—for example, a recent survey found that 29 % of individuals with Turkish migration backgrounds living in Germany only use Turkish television (Gerhard and Beisch, 2011)—can be motivated to consume a host country’s news. This strategy focuses on a direct change of news consumption based on (more or less) deliberate reasoning. For example, an information campaign accompanied by free newspaper subscriptions could be used to trigger thoughts on the fact that exposure to opposing arguments is indispensable for democracy’s ideal of a marketplace of ideas. Accordingly, an individual may reflect on his/her own (explicit) attitudes and may consciously decide to counteract a possible confirmation bias. Such a strategy affords an individual’s ability (time, cognitive resources) and motivation to initiate this strategic reasoning relevant for behavior change. Unfortunately, news users often lack the time, motivation, or cognitive capacity to select specific content in a deliberate way (Arendt et al., 2016; Knobloch-Westerwick, 2015).
Second, a change in attitudes toward the host country may elicit a beneficial indirect effect on German news use. Implicit attitudes may be an especially worthwhile target. As already noted, automatic affective reactions to news items elicit a selection tendency within a few hundred milliseconds of exposure. Importantly, these automatic responses seem to enter the decision-making stream moments before conscious reasoning can take place (Lodge and Taber, 2013), as required by the first strategy. A potent research question worth testing in future studies is whether changing (automatic) attitudes toward the host country would positively contribute to news choice. Politicians can, for example, support programs in which successful individuals with migration backgrounds (e. g., sports celebrities) go into schools to influence attitudes toward the host country. Importantly, if these role models explicitly refer to the host country’s news media and, for example, promote newspaper subscriptions, both (direct and indirect) effects may emerge, possibly eliciting the greatest success. This combined strategy can be supported by successful journalists with migration backgrounds.
Limitations
There are several limitations worth noting. First, the causal relationship between attitudes and news choice is unclear. We prefer the interpretation that attitudes causally influence news choice. This is in line with the selective-exposure theory outlined above. However, it is also possible that news choice (as a stable trait) has influenced implicit and explicit attitudes over the course of the days, weeks, months, or even years prior to the data collection for our study. We cannot rule out this possibility. In fact, we argued that both factors are likely to reciprocally influence each other. Thus, the “selective exposure” and the “effects” causal interpretations do not rule each other out.
Second, the news-choice measure may elicit reliability concerns. Although internal consistency was low, it was still at an acceptable level. The low value was presumably due to the low number of news-choice trials. Unfortunately, we were unable to use more choice trials due to time constrains. The fact that we found significant effects supports our assumption about the adequacy of the measure. It has to be noted that the effects of both attitude constructs might be even more pronounced if a more reliable news-choice measure were to have been used.
Third, we used the affective misattribution procedure for the measurement of implicit attitudes. Importantly, this procedure is a relative attitude measure. Implicit attitudes were assessed as automatic affective reactions toward Germany relative to other countries. The question arises as to whether it was a positive attitude toward Germany that influenced news choice in favor of German news items (our preferred interpretation) or whether a negative attitude toward other countries influenced the selection of news items. This is up to future research to determine.
Conclusions
Previous research on attitude-based selective exposure has used explicit attitudes as a predictor variable for news choice. Although the present study is consistent with the assumption that overtly-expressed evaluations based on reasoning influence news choice (i. e., explicit attitudes), recent research has theorized that automatic affective reactions are also able to predict news choice (i. e., implicit attitudes). The present study provides supporting evidence for the assumption that both implicit and explicit attitudes predict news choice. Importantly, an attitude-based selection bias was also related to voting, pointing to the societal relevance of an attitude-based bias in selective exposure. Future selective exposure research should more strongly focus on the reciprocal multi-causality of media user’s (automatic and deliberate attitudinal) predispositions, selective exposure, and corresponding effects.
References
Arendt, F., Northup, T., & Camaj, L. (2017). Selective exposure and news media brands: Implicit and explicit attitudes as predictors of news choice. Media Psychology, 22, 526–543. doi:10.1080/15213269.2017.133896310.1080/15213269.2017.1338963Search in Google Scholar
Arendt, F., Steindl, N., & Kümpel, A. (2016). Implicit and explicit attitudes as predictors of gatekeeping, selective exposure, and news sharing: Testing a general model of media-related selection. Journal of Communication, 66, 717–740. doi:10.1111/jcom.1225610.1111/jcom.12256Search in Google Scholar
Arnold, A.-K., & Schneider, B. (2007). Communicating separation? Journalism: Theory, Practice & Criticism, 8(2), 115–136. doi:10.1177/146488490707480710.1177/1464884907074807Search in Google Scholar
Berry, J. W. (1997). Immigration, acculturation, and adaptation. Applied Psychology, 46, 5–34. doi:10.1111/j.1464–0597.1997.tb01087.x10.1111/j.1464-0597.1997.tb01087.xSearch in Google Scholar
Bonfadelli, H., Bucher, P., & Piga, A. (2007). Use of old and new media by ethnic minority youth in Europe with a special emphasis on Switzerland. Communications, 32, 141–170. doi:10.1515/COMMUN.2007.01010.1515/COMMUN.2007.010Search in Google Scholar
Bonfadelli, H., & Moser, H. (2006). Medien und Migration: Europa als multikultureller Raum? [Media and migration: Europe as a multicultural space?] Wiesbaden, Germany: VS.Search in Google Scholar
Bundesamt für Migration und Flüchtlinge (2017). Das Bundesamt in Zahlen 2016. Asyl, Migration und Integration [The Federal Office in numbers 2016. Asylum, migration and integration]. Nuremberg, Germany: Bundesamt für Migration und Flüchtlinge.Search in Google Scholar
Buss, D. (2009). Evolutionary psychology: The new science of the mind. Boston, MA: Pearson.Search in Google Scholar
Chaffee, S. H., Nass, C. I., & Yang, S. M. (1990). The bridging role of television in immigrant political socialization. Human Communication Research, 17, 266–288. doi:10.1111/j.1468-2958.1990.tb00233.x10.1111/j.1468-2958.1990.tb00233.xSearch in Google Scholar
DellaVigna, S., & Kaplan, E. (1997). The Fox News effect: Media bias and voting. The Quarterly Journal of Economics, 122, 1187–1234.10.3386/w12169Search in Google Scholar
DESTATIS – Statistisches Bundesamt (2018). Persons with a migration background. Retrieved December 12th, 2017 from https://www.destatis.de/EN/FactsFigures/SocietyState/Population/MigrationIntegration/Methods/MigrationBackground.html.Search in Google Scholar
Deutschlandfunk (2017, January 9). Günter Seufert (SWP) – Politologe: Präsidialsystem wäre Ende der Demokratie in der Türkei [Guenter Seufert (SWP) – Political scientist: Presidential system would be the end of democracy in Turkey]. Retrieved December 12th, 2017 from http://www.deutschlandfunk.de/guenter-seufert-swp-politologe-praesidialsystem-waere-ende.694.de.html?dram%3Aarticle_id=375892 2017.Search in Google Scholar
Druckman, J., & Parkin, M. (2005). The impact of media bias: How editorial slant affects voters. The Journal of Politics, 67, 1030–1049. doi:10.1111/j.1468-2508.2005.00349.x10.1111/j.1468-2508.2005.00349.xSearch in Google Scholar
Festinger, L. (1957). A theory of cognitive dissonance. Evanston, IL: Row Peterson.10.1515/9781503620766Search in Google Scholar
Galdi, S., Gawronski, B., Arcuri, L., & Friese, M. (2012). Selective exposure in decided and undecided individuals: Differential relations to automatic associations and conscious beliefs. Personality and Social Psychology Bulletin, 38, 559–569. doi:10.1177/014616721143598110.1177/0146167211435981Search in Google Scholar
Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in evaluation: An integrative review of implicit and explicit attitude change. Psychological Bulletin, 132, 692. doi:10.1037/0033-2909.132.5.692.10.1037/0033-2909.132.5.692Search in Google Scholar
Geißler, R., & Weber-Menges, S. (2013). Medien und Integration. Mediennutzung und Vorstellungen zur medialen Integration bei Migranten [Media and integration. Media use and media-based integration]. In H. Brinkmann & H. Uslucan (Eds.), Dabeisein und Dazugehören. Integration in Deutschland [To be a part and have a sense of belonging. Integration in Germany] (pp. 273–289). Wiesbaden, Germany: VS.10.1007/978-3-531-19010-5_14Search in Google Scholar
Gerhard, H., & Beisch, N. (2011). Fernsehnutzung von Migranten 2011. Ergebnisse der ARD/ZDF-Studie Migranten und Medien 2011 [TV use of migrants 2011. Results of the ARD/ZDF study migrants and media 2011]. Media Perspektiven, 10, 479–492.Search in Google Scholar
Greenwald, A. G., Banaji, M., Rudman, L., Farnham, S., Nosek, B., & Mellot, D. (2002). A unified theory of implicit attitudes, stereotypes, self-esteem, and self-concept. Psychological Review, 109, 3–25. doi:10.1037//0033-295X.109.1.310.1037/0033-295X.109.1.3Search in Google Scholar
Halm, D., & Sauer, M. (2017). Muslime in Europa integriert, aber nicht akzeptiert? [Muslims in Europe: Integrated but not accepted?] Gütersloh, Germany: Bertelsmann Stiftung.Search in Google Scholar
Hart, W., Albarracin, D., Eagly, A., Brechan, I., Lindberg, M., & Merrill, L. (2009). Feeling validated versus being correct: A meta-analysis of selective exposure to information. Psychological Bulletin, 135, 555–588. doi:10.1037/a001570110.1037/a0015701Search in Google Scholar
Haug, S., Müssig, S., & Stichs, A. (2009). Muslimisches Leben in Deutschland – im Auftrag der Deutschen Islam Konferenz [Muslim life in Germany – on behalf of the German Islam Conference]. Retrieved December 12th, 2017 from http://www.bmi.bund.de/SharedDocs/Downloads/DE/Themen/Politik_Gesellschaft/DIK/vollversion_studie_muslim_leben_deutschland_.html.Search in Google Scholar
Hayes, A. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press.Search in Google Scholar
Iyengar, S., & Hahn, K. (2009). Red media, blue media: Evidence of ideological selectivity in media use. Journal of Communication, 59, 19–39. doi:10.1111/j.1460-2466.2008.01402.x.10.1111/j.1460-2466.2008.01402.xSearch in Google Scholar
Kim, Y. Y. (1988). Communication and cross-cultural adaptation: An integrative theory. Philadelphia, PA: Multilingual Matters.10.1002/9781118783665.ieicc0041Search in Google Scholar
Knobloch-Westerwick, S. (2015). Choice and preference in media use: Advances in selective exposure theory and research. New York, NY: Routledge.10.4324/9781315771359Search in Google Scholar
Knobloch-Westerwick, S., & Johnson, B. K. (2014). Selective exposure for better or worse: Its mediating role for online news’ impact on political participation. Journal of Computer-Mediated Communication, 19, 184–196. doi:10.1111/jcc4.1203610.1111/jcc4.12036Search in Google Scholar
Kunst, J. R., Tajamal, H., Sam, D. L., & Ulleberg, P. (2012). Coping with Islamophobia: The effects of religious stigma on Muslim minorities’ identity formation. International Journal of Intercultural Relations, 36, 518–532. doi:10.1016/j.ijintrel.2011.12.01410.1016/j.ijintrel.2011.12.014Search in Google Scholar
Liu, W., & Gastil, J. (2014). Pathways of immigrant political socialization: Examining the role of news media, social connections, and community interaction. Journal of Intercultural Communication Research, 43, 238–263. doi:10.1080/17475759.2014.93921410.1080/17475759.2014.939214Search in Google Scholar
Lodge, M., & Taber, C. (2013). The rationalizing voter. New York, NY: Cambridge University Press.10.1017/CBO9781139032490Search in Google Scholar
Mahajan, N., Martinez, M., Gutierrez, N., Diesendruck, G., Banaji, M., & Santos, L. (2011). The evolution of intergroup bias: Perceptions and attitudes in rhesus macaques. Journal of Personality and Social Psychology, 100, 387–405. doi:10.1037/a002245910.1037/a0022459Search in Google Scholar
Müller, D. (2009). Uyum statt entegrasyon? Zur Europa-Ausgabe der türkischen Zeitung Hürriyet [Uyum instead of entegrasyon? The European issue of the Turkish newspaper Hürriyet]. In R. Geißler & H. Pöttker (Eds.), Massenmedien und die Integration ethnischer Minderheiten in Deutschland (pp. 299–315). Bielefeld: transcript Verlag. doi:10.14361/9783839410271-01110.14361/9783839410271-011Search in Google Scholar
Noelle-Neumann, E. (1973). Kumulation, Konsonanz und Öffentlichkeitseffekt. Ein neuer Ansatz zur Analyse derWirkung der Massenmedien [Accumulation, consonance, and the publicness effect. A new approach to the analysis of media effects]. Publizistik, 18, 26–55.Search in Google Scholar
Nygren, G., Glowacki, M., Hök, J., Kiria, I., Orlova, D., & Taradai, D. (2016). Journalism in the crossfire: Media coverage of the war in Ukraine in 2014. Journalism Studies. Advance online publication. doi:10.1080/1461670X.2016.125133210.1080/1461670X.2016.1251332Search in Google Scholar
Olson, M. A., & Fazio, R. H. (2009). Implicit and explicit measures of attitudes: The perspective of the MODE model. In R. E. Petty, R. H. Fazio & P. Briñol (Eds.), Attitudes: Insights from the new implicit measures (pp. 19–63). New York, NY: Psychology Press.Search in Google Scholar
OSCE Office for Democratic Institutions and Human Rights (2017). Turkey, Constitutional Referendum. 16 April 2017: Final Report, Warsaw. Retrieved December 12th, 2017 from http://www.osce.org/odihr/elections/turkey/324816?download=true.Search in Google Scholar
Pariser, E. (2011). The filter bubble: What the internet is hiding from you. London, England: Penguin Books Ltd.Search in Google Scholar
Payne, K., Cheng, C., Govorun, O., & Stewart, B. (2005). An inkblot for attitudes: Affect misattribution as implicit measure. Journal of Personality and Social Psychology, 89, 277–293. doi:10.1037/0022-3514.89.3.27710.1037/0022-3514.89.3.277Search in Google Scholar
Peeters, A. L., & d’Haenens, L. (2005). Bridging or bonding? Relationships between integration and media use among ethnic minorities in the Netherlands. Communications, 30, 201–231. doi:10.1515/comm.2005.30.2.20110.1515/comm.2005.30.2.201Search in Google Scholar
Pew Research Center (2017). Europe’s growing Muslim population. Retrieved December 12th, 2017 from Pew Research Center, http://www.pewforum.org/wp-content/uploads/sites/7/2017/11/FULL-REPORT-FOR-WEB-POSTING.pdf.Search in Google Scholar
Roman, N., Wanta, W., & Buniak, I. (2017). Information wars: Eastern Ukraine military conflict coverage in the Russian, Ukrainian and U.S. newscasts. International Communication Gazette, 79, 357–378. doi:10.1177/174804851668213810.1177/1748048516682138Search in Google Scholar
Sauer, M. (2010). Mediennutzungsmotive türkeistämmiger Migranten in Deutschland [Motives of media use of Turkish immigrants in Germany]. Publizistik, 55, 55–76. doi:10.1007/s11616-010-0074-510.1007/s11616-010-0074-5Search in Google Scholar
Slater, M. D. (2007). Reinforcing spirals: The mutual influence of media selectivity and media effects and their impact on individual behavior and social identity. Communication Theory, 17, 281–303. doi:10.1111/j.1468-2885.2007.00296.x10.1111/j.1468-2885.2007.00296.xSearch in Google Scholar
Stichs, A. (2016). Wie viele Muslime leben in Deutschland? Eine Hochrechnung über die Anzahl der Muslime in Deutschland zum Stand 31. Dezember 2015 [How many Muslims live in Germany? A projection of the number of Muslims in Germany]. Retrieved December 12th, 2017 from https://www.bamf.de/SharedDocs/Anlagen/DE/Publikationen/WorkingPapers/wp71-zahl-muslime-deutschland.pdf?__blob=publicationFile.Search in Google Scholar
Sunstein, C. (2009). Republic.com 2.0. Princeton, NJ: Princeton University Press.Search in Google Scholar
Trebbe, J. (2007). Types of integration, acculturation strategies and media use of young Turks in Germany. Communications, 32, 171–191. doi:10.1515/COMMUN.2007.01110.1515/COMMUN.2007.011Search in Google Scholar
Trebbe, J., & Weiß, H.-J. (2007). Integration als Mediennutzungsmotiv? Eine Typologie junger türkischer Erwachsener in Nordrhein-Westfalen [Integration as motif for media use? A typology of young Turkish adults in North Rhine-Westphalia]. Media Perspektiven, 3, 136–141.Search in Google Scholar
Valkenburg, P. M., & Peter, J. (2013). The differential susceptibility to media effects model. Journal of Communication, 63, 221–243. doi:10.1111/jcom.1202410.1111/jcom.12024Search in Google Scholar
Venice Commission (2017). Opinion on the amendments to the constitution adopted by the Grand National Assembly on 21 January 2017 and to be submitted to a national referendum on 16 April 2017. Opinion No. 875/2017. CDL-AD(2017)005.Search in Google Scholar
Yüksek Seçim Kurulu [Supreme Board of Elections] (2017). Data on the Constitutional Referendum 2017 [Data]. Retrieved Retrieved December 12th, 2017 from http://www.ysk.gov.tr/.Search in Google Scholar
Zillmann, D., & Bryant, J. (1985). Selective exposure to communication. Hillsdale, NJ: Lawrence Erlbaum Associates.10.4324/9780203056721Search in Google Scholar
Appendix
Translation of headlines used in the news-choice trials:
Question asked in each trial: “Which one out of the two articles would you prefer to read?”
Trial 1
Süddeutsche Zeitung: Big protests against Erdoğan in Istanbul.
Sabah: Gülen protests in Istanbul are smaller than noted by protest organizers.
Trial 2
Die Welt: Merkel talks about human rights violations during a state visit to Saudi Arabia.
AI-Madina: After visiting Merkel, Saudi Arabian and German trade relations are consolidated.
Trial 3
Bild Zeitung: Terrorist attack in Jerusalem: Palestinian detonated car bomb.
The Palestine Chronicle: Israeli army kills Palestinian after a car explosion in Jerusalem.
Trial 4
ProSieben: Freedom of press in Turkey becomes more and more restricted.
Aktüel Televizyonu (ATV): Two journalists arrested in Istanbul.
Trial 5
ZDF: American President worried about rocket test in Iran.
Mohabat TV: Further provocation by USA after rocket test.
Trial 6
ARD: Increased Islamist-motivated terrorist attacks in European cities.
Al Manar TV: Europe holds Islam responsible for terror.
Note. Although we randomized the order in which both headlines appeared within each trial, we present the German news item in the first position here in the Appendix.
© 2020 Arendt, et al., published by De Gruyter
This work is licensed under the Creative Commons Attribution 4.0 International License.
