https://orcid.org/0000-0001-7392-9791
Abstract
In contemporary high-choice media environments, the issue of media trust and its impact on people's media use has taken on new importance. At the same time, the extent to which people trust the news media and how much it matters for their use of different types of media is not clear. To lay the groundwork for future research, in this article we offer a focused review of (a) how news media trust has been conceptualized and operationalized in previous research and (b) research on the extent to which news media trust influences media use, and (c) offer a theoretically derived framework for future research on news media trust and its influence on media use.
Introduction
From a democratic perspective, a key function of news media is to ‘aid citizens in becoming informed’ (Holbert, 2005, p. 511). For the news media to fulfill this function, an important prerequisite is that they provide people with the kind of information they need to be free and self-governing (Kovach & Rosenstiel, 2014; Strömbäck, 2005). In itself, this is however not enough. Of key importance is also that people use, and that they trust, the news media. After all, even a perfectly informative news media environment is of little democratic use if citizens by and large do not consume the news or if they do not trust the news.
At the same time, research suggests that media trust is either falling or that many citizens do not trust the news media. In the US, for example, Gallup has shown that the share expressing a great deal or a fair amount of ‘trust and confidence’ in ‘the mass media’ declined from 68 percent in 1968 to 32 percent in 2016 (Jones, 2018), and although it has rebounded since, only 12 percent report ‘a great deal’ of confidence in the media (Guess et al., 2018). Other comparative research suggests that the share trusting ‘most news most of the time’ is about 49 percent across all countries investigated (Newman et al., 2019, p. 20). While claims that media trust in general is falling are exaggerated (Hanitzsch et al., 2018), it is thus clear that news media trust is at least fragile.
Adding to the above, the transformation into high-choice media environments (Prior, 2007; Van Aelst et al., 2017) has brought with it several new or increased challenges for traditional news media and news media trust. First, like never before, news media today face competition for people's attention from a myriad of other sources of information. Second, many of news media's newer competitors are so-called alternative and partisan media, in which attacks on traditional news media for being untrustworthy is a prominent feature (Jamieson & Cappella, 2008; Ladd, 2012). Third, digital and social media have made political and other social actors less dependent on news media to reach the public, allowing them to by-pass the news media but also providing channels for attacks on the news media (Groshek & Koc-Michalska, 2017). Fourth, there is probably more so-called ‘fake news’ (Egelhofer & Lecheler, 2019), disinformation and misinformation circulating in the public arena than ever (Benkler et al., 2018; Kavanagh & Rich, 2018). There are also indications that leading politicians are less afraid than before of being caught with providing misinformation, with president Trump as a prime example (PolitiFact, 2019). Fifth, it is an established fact that people tend to prefer attitude-consistent information (Flynn et al., 2017) and to engage in motivated reasoning and skepticism when encountering challenging information (Kunda, 1990; Taber & Lodge, 2006). This holds particularly true for partisans and those whose worldviews are challenged by attitude-discrepant information (Lewandowsky et al., 2012). Those are also the ones most likely to fall prey for the hostile media phenomena, i.e. the tendency to perceive news media as being hostile towards one's own side while favoring the other side in a political conflict (Hansen & Kim, 2011; Vallone et al., 1985).
With all those challenges facing contemporary news media, it has become increasingly important for news media to be trusted. After all, why would people otherwise have more faith in the veracity of information coming from news media compared to that coming from other information sources? How could news media otherwise shield themselves from accusations from political actors or partisan media? And why would people otherwise choose to use news media when they can get information from other sources that will provide them with information that confirm their own attitudes and beliefs? In fact, there is some research showing that lack of trust in news media is related to less use of these and greater use of non-mainstream information sources (Kalogeropoulos et al., 2019; Ladd, 2012; Mourão et al., 2018; Tsfati & Cappella, 2003, 2005), suggesting that there is a linkage between news media trust and selective (non)exposure to news media.
At the same time, the extent to which (a) people trust the news media and (b) how much it matters for their use of these and other types of media is not clear. One reason is that too often, studies rely on single or too general indicators of media trust, while another is a lack of common conceptualizations and operationalizations of media trust (Engelke et al., 2019; Fischer, 2016). In addition, there are surprisingly few studies addressing the impact of news media trust on people's media use. Taken together, this suggests that there is a need for a reassessment of research on news media trust and its consequences in terms of how it relates to media use.
Against this background, the purpose of this article is twofold. The first is to offer a focused review of (a) how news media trust has been conceptualized and operationalized in previous research and of (b) research on the extent to which news media trust influences the use of traditional news media versus non-mainstream media.1 Based on that, the second is to suggest a framework for further research on news media trust and its influence on people's use of traditional news media as well as non-mainstream media.
Before proceeding, two caveats should be noted. First, in the literature, ‘media trust’ and ‘news media trust’ are often used interchangeably, but most of the time, the terms refer to trust in traditional news media such as newspapers. In this article, if nothing else is stated, we refer to traditional news media such as newspapers, television news and radio news – in their offline or online formats – when writing about news media and news media trust. Second, while a full understanding of the linkages between people's trust in news media and their media use would require equal attention to conceptualizations and operationalizations of news media trust and media use, in this article, we will focus on news media trust.
Conceptualizations of media trust
In the relevant literature, media trust is often discussed alongside related concepts such as media credibility and media trustworthiness (Engelke et al., 2019; Kiousis, 2001; Kohring & Matthes, 2007; Otto & Köhler, 2018; Tsfati & Cappella, 2003). Its opposite is usually conceptualized as distrust, media cynicism or media skepticism. As other types of trust, media trust describe a relation between two sides: ‘a trustor, the side that places trust, and a trustee, the side being trusted’ (Tsfati & Cappella, 2003, p. 505). Inherent in all such relationships is a degree of uncertainty, making the credibility of the trustee imperative for understanding the degree to which people trust the trustee. As Kohring (2019, p. 1) explains, ‘news media users do not have at their disposal the resources and capabilities to evaluate thoroughly the reliability of news … Thus, somehow they have to find clues to legitimate their trust and to compensate for this inevitable risk.’ Perceptions of media credibility can thus be conceptualized as encapsulating the clues that people use to evaluate their trust in media. This is one reason why news credibility is closely linked to, and often discussed interchangeably with, news media trust. Another important part of virtually all conceptualizations of both trust in general and news media trust is the expectation on the part of the trustor that interactions will lead to gains rather than losses for the trustor (Gambetta, 1988; Tsfati & Cappella, 2003; Patterson, 1999; Warren, 1999).
Research on media credibility dates back to Hovland and colleagues’ research on source credibility, which investigated how different source characteristics influence people's willingness to change their attitudes toward different issues (Hovland et al., 1953). Their research focused mainly on individual communicators, although they noted that ‘the impact of a message probably depends also upon the particular publication or channel through which it is transmitted’ (p. 19). Since then, it is common to distinguish between source credibility and medium credibility (Kiousis, 2001).
According to Hovland et al. (1953), source credibility consists of two main components: expertise and trustworthiness. As noted by Kohring and Matthes (2007, p. 233), it remains unclear however whether expertise and trustworthiness should be conceptualized as dimensions of or reasons for credibility. More importantly, subsequent research suggests that media credibility includes more components than expertise and trustworthiness (Gaziano & McGrath, 1986; Kohring & Matthes, 2007; Meyer, 1988; West, 1994). Some components that have been identified are the degree to which the media are perceived to be fair, unbiased and accurate (Gaziano & McGrath, 1986; Meyer, 1988; West, 1994). Worth noting, in some of the research attempting to develop and validate a scale for measuring media credibility, media trust is conceptualized as a dimension of media credibility rather than the other way around (Meyer, 1988; West, 1994).
More recently, several attempts have been made at developing and validating a multidimensional scale for the measurement of media trust. Beginning with Kohring and Matthes (2007), they build upon the notion that all trust relationships involve a certain degree of uncertainty where one social actor needs another social actor but cannot be sure how that second actor – be it an individual, organization or an institution – will behave in the future (p. 238). Furthermore, they assume that the most important function of news media is to select and convey the kind of information people need to understand politics and society (Kovach & Rosenstiel, 2014). As news reporting by necessity is selective, they argue that selectivity should be the basis for analyses of news media trust: ‘when people put their trust in news media, they take a certain risk. This is because journalists selectively choose some information over other information. Therefore, when trusting news media, people trust in specific selections’ (Kohring & Matthes, 2007, p. 239). More specifically, they posit that trust in news media consists of four separate dimensions: trust in the selectivity of topics, trust in the selectivity of facts, trust in the accuracy of depictions, and trust in journalistic assessment.
Another attempt at examining first- and second-order factor structures for news credibility was undertaken by Yale et al. (2015). More specifically, they focused on whether news credibility consists of several, empirically separate subdimensions, as suggested by some earlier studies (Abdulla et al., 2004; Gaziano & McGrath, 1986; Kohring & Matthes, 2007). Using experiments with individual news articles as stimuli and confirmatory factor analysis, they found insufficient discriminant validity among the subdimensions investigated, leading them to conclude that it is not possible to empirically establish that media credibility or media trust consists of different subdimensions. Instead, they recommend that any credibility scale ‘must be treated as a single-factor measure when used as a variable’ (p. 166).
The most recent study in this area was done by Prochazka and Schweiger (2019), in which they test both the trust in news media scale by Kohring and Matthes (2007) and the news credibility scale by Yale et al. (2015). Among other things, they found that both scales entailed a good model fit when applied to general trust in news media, but also – similar to Yale et al. (2015) – problems with discriminant validity among the subdimensions. This suggests that while it theoretically makes sense to conceptualize news media trust as consisting of different subdimensions, empirically, media trust should rather be considered a higher-order factor that need to be measured by multi-item scales but not different dimensions that are measured and used as separate variables (Prochazka & Schweiger, 2019).
Another thorny issue is related to the trustee, i.e. what ‘media’ as in ‘media trust’ refers to. Early on, research noted that trust in and the credibility of television and newspapers tend to differ, although there is usually a correlation between the two (Gaziano & McGrath, 1986; Kiousis, 2001; Metzger et al., 2003; Newhagen & Nass, 1989; Westley & Severin, 1964). Recent research has furthermore established that levels of media trust differ depending on whether it refers to news overall, news that people use, or news in digital and social media (Newman et al., 2019), whether it refers to an unspecified referent (such as ‘the press’ or ‘the media’) or specified news sources (Daniller et al., 2017; Eberl, 2019), and whether it refers to commercial versus public service broadcast news (Matsa, 2018). Media trust might also differ depending on whether it refers to national versus local media, across individual news media, and across the coverage of different topics (Matsa et al., 2018; Metzger et al., 2003; Mitchell et al., 2014; Mitchell et al., 2016).
Hence, it is clear that the concept of media is polysemic and might refer to many – partly overlapping – different facets of media, in particular in contemporary high-choice media environments with a greater media abundance than ever. This is important also when assessing time trends, in particular when questions relate to unspecified media. Assuming an accessibility bias is at work, in earlier low-choice media environments, people were likely thinking about quite similar mainstream news media when responding to questions about media trust (Ladd, 2012; Tsfati, 2002; Tsfati & Peri, 2006). In contemporary media environments, the media they are thinking about might be more diverse as well as more partisan or high-profile, in particular in countries where such media are prominent. In fact, Daniller et al. (2017) suggest that as much as 60 percent of the decline in Americans’ trust in ‘the press’ can be explained by shifts in the most accessible referent – i.e. what media people are thinking of – when responding to questions about media trust (p. 81). This is a reminder that ‘Even when the survey wording for an item remains constant, the meaning of answers to a survey question can still change’ (Daniller et al., 2017, p. 82).
Summing up, this review shows that there is significant variability in terms of how ‘trust’ as well as ‘news media’ in the context of ‘news media trust’ has been conceptualized (see also Engelke et al., 2019), and although it makes sense theoretically to think of news media trust as consisting of several subdimensions, empirically the conclusion must be that it is difficult to measure subdimensions of news media trust with sufficient discriminant validity (Prochazka & Schweiger, 2019; Yale et al., 2015). Having said so, at the broadest conceptual level, there is significant consensus that news media trust refers to the relationship between citizens (the trustors) and the news media (the trustees) where citizens, however tacit or habitual, in situations of uncertainty expect that interactions with the news media will lead to gains rather than losses (Gambetta, 1988; Kohring & Matthes, 2007; Ladd, 2012; Prochazka & Schweiger, 2019; Tsfati & Cappella, 2003; Warren, 1999; Yale et al., 2015).
Operationalizations of media trust
Turning to operationalizations of media trust, much research has focused on media trust at the general or institutional level, asking about people's trust or confidence in unspecified types of media such as ‘the press’ or ‘mass media’. One example is the Gallup poll, which have asked Americans about their media trust since the early 1970s. In these surveys, respondents are asked
This survey item has also been used by others, such as the Poynter Media Trust Survey (Guess et al., 2018). Another example is the General Social Survey in the US, in which respondents are asked: ‘I am going to name some institutions in this country. As far as the people running these institutions are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them.’ In this question battery, one of the institutions are ‘the press.’ A similar question is used in the World Values Surveys, asking:In general, how much trust and confidence do you have in the mass media – such as newspapers, TV and radio – when it comes to reporting the news fully, accurately and fairly – a great deal, a fair amount, not very much or none at all. (Jones, 2018)
This question is asked for ‘the press’ and ‘television,’ and is often used in cross-national studies on media trust (Hanitzsch et al., 2018; Tsfati & Ariely, 2014). A similar wording is used in the European Values Study.I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it a great deal of confidence, quite a lot of confidence, not very much confidence or none at all.
Another source of cross-national data is the Eurobarometer, which also focuses on unspecified media types. In these surveys, media trust is measured by the question:
In more recent surveys, this question is asked for radio, television, the written press, the internet, and online social networks (European Commission, 2018). A more recent addition to cross-national research is the Reuters Digital News Report, in which media trust is tapped by the question: ‘Please indicate the level of agreement with the following statements: I think you can trust most news/most news I consume/news in social media/news in search engines most of the time’ (Newman et al., 2019; Fletcher & Park, 2017; Kalogeropoulos et al., 2019). Important to note is that the question used by the Digital News Report asks about the level of trust in the news rather than in the news organizations or news institutions. This is similar to the Pew Research Center, that asks respondents ‘How much, if at all, do you trust the information you get from … ?’ The question is asked for national and local news organizations separately, as well as for social media. The response alternatives are ‘a lot,’ ‘some,’ ‘not too much,’ and ‘not at all’ (Mitchell et al., 2016).I would like to ask you a question about how much trust you have in certain media and institutions. For each of the following media and institutions, please tell me if you tend to trust it or tend not to trust it.
Turning to individual studies, there are many different operationalizations of media trust. Some of them include items from News Credibility Scale, developed by Gaziano and McGrath (1986). This includes 12 items asking respondents to ‘think about the daily’ newspaper or television news show ‘that you are most familiar with’ and, using a five-point scale anchored with word pairs with opposite meaning, state the value ‘that best represents how you feel about the’ daily newspaper or television news show they have in mind. The word pairs used were fair–unfair, biased–unbiased, tells – does not tell the who story–, accurate–inaccurate, invades–respects people's privacy, does–does not watch after readers’/viewers’ interests, is–is not concerned about the community's well-being, does–does not separate fact and opinion, can–cannot be trusted, is concerned about the public interest–is concerned about making profits, factual–opinionated, and has well-trained–poorly trained reporters (Gaziano & McGrath, 1986, pp. 454–455). Tsfati and Cappella (2003), for example, used four of these items (fair, accurate, tell the whole story, can be trusted), in addition to an item asking whether the media care more about being the first to report a story or about being accurate in reporting the story and one asking whether media help society or get in the way of society's solving its problems (adapted from Cappella & Jamieson, 1997) (see also Tsfati, 2010; Tsfati & Cappella, 2005). Kiousis (2001) also used five items from the News Credibility Scale (the media's factualness, motivation by public interest or commercial considerations, whether they invade or respects people's privacy, its concern for the community and whether they can be trusted), but measured on a four-point scale and for newspapers, television news and online news, respectively. Others have used Meyer's Credibility Index (1988), which consist of five statements from Gaziano and McGrath (1986): fair–unfair, unbiased–biased, tells–does not tell the whole story, accurate–inaccurate and can–cannot be trusted (Turcotte et al., 2015).
Quite many studies use single items to measure media trust, even if asked for different types of news media. Jones (2004) and Lee (2010), for example, used an item from the American National Election Study asking respondents ‘How much of the time do you think you can trust the news media to report the news fairly?,’ with the response alternatives ‘just about always, most of the time, only some of the time, almost never, and none of the time.’ Somewhat similarly, Ardèvol-Abreu and Gil de Zúniga (2017) used the single item ‘how much do you trust the news’ from alternative media, social media sites, mainstream news media and news aggregators, respectively, using a ten-point scale ranging from do not trust to trust completely. Less focused on the news that media report, Hopmann et al. (2015) on their end measured media trust by asking people ‘generally speaking, how much trust do you have in … ,’ Swedish national public service television and radio, respectively, morning newspapers, tabloids, and journalists. The response alternatives ranged from no trust at all (1) to very high trust (5). More focused on the content of news, Elvestad et al. (2018) used the measure ‘How much do you trust the news distributed by the sources listed below’ (on a four-point scale ranging from no trust at all to complete trust), including public service and commercial broadcasters and newspapers. As noted, several studies also use data from the World Values Survey (WVS) or the European Values Study (EVS) (Ariely, 2015; Hanitzsch et al., 2018; Tsfati & Ariely, 2014).
Then there are some studies that combine questions about news content, journalists and media corporations or outlets. Williams (2012), for example, assessed media trust by five items measured on a 10-point scale ranging from strongly disagree to strongly agree: ‘I trust the information that I get from the news media/the Internet,’ ‘Individuals reporting the news stories are helpful to others,’ ‘Individuals reporting news stories can be trusted,’ ‘Media corporations can be trusted’ and ‘Media corporations help solve social problems.’ Jackob (2010) similarly mixed measures of the extent to which respondents trust media reports with their trust in various professions and institutions.
While this review of operationalizations could be extended, it would not change the most important conclusions that can be drawn. First and most importantly, there is no agreed-upon measurement or operationalization of media trust (Engelke et al., 2019; Fischer, 2016, 2018). Second, there is great variability in terms of whether media trust refers to generalized news media trust or trust in different types of media, specific news outlets, in the news that is reported, or in journalists. Third, there is great variability in whether ‘trust’ refers to media as institutions or organizations, the people working in these institutions or organizations, or the news reported by various media. When questions deal with the news reported by various media, it also differs whether or not a specification is made in terms of whether the news coverage is, for example, comprehensive, accurate and fair. Fourth, there are only few studies that investigate if or how trust varies depending on the topic of news stories. Fifth, there are virtually no studies that link trust in media content to trust in journalists (but see Andersson, 2017), media organizations or brands, and media at the more institutional or general level.
On top of these limitations, the nature of mistrust as measured by media trust items is not clear (Engelke et al., 2019). Following Cappella and Jamieson (1997), an important distinction can be made between media skepticism – distrust based on close examination and consideration – and media cynicism – distrust without examination, based on a disposition not to trust. With respect to this distinction, we do not know whether the different scales or items used tap the former or the latter. This is important in relation to the impact of news media trust (and the lack thereof) on media use, as it impinges on the extent to which people avoid certain media because of experiences that they do not provide trustworthy information or because of hearsay or presumptions. In fact, overall there is quite limited research on the reasons people have for not trusting news media (but see Newman & Fletcher, 2017). The extent to which different scales reflect the concept of trust versus general attitudes (‘the affect for or against a psychological object,’ Thurstone, 1931) towards media is also unclear. Ladd (2012, pp. 94–96), for example, demonstrates that the trust item behaves similarly to, and is rather strongly correlated with, a more general media thermometer rating, and questions asking respondents to evaluate the performance of media coverage of a specific scandal.
The end result is that despite extensive research, our knowledge of news media trust might be more limited than what appears at first glance. Presumably this holds true also for the effects of media trust – including the effects on media use.
Connecting media trust to media use
Turning to the consequences of media trust, a common assumption in the general trust literature is that trust matters and has behavioral consequences, both on the individual and at the societal level (Easton, 1965; Gambetta, 1988; Putnam, 2000; Warren, 1999). This assumption is backed by empirical research demonstrating that trust affects a range of attitudes and behaviors such as civic engagement (Putnam, 1993), taxpaying (Batrancea et al., 2019), and teamwork (De Jong et al., 2016), to mention some examples. Based on this, it could be – and often is – assumed that levels of news media trust influences people's news media use as well. This holds particularly true as the relationship between individuals and the news media involves a certain degree of uncertainty in the sense that people usually cannot verify news media reports with non-media sources – although digital media have made this easier than before.
One a theoretical level, Tsfati and Cappella’s (2003, p. 508) theory connecting news media trust and assumes audience rationality, and in particular utility maximization. It further assumes that people turn to news in order to obtain accurate information about the world, and given utility maximization and the impossibility of attending to all news, all the time, that people have an incentive to ignore many stimuli. Given these theoretical assumptions and the definition of trust as ‘the expectation that the interaction with the trustee would lead to gains rather than losses’ (Tsfati & Cappella, 2003, p. 506), it does not make much sense to follow media that one does not trust. Theoretically, a correlation between news media trust and exposure can be expected.
However, it is also a fact that in many cases, people's media use is ritualized rather than instrumental and habitual rather than active (Rubin, 2009; Ruggiero, 2000), and that people use media, even news media and news genres, for other purposes than getting informed. These include diversion and entertainment, as substitute companionship, for social utility, and to serve personal or social identity needs (Blumler, 1979; Katz et al., 1973; Rubin, 2009; Tsfati & Cappella, 2005). In these cases, for example when people are seeking diversion or to understand diverse perspectives, the level of media trust might matter less compared to when media are used for informational and surveillance purposes. Also important are people's need for cognition (Tsfati & Cappella, 2005), their political or ideological preferences, and that people tend to prefer information that is attitude-consistent (Flynn et al., 2017; Garrett et al., 2013; Knobloch-Westerwick, 2014; Stroud, 2011). This might lead people to select media that are aligned with their political views rather than the media that they otherwise trust the most, although it is also a fact that people tend to trust news media that are aligned with their political views. In the US, for example, liberals tend to trust and use media such as PBS, while conservatives tend to trust and use media such as Fox News (Mitchell et al., 2014).
Beyond the individual level, research also shows that news media use is shaped by structural and semi-structural factors as well as situational ones (Althaus et al., 2009; Hallin & Mancini, 2004; Hartmann, 2009; Norris, 2002; Shehata & Strömbäck, 2011; Webster, 2014; Wonneberger et al., 2011). The media systems and the overall supply of different types of media, media platforms and media content also matter, as do the contexts in which people use news media and what alternatives to news media there are (Hartmann, 2009; Skovsgaard et al., 2018; Tsfati & Cappella, 2003; Webster, 2014).
In other words, the relationship between media trust and media use is complex. The question then is whether and how much news media trust matters for people's use of news media. Somewhat surprisingly, there is rather limited research directly addressing this question. There are some exceptions though. One example is Tsfati and Cappella (2003), who investigated the associations between mainstream media skepticism and exposure to mainstream news and non-mainstream news. In that study, non-mainstream news exposure was defined as political talk radio and internet. Among other things, their results showed that media skepticism was negatively associated with exposure to mainstream news media and positively associated with exposure to non-mainstream news media, and that media skepticism was associated with news diets consisting of a larger share of non-mainstream news media. These results lend some support to the notion that media trust influences the use of news media. At the same time, the associations between media trust and exposure were quite modest. As the authors note, ‘media skepticism is indeed associated with exposure, but it is definitely not the only factor that comes into play in the exposure decision’ (p. 518). In a follow-up study, Tsfati (2010) also found that media skepticism was positively associated with exposure to non-mainstream news sites. The same pattern was found by Jackob (2010). In both cases, the associations were again quite modest. In a more recent comparative study using data from the Digital News Report, Fletcher and Park (2017, p. 1291) also found that ‘very low trust is significantly associated with a preference for non-mainstream news sources,’ while the opposite pattern held for very high trust. The strength of the association varied across countries however, and in several cases was not significant. Similar results were found by Prochazka and Schweiger (2019) when testing different scales of media trust, by Mourão et al. (2018) when investigating the linkage between media trust on the one hand and different news repertoires on the other, and by Yuan (2011) when investigating the linkage between media credibility and the likelihood to include different media types in their news media repertoires. On the other hand, a study by Kalogeropoulos and colleagues (2019), also using data from the Digital News Report, found that using non-mainstream media (defined in this study as digital-born news outlets and news on social media) was positively – not negatively – associated with trust in the news. However, use of social media as the main source of information was associated with lower levels of trust (see also Schranz et al., 2018).
Nevertheless, overall previous research suggests that media trust is associated with greater use of news media while media distrust is associated with greater use of non-mainstream news sources, but that the relationship between media trust and media use is quite modest (see also Ladd, 2012). Three caveats should be added though. The first is that there is rather limited research, and that operationalizations of media trust as well as non-mainstream news sources differ. In addition, there are no studies on how media trust at different levels of analysis might influence media use, hindering a more nuanced and detailed understanding of the relationship between media trust and use. In fact, most studies correlate media trust on one level with media use on another (e.g, in Tsfati & Cappella, 2003 and in Kalogeropoulos et al., 2019, trust is measured on the general level and use on the media type level). This lack of correspondence does not suit the theory's rational choice foundations, and may in and of itself explain the modest correlations (Tsfati & Cappella, 2003).
The second caveat is that most albeit not all studies were done before the more recent transformation into media environments characterized by, among other things, greater abundance of alternative and partisan media, social media, and more frequent attacks on news media. Social media, in particular, add a new layer of diffusion, that includes journalists and other actors disseminating both mainstream and non-mainstream news content. Trust in the proximate disseminator (be it a journalist or a social media friend) potentially serves as a cue to the credibility of the both the information and the more distal disseminator (the news organization).
The third caveat is that almost all previous studies in this area rely on cross-sectional data. Hence, the direction of causality – i.e. the extent to which media trust influences media use versus media use influences media trust – is not clear. What limited research there is addressing this suggests that the association between media trust and media use runs in both directions (Ladd, 2012; Tsfati, 2002), with some evidence suggesting that for traditional news media, the causal path goes from media use to media trust, while for non-mainstream media, the causal path goes in the other direction (Tsfati, 2002). Perhaps most interesting in this respect is the instrumental approach taken by Ladd (2012), who used respondents’ average daily commute as a predictor of exposure to talk radio and found that it predicted mistrust in traditional news media. One explanation for this pattern might be the strong anti-media rhetoric of many non-mainstream or partisan media (Tsfati, 2002; Jamieson & Cappella, 2008), leading users to trust and use traditional news media less. More research using longitudinal data is however needed before firm conclusions can be drawn with respect to the causal and reciprocal relationship between media trust and media use. One theoretical framework that would be highly suitable in that context would be the reinforcing spirals model (Slater, 2007).
Towards a framework for studying media trust and its impact on media use
Based on the review above, our assessment is that there is a need for further research both on news media trust in itself and on its causal relation to media use in contemporary high-choice media environments. Such research should furthermore be theory-driven and based on a solid conceptualization of and, ideally, validated measurements of media trust. The same, of course, applies for measures of media use, although it is beyond the scope of this article to address that.
In terms of conceptualizing media trust, one important aspect is related to the concept of media. As shown, sometimes questions tapping media trust refer to unspecified (news) media in general, sometimes to different media types, sometimes to media as institutions, sometimes to individual media outlets, sometimes to journalists, and sometimes to the content or topic of media coverage. How (news) media trust at these different levels of analysis are related to each other remains unclear, and is hence an important research problem that should be addressed in future research. Equally important is to be clear about what level of analysis (news) media trust refers to. Towards that end, in Figure 1 we propose a framework for conceptualizing (news) media trust at different levels of analysis.
Conceptualizing media trust at different levels of analysis.
The starting point for this framework is the individual users, the trustors. At the very bottom we have placed news media content. This refers to the news stories that individual users, in the most concrete terms, are exposed to when using different media. In terms of measurements, this level could refer to ‘news from’ the news media when covering different topics. Climbing up the ladder of abstraction, we have placed trust in journalists, as they are the ones closest to the production of the news that the media report. These are also quite visible to users and hence quite concrete. Journalists are in turn nested within different media organizations and outlets, which we refer to as individual media brands in the figure. These, in turn, belong to different types of media, such as television or newspapers. This level corresponds to the institutional level. Then, at the highest level of abstraction, we have placed news media in general. It is a deliberate choice to here talk about news media rather than media in general, as the diversity of the latter category makes it close to meaningless in terms of measuring trust.
The arrows at each side run in both directions, to illustrate that trust at one level of analysis might influence trust at a higher or lower level of analysis, although to date there is insufficient research to know if, and if so how or to what extent, trust at different levels of analysis influences trust at another level.
Another important aspect is related to what is meant by trust, aside from the broader notion that media trust refers to the relationship between citizens (the trustors) and the media (the trustees) where citizens, in situations of uncertainty, expect that interactions with the news media will lead to gains rather than losses. One key problem in previous research is that the measures used often leave the meaning of trust unspecified, which leaves it to respondents to interpret the concept of trust. In extension, this means that respondents might have quite different things in mind when responding to the very same questions of media trust. To the extent that trust is specified, the focus is usually on the objects of trust (see above review). While specifications are preferable compared to leaving the meaning of trust unspecified, as they narrow down the degree of freedom when interpreting the questions, there are however important conceptual differences between, for example, ‘the people running’ media institutions and ‘reporting the news fully, accurately and fairly.’ The end result is not only that results are quite hard to interpret in substantial terms – as we do not know what people mean when responding that they trust or do not trust media at different levels of analysis – but also insufficient comparability across studies and hence cumulativity of findings (c.f. Esser et al., 2012).
In order to increase conceptual clarity, comparability across studies, cumulativity of findings, and our understanding of how news media trust matters both in general and for people's news media use, what is needed is a specification of news media trust that both (a) stays close to the broader definition of trust and (b) the specific nature and function of news media in democratic societies. From that perspective, it follows that the focus should be on trust not in media as institutions or in the people running media institutions, but trust in the information coming from news media at different levels of analysis. There are several reasons why this should be the focus. First, a key democratic function of news media is to provide people with the kind of information they need to ‘be free and self-governing’ (Kovach & Rosenstiel, 2014, p. 12; McQuail, 1992; Patterson, 2013; Strömbäck, 2005). That presupposes that the information provided corresponds to reality and is factually verified. Hence, Kovach and Rosenstiel call journalism ‘a discipline of verification,’ which ultimately is what sets journalism apart from other types of media content such as, for example, entertainment, propaganda, or art (Kovach & Rosenstiel, 2014, p. 71). Second, within the journalistic community, providing factual and reliable information to the public is generally thought of as the most important role of journalism (Hanitzsch et al., 2011; Weaver & Willnat, 2012). Thus, focusing on trust in the information reported by news media at different levels of analysis offers a way of investigating the extent to which news media is perceived as living up to its ideals according both to democratic theory and the journalistic community at large. Third, accepting the veracity of the information is the risk people take when they consume news. When people act upon this information in their daily lives (when voting, buying or selling stocks, planning trips etc.) they risk taking the wrong decision, and this risk is the most central element in the definition of trust (Coleman, 1990; Gambetta, 1988; Mayer et al., 1995). Fourth, it is no coincidence that political actors and non-mainstream or partisan media that attack traditional news media for delivering ‘fake news’ or otherwise seek to undermine the legitimacy of traditional news media focus on the trustworthiness of the information coming from traditional news media (Benkler et al., 2018; Egelhofer & Lecheler, 2019; Jamieson & Cappella, 2008). That illustrates that what matters is ultimately not the other roles of news media as institutions or organizations but the information coming from these. Fifth, virtually all research on news media trust and news media credibility in fact shares an understanding that what matters is trust in the information coming from news media (Gaziano & McGrath, 1986; Kohring & Matthes, 2007; Ladd, 2012; Tsfati & Cappella, 2003; Yale et al., 2015), although that is not always sufficiently mirrored in the measurements used.
Beyond these normative and logical reasons, the most important reason for the proposed focus on information is theoretical. The theory connecting trust in media and news use assumes (Tsfati & Cappella, 2003, Assumption 2, p. 508) that the main motivation for spending time and energy on news consumption is the necessity of acquiring information about the non-immediate environment. While this information does not have to be full or fully accurate, it is required for adjusting our behavior to changes in the environment in order to reach our goals. As information is at the core of the trust-news-use theoretical proposition, its incorporation into the measurement of the theory's main construct is required.
Based on this, we propose that basically all measures of trust in news media should specify that what matters is people's trust in the information coming from news media, regardless of the level of analysis. More specifically, in Table 1 we offer our proposed specified measures of trust in the information coming from news media at each level of analysis.
Proposed measures of trust in information coming from news media at different levels of analysis.
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you trust information from the news media in [country]?’ |
| Media Type | ‘Generally speaking, to what extent do you trust information from the following media types in [country]?’ (e.g. newspapers, radio, television, online news sites etc,) |
| Individual media brands | ‘Generally speaking, to what extent do you trust information from the following media?’ (e.g. CNN, Fox News Network, Wall Street Journal, MSNBC) |
| Journalists | ‘Generally speaking, to what extent do you trust information from the following groups of people?’ (e.g. journalists, politicians, scientists, religious leaders) |
| Media content | ‘Generally speaking, to what extent do you trust information from the news media in [country] when they cover the following topics:’ (e.g. economy, foreign affairs, health, the environment, crime) |
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you trust information from the news media in [country]?’ |
| Media Type | ‘Generally speaking, to what extent do you trust information from the following media types in [country]?’ (e.g. newspapers, radio, television, online news sites etc,) |
| Individual media brands | ‘Generally speaking, to what extent do you trust information from the following media?’ (e.g. CNN, Fox News Network, Wall Street Journal, MSNBC) |
| Journalists | ‘Generally speaking, to what extent do you trust information from the following groups of people?’ (e.g. journalists, politicians, scientists, religious leaders) |
| Media content | ‘Generally speaking, to what extent do you trust information from the news media in [country] when they cover the following topics:’ (e.g. economy, foreign affairs, health, the environment, crime) |
Proposed measures of trust in information coming from news media at different levels of analysis.
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you trust information from the news media in [country]?’ |
| Media Type | ‘Generally speaking, to what extent do you trust information from the following media types in [country]?’ (e.g. newspapers, radio, television, online news sites etc,) |
| Individual media brands | ‘Generally speaking, to what extent do you trust information from the following media?’ (e.g. CNN, Fox News Network, Wall Street Journal, MSNBC) |
| Journalists | ‘Generally speaking, to what extent do you trust information from the following groups of people?’ (e.g. journalists, politicians, scientists, religious leaders) |
| Media content | ‘Generally speaking, to what extent do you trust information from the news media in [country] when they cover the following topics:’ (e.g. economy, foreign affairs, health, the environment, crime) |
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you trust information from the news media in [country]?’ |
| Media Type | ‘Generally speaking, to what extent do you trust information from the following media types in [country]?’ (e.g. newspapers, radio, television, online news sites etc,) |
| Individual media brands | ‘Generally speaking, to what extent do you trust information from the following media?’ (e.g. CNN, Fox News Network, Wall Street Journal, MSNBC) |
| Journalists | ‘Generally speaking, to what extent do you trust information from the following groups of people?’ (e.g. journalists, politicians, scientists, religious leaders) |
| Media content | ‘Generally speaking, to what extent do you trust information from the news media in [country] when they cover the following topics:’ (e.g. economy, foreign affairs, health, the environment, crime) |
Of course, oftentimes when not having to rely on secondary data, it is more appropriate to use multiple items to measure trust, and use those to build a composite trust scale. As discussed above, in previous research a number of different multidimensional scales, made up of more or less different measures, have been proposed and in some cases validated (Abdulla et al., 2004; Gaziano & McGrath, 1986; Kohring & Matthes, 2007; Meyer, 1988; West, 1994). As shown by recent studies (Prochazka & Schweiger, 2019; Yale et al., 2015), it has however been difficult to establish with sufficient discriminant validity that news media trust consists of several different subdimensions. Yale et al. (2015) thus note that ‘people seem to evaluate news credibility more heuristically’ (p. 167) and recommend that any scale should be treated as a single-factor measure when used as a variable (p. 166). Similarly, Prochazka and Schweiger (2019) note that ‘the current path in trust and credibility research to find underlying factor structures of the concepts might be partly misleading’ (p. 40). Based on these considerations, we propose a focus on the perceptions of news when using multiple items to measure media trust. Here, the credibility items from the Gaziano and McGrath (1986; see also Meyer, 1988; West, 1994) News Believability Scale offer a standardized, quite efficient and established approach for focusing on trust in the information coming from news media when using multi-item batteries (see also Yale et al., 2015). Convergent and discriminant validity of the Gaziano and McGrath items has also been demonstrated (Tsfati, 2003, p. 73). Compared to the scale suggested by Kohring and Matthes (2007), this approach requires fewer items and is thus more efficient. In addition, the items can be adapted to the measurement of media trust on each level of analysis (see Table 2), although for efficiency reasons, they might be more relevant to use when measuring news media trust in general, trust in different media types or trust in journalists than when measuring at the other levels of analysis. In line with our emphasis on trust in the information coming from news media, we suggest however that the items should be adapted to explicate that the statements deal with the news media's coverage rather than the news media as organizations or institutions.
Using multiple items to measure media trust at different levels of analysis.
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the news media in [country]?’
|
| Media Type | ‘Generally speaking, to what extent do you agree or disagree with the following statements about newspapers/television/radio in [country]?’
|
| Individual media brands | ‘Generally speaking, to what extent do you agree or disagree with the following statements about [media x, y, z, etc.]’?
|
| Journalists | ‘Generally speaking, to what extent do you agree or disagree with the following statements about journalists in [country]?’
|
| Media content | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the media's coverage of [topic x, y, z, etc.]?’
|
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the news media in [country]?’
|
| Media Type | ‘Generally speaking, to what extent do you agree or disagree with the following statements about newspapers/television/radio in [country]?’
|
| Individual media brands | ‘Generally speaking, to what extent do you agree or disagree with the following statements about [media x, y, z, etc.]’?
|
| Journalists | ‘Generally speaking, to what extent do you agree or disagree with the following statements about journalists in [country]?’
|
| Media content | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the media's coverage of [topic x, y, z, etc.]?’
|
Using multiple items to measure media trust at different levels of analysis.
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the news media in [country]?’
|
| Media Type | ‘Generally speaking, to what extent do you agree or disagree with the following statements about newspapers/television/radio in [country]?’
|
| Individual media brands | ‘Generally speaking, to what extent do you agree or disagree with the following statements about [media x, y, z, etc.]’?
|
| Journalists | ‘Generally speaking, to what extent do you agree or disagree with the following statements about journalists in [country]?’
|
| Media content | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the media's coverage of [topic x, y, z, etc.]?’
|
| Level of analysis | Proposed measure |
|---|---|
| News media in general | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the news media in [country]?’
|
| Media Type | ‘Generally speaking, to what extent do you agree or disagree with the following statements about newspapers/television/radio in [country]?’
|
| Individual media brands | ‘Generally speaking, to what extent do you agree or disagree with the following statements about [media x, y, z, etc.]’?
|
| Journalists | ‘Generally speaking, to what extent do you agree or disagree with the following statements about journalists in [country]?’
|
| Media content | ‘Generally speaking, to what extent do you agree or disagree with the following statements about the media's coverage of [topic x, y, z, etc.]?’
|
The key point is that the focus should be on trust in the information coming from news media regardless of the level of analysis and, when using multiple items to form a trust scale, adapting items used to measure the concept that comes closest, which is (perceptions of) the credibility of news. This would allow scholars to get a better understanding of the extent to which people trust the news media in the dimension that matters the most from the perspective of democratic theory as well as the journalistic community and those who seek to delegitimize the media, while staying true to the core of previous research on news media trust. In extension, the framework proposed here would also allow for more fine-grained and focused research on whether, how and to what extent media trust at different levels of analysis is related and how it influences citizens’ use of news media. In addition, we believe it would also open up new avenues for research on the reasons that people have for not trusting the news media (c.f. Newman & Fletcher, 2017).
Discussion and conclusions
Following both the public debate and the scholarly literature, it is often assumed that media trust is falling virtually everywhere. The perhaps most important reason is the long-term trend with respect to Americans’ shrinking trust in the ‘press,’ which then is extrapolated, and that people – even scholars – tend to infer the accuracy and consensus of opinion from the number of times it has been repeated (Weaver et al., 2007). The fact though is that media trust – in terms of levels as well as trends – differs across countries (Newman et al., 2019; Hanitzsch et al., 2018).
Nevertheless, it is abundantly clear that many people do not trust traditional news media. It is also clear that the transformation into high-choice media environments has brought with it a host of new and exacerbated challenges threatening to undermine news media trust, regardless of whether it is already on decline or thus far has been more stable. And although there are important differences across countries and media environments, there are more so-called non-mainstream and partisan media that compete with traditional news media. Such non-mainstream media in many cases even actively engage in attempts to undermine trust in traditional news media – something which, to an increasing extent, is supported by politicians who also themselves perform attacks on news media and accuse them of producing ‘fake news.’ Where political polarization (Hopkins & Sides, 2015) and more or less authoritarian political populism (Aalberg et al., 2017; Moffitt, 2016; Norris & Inglehart, 2019) in addition is on the rise, it becomes even harder for traditional news media to stay above the fray and be seen as credible when claiming to provide news that is truthful and objective or neutral in intent (if not necessarily in consequences) and thus deserves to be trusted more than information from partisan actors (see Ladd, 2012, for a discussion regarding the connection between media trust and political polarization). In addition, there are – although the evidence is not consistent – indications that trust in various sets of experts and knowledge institutions is declining (Nichols, 2017), among which news media is but one. In combination with people's tendency to prefer attitude-consistent information and engage in motivated reasoning (Flynn et al., 2017; Kunda, 1990; Lewandowsky et al., 2012; Taber & Lodge, 2006), this might contribute not only to less use of news media but also more widespread misperceptions and increasing knowledge resistance (Klintman, 2019).
In light of this, it has become increasingly important to investigate and understand not just news media trust in itself, but also if, how and to what extent it influences people's use of traditional news media and other types of information more broadly. Research suggests a reciprocal relationship between media trust and news media use with modest correlations, but the number of studies is limited, many were done before the transformation into contemporary high-choice media environments, the measures used were arguably imperfect, and in addition inconsistent within and across studies as to the level of analysis of the object of trust. Furthermore, these studies were mainly based on cross-sectional data with all the limits that such data entail in terms of understanding causal relationships.
To help remedy this situation, in this article we have offered a focused review of how news media trust has been conceptualized and operationalized in previous research. We have also offered a theoretically derived framework and specific measures for investigating news media trust at different levels of analysis. An important aspect of this framework is our suggestion to focus on trust in the information coming from news media rather than on media as institutions or organizations. Altogether, we believe that our proposed framework will allow research not only to investigate how news media trust at different levels of analysis are related to each other, but also if, how and to what extent news media trust at different levels of analysis are related to news media use. It might very well be the case that news media trust at one level of analysis does not have much of an impact on news media use, while news media trust at another level has a more significant impact. The same, of course, holds true for the effects of media use on news media trust. It could also be the case that the associations between media trust and use will be considerably higher when both constructs (trust and news consumption) will refer to specific referents at specific levels. Finally, it could be the case that trust in the media coverage of certain topics – for example, more politically polarized topics – have more influence on media trust at higher levels of abstraction than trust in the coverage of other topics.
While we believe the offered framework and measures represents an important step forward in research on news media trust and its impact on media use, ultimately it needs to be investigated+empirically before it can be evaluated. In light of that, our hope is that this article will help stimulate more, and more theory-driven, research in this area. The game is afoot, to quote Shakespeare.
Notes
The review was systematic in the sense that we systematically searched for relevant publications using ‘media trust,’ ‘media trustworthiness’ and ‘media credibility’ as keywords, but the review presented here is focused in the sense that we focus on those articles that are most relevant in terms of how news media trust has been conceptualized and operationalized and on the extent to which news media trust influences the use of different types of media.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
This work is part of the research project 'Knowledge Resistance: Causes, Consequences, Cures,' funded by Riksbankens Jubileumsfond for the Advancement of the Humanities and Social Sciences.
ORCID
Jesper Strömbäck http://orcid.org/0000-0001-7392-9791
Yariv Tsfati http://orcid.org/0000-0001-7395-3257
Hajo Boomgaarden http://orcid.org/0000-0002-5260-1284
Alyt Damstra http://orcid.org/0000-0001-7753-018X
Elina Lindgren http://orcid.org/0000-0002-9032-1347
