Dispersing the Opacity of Transparency in Journalism on the Appeal of Different Forms of Transparency to the General Public

This study investigates the kinds of transparency that appeal to di ﬀ erent parts of the public and the extent to which transparency can be a remedy for declining trust in journalism. It uses a representative survey of Swedes, and the results show that there are three distinct forms of transparency, including the previously unreported ambient transparency, and that they appeal to di ﬀ erent people. News consumption or social media use has little or no e ﬀ ect on transparency. The strongest positive e ﬀ ect on transparency comes from appreciation of the current quality of journalistic performance, high trust in journalists and media, and having news media and authorities as the preferred channels of information. Those most skeptical about journalism are also least positive about transparency. The results suggest that transparency has very limited reach as a cure for declining trust in, and the trustworthiness of, journalism, possibly since the acts of transparency themselves remain non-transparent.

Transparency in journalism, broadly speaking the idea that news organizations are open to the public in how news are being made, has been intensively discussed within journalism studies and practice over the last 20 years. The ongoing digitalization of journalism has only added to its relevance. Digital publishing has both created a need for transparency in journalism, due to speed-induced accuracy problems, and increased the potential for user-driven criticism, but it also provides unlimited space in which to correct and explain publishing decisions, as well as engaging in dialogue with the public (Karlsson 2011). These discussions are also heavily linked to social media platforms beyond the control of traditional news media (Lasorsa 2012;Lasorsa, Lewis, and Holton 2012;Revers 2014), prompting Revers (2014, 823) to view Twitter "as a carrier of the ethic of transparency".
Transparency in journalism has been embraced by academics, journalists and the public alike. Journalists have elevated it to a central norm in contemporary journalism and it is so established that it has been included in ethical/professional guidelines (Vos and Craft 2017). For academics it is a tool to (re)build relationships with the public and a awareness. Research implications for transparency extend into the heart of what journalism studies and practice are and should be.
The idea of transparency as a way to improve journalism and form a stronger relationship with audiences has been at the forefront of journalism studies and practice for at least 20 years. The most detailed, extended and authoritative account of what they call the rule or spirit of transparency is perhaps made in Kovach and Rosenstiel's highly cited book, the first edition of which was printed in 2001, The Elements of Journalism: What Newspeople Should Know and the Public Should Expect (Kovach and Rosenstiel 2001;Kovach and Rosenstiel 2007). This spirit both invites dialogue with the public, and is about telling and showing the public how journalism is done. Kovach and Rosenstiel predict that in the end this is a way to reconnect with the public. Other researchers have built on this research and suggested that transparency can be linked to increased trust, trustworthiness, accountability and authenticity (Carvalho, Chung, and Koliska 2018;Karlsson 2010;Karlsson 2011;Morton 2015;Phillips 2010;Singer 2007Singer , 2005, that a more transparent journalism requires new practices and skillsets among journalists (Bivens 2008), and that transparency can be seen as a defensive move to preserve journalistic authority (Allen 2008;Karlsson 2010).
The next two sections detail the antecedents of this thinking and the prerequisites for this to also work in practice.

Transparency as Uncertainty Reducer and Trust Increaser
Uncertainty and vulnerability are preconditions for trust, because if people really know about the state of things, then they would not have to be vulnerable and trust others to inform them (Hall et al. 2001; Lewis and Weigert 1985;Mayer, Davis, and Schoorman 1995). Trust thus serves the dual function of reducing uncertainty and replacing knowledge at the same time (Kohring and Matthes 2007).
Uncertainty is also an inevitable aspect of news consumption, as news consumers cannot usually verify the events themselves (Tsfati and Cappella 2003). Transparency, then, can be considered a tool to allow the public to increase their knowledge about how journalism gets done, thereby decreasing uncertainty and, it is hoped and predicted, decreasing the risk involved in trusting journalism. As Kovach and Rosenstiel argue (2007, 92), "It [transparency] allows the audience to judge the validity of the information, the process by which it was secured, and the motives and biases of the journalist providing it". Transparency, ideally, reduces the risk one takes in trusting others. Transparency could thus appeal to low-trust members of the public who, all other things being equal, do not trust journalism and the news media under regular conditions, but may eventually, find journalism more trustworthy through acts of transparency. In order to better understand the role of journalistic transparency in relation to media trust, we therefore need to scrutinize how members of the public view it. But what do people trust in others, when they trust? How is a transparent journalism connected to that "what"? A closer look is needed at journalistic performance in order to address those questions.

Linking Trust to Transparent Performances, Countering Vulnerability
Trust in journalism is usually linked to institutions and people; for example, the extent of which the public (the trustor) trust in journalists or news media organizations (the trustee).
While trust can be general, it is also a question of what the public trust journalists to do specifically (c.f. Mayer, Davis, and Schoorman 1995, 729). The vulnerability underpinning trust is linked to expectations that " … the other will perform a particular action important to the trustor … " (Mayer, Davis, and Schoorman 1995, 712; see also Rousseau et al. 1998) according to certain standards. It is a pointed vulnerability that cannot be alleviated by blind trust. Relatedly, Lewis and Weigert (1985, 981) argue that trust in profession-client relationships must be built on "an adequate symbolic representation of the competence and integrity of the professional" (e.g., a performance). Trusting, as something a member of the public does, then, is related to the evaluation of expected performances and characteristics by, in this case, professional journalists. An evaluation is essentially asking the questionare these performances trustworthy, based on the currently available information, to such an extent that I can let my guard down and be willing to vulnerable to the unmonitored and uncontrolled actions of those journalists or media outlets (c.f. Mayer, Davis, and Schoorman 1995)? This turns attention to the performative dimension of building trust in highlighting how the job gets done, and that trust "is a product of behaviours" (Fletcher andPark 2017, 1283). Because transparency is essentially about doing journalism differently, this inevitably means changing journalistic performance, and how it appears to the public through various "rituals of transparency" (Karlsson 2010). The key theoretical question underlining this whole strand of research is to what extent a more transparent journalistic performance affects the trustworthiness of journalism leading to an increased trust in journalism by the public?
In this context it will be interesting to probe the relationship between attitudes towards the standards of current journalistic performance and transparency. Since there is an implicit suggestion in previous theorizing that those critical of how journalism is currently done (e.g., non-transparent, falling short of fulfilling its societal purposes) will be positive about a more transparent journalism.
The remainder of the theory section will turn its attention to previous research that has tried to unpack different forms of transparency in journalism, how members of the public view journalistic transparency, and its proposed interweaving with social media platforms.

Forms of Transparency
Up until this point, many scholars have felt compelled to theorize about what transparency can do for journalism, and there is no shortage of hope and predictions that transparency is a medicine for journalism trust issues vis-a-vis the public.
Neither term, however -"transparency" or "public"is always explicitly granulated to a level where they can be investigated empirically, but there is growing body of research seeking to define and investigate transparency in journalism. A commonly used distinction divides the general openness that transparency proposes into two specific formsdisclosure and participatory transparency (Karlsson 2010). These forms will be detailed in the method, but disclosure transparency involves various techniques illustrating to the public how and why the news is being made, by, for instance, explaining the news selection process and so opening up news production for insights. Participatory transparency is understood as inviting outsiders (such as the audience) to partake in various stages of the news production by, for example, commenting or sending in images of events. So far, the understanding of different forms of transparency has been advanced from what journalists and news media do. Should the point of departure instead be from the public, then new, other, and conflicting forms of transparency might emerge.

A Lack of Perspectives from the Public
Despite a large and growing body of studies, research on transparency in news rarely involves the public This is problematic as neither journalism in general, nor transparency in particular, can be properly understood without taking into consideration the news consumers " … as active and intrinsic components of journalism as a cultural practice" (Carlson et al. 2018, 12). In other words, the public have a stake in determining what is considered proper journalism. Should various levels of public anticipation not be met or swayed, then journalism cannot possess or exercise authority (c.f. Carlson 2017; Karlsson and Clerwall 2019;Sjøvaag 2010). Whilst there is a shortage of studies there are some that offer advice.
When investigating the article topics and political leanings, an experimental US study found that transparency had a small positive effect on trustworthiness (Curry and Stroud 2019), however, another US-based experiment found that objectivity produced greater trustworthiness than transparency (Tandoc and Thomas 2017). In Europe, survey studies in the Netherlands have shown that the public is relatively positive towards participatory transparency in news (without specifically using that term), but that only a minority is interested in actually being involved in making the news (van der Wurff and Schoenbach 2014; van der Wurff and Schönbach 2014). van der Wurff and Schönbach (2014) also reported that activities related to disclosure transparency (e.g., admitting mistakes, explain how news items are made) are appreciated by the public. The key take-away from studies conducted in Sweden (Karlsson and Clerwall 2018;Karlsson, Clerwall, and Nord 2014) is that transparency in news is not a big issue for the public, but that they are more appreciative of disclosure than participatory transparency.
With one exception (Curry and Stroud 2019) there have been no similar studies on transparency. Curry and Stroud (2019) investigated the moderating effects on transparency and found that neither article topic, nor participant's political inclination had an impact on how the participants viewed transparency. It is still reasonable to propose that transparency will be influenced by different forms of publics, however, for instance based on demographics and attitudes, as we know from studies on media credibility in general that there is a correlation between credibility assessment and media use (Curry and Stroud 2019;Kiousis 2001) where those more skeptical towards news media will use it less (Tsfati and Cappella 2003). Gender, age and education have also been investigated and occasionally found to matter in relation to media exposure and credibility (Tsfati and Ariely 2014;Tsfati & Capella 2003). From research on governmental transparency we know that, for instance, internet use, age, gender, education, frequency of contacting government, confidence in, or suspiciousness towards, governmental officials are important determinants explaining citizens attitudes and demands towards different forms of transparency (Cuillier and Piotrowski 2009;Piotrowski and Van Ryzin 2007). Although the forms of transparency measured in that research area are not directly translatable to the context of transparency in journalism the results do point to important theoretical dimensions in explaining individual's relation towards transparency performed by societal institutions.
Generally, previous empirical studies involving the public show that transparency in journalism has a small effect on how the public views the trustworthiness (primarily source and message credibility) of journalism.

Information Distribution and Performing Transparency on Social Media Platforms
The changed media landscape entails a higher reliance on various platforms for news/ information distribution, meaning that people will search for, and find, information from various actors, both journalistic and non-journalistic. Building on the relationship between different forms of publics and the connection to news exposure, media diets and media skepticism (e.g., skeptics seeking alternatives to legacy media) observed in previous research (Fletcher and Park 2017;Tsfati and Cappella 2003), it would also be pertinent to examine the relationship between the places people seek information and their views on transparency in news. The wealth of information sources online means that is much easier to find alternatives to mainstream media. Theoretically, it is interesting to see whether people's preferred source of information is related to journalistic transparency. For instance, if there is a positive relationship between attitudes towards transparency in journalism and an inclination to search for information outside the journalistic realm, then a more transparent journalism might appeal, enticing them to go (back) to news outlets.
Related to changing patterns in information distribution, some previous research (Lasorsa 2012;Lasorsa, Lewis, and Holton 2012;Revers 2014) has underlined the role of social media, especially Twitter, in relation to transparency, pointing to Twitter "as a carrier of the ethic of transparency" (Revers 2014, 823). Transparency in journalism is thus to some extent performed on social media platforms and it is therefore pertinent to explore the relationship between attitudes to social media, social media use and transparency. According to previous research, there is an implicit theoretical proposition predicting a positive correlation between the use of, and attitudes towards, social media and journalistic transparency, as the positive virtue of transparency is performed on social media, visible to people who are positive towards, and using social media. Following this, it is of interest to investigate the relationship between preferred media use (e.g., news consumption, social media use, preferred channel of communication) and transparency in journalism.

Research Questions
Although the previous research reports that the public are not exactly enthused to find transparency there are at least three reasons to continue empirical research into how the public views transparency in journalism. First, in the long term, and since journalistic practice and norms have changed before, journalistic transparency can be established as a new norm if other factors in society work towards that goal. Journalism does not invent its norms, ethics or professional skills in a vacuum (Ward 2019), but builds on what is available and acceptable within a specific historical and cultural context. Second, although the effects of transparency in journalism are minor and it does not seem to be at the forefront of people's minds, when there is a transparency effect, it is generally positive. Third, despite the small return that research has yielded so far, there is still persistent interest in, and hope tied to, transparency in journalism, as outlined in the literature noted above.
Previous research is not able to guide us to propose hypotheses about how media trust will affect views on transparency in journalism but it does indicate trust as a good place to start looking for differences. In particular, it is interesting to see whether some transparency techniques are more appealing to low-trust publics than others. After all, it is those people that transparency is supposed to convert from skeptics to believers, and for that to happen, transparency must appeal to those with less trust in news media. Drawing on the theoretical framework and previous research outlined above, the following research questions are posed: RQ1 What is the relationship between age, gender, education and attitudes towards transparency in journalism?
RQ2 What is the relationship between trust in media (RQ2a), the assessment of journalistic performance (RQ2b) and attitudes towards transparency in journalism?
RQ3 What is the relationship between attitudes towards social media (RQ3a), social media use (RQ3b) and attitudes towards transparency in journalism?
RQ4 What is the relationship between preferred channel of information and attitudes towards transparency in journalism?
RQ5 What is the relationship between news consumption and attitudes towards transparency in journalism?

Method
A representative Swedish national survey (N = 2091) was used to inform the research questions and is a part of a larger research project on transparency and credibility consisting in total of 91 questions spread over 19 areas of enquiry (all not reported here). The mean time to answer the survey was 12 min. Some parts of the material have been presented in previous publications where the method is discussed in more detail (Karlsson and Clerwall 2018;Karlsson, Clerwall, and Nord 2017). The data were collected in cooperation with polling company TNS/Sifo during March 10th and 24th 2014 using their panel of a representative sample of Swedes. Invitations to the online survey were sent to 7918 persons to ages 16 and upwards. The polling company offers an incentive program for their panelists (e.g., collecting credits to either give away to charity or receive a small remuneration such as movie tickets and gift vouchers). A 26.4% response rate produced 2091 answers. The respondents were slightly older (M = 51, SD = 16) than Swedes in general and there were also more women responding than men (52% women, 48% men).
The survey builds on previous work (Karlsson 2010) that has identified disclosure and participatory transparency as distinct expressions of a general openness. The respondents were asked to express their attitudes towards 12 transparency techniques serving as dependent variables. Nine variables were connected to disclosure transparency (e.g., explaining and detailing the news production process) asking the respondents to voice their opinions on a 7-point Likert-type scale (1-very bad to 7-very good) regarding what they thought about when news media: (1) Tells them there is an error is news stories (M = 4.80, SD = 1.90), (2) Explains why there was an error (M = 4.49, SD = 1.96), (3) Notes when a news story was published or updated (M = 4.56, SD = 1.52), (4) Explains why a certain news story was published (M = 4.11, SD = 1.77), (5) Explain why a particular framing was used in a news story (M = 3.91, SD = 1.95), (6) Lets journalists personal opinions be expressed in the news stories (M = 3.45, SD = 1.54), (7) Let journalists personal opinions be expressed at other places on the news sites (but not in news stories) (M = 3.83, SD = 1.53), (8) Use hyperlinks to other content on the website (e.g., internal linking) (M = 4.01, SD = 1.55), (9) Use hyperlinks to link to content on other websites (e.g., external linking) (M = 4.04, SD = 1.51).
Three variables were linked to participatory transparency (e.g., opening up news production for people outside the news organization to take part in), again on a 7-point Likert-type scale (1-very bad, 7-very good), where the respondents were asked to show what they thought about letting members of the audience: (10) Comment on news stories on news sites (M = 4.45, SD = 1.55), (11) Have their images published on news sites (M = 3.98, SD = 1.48), (12) Publish their own news stories on news sites (M = 3.30, SD = 1.56).
The independent variables (means and standard deviation for the independent variables with ordinal scale can be found in Appendix) are in two groups, where the first are basic demographic variables: gender, age (years), education (in the Swedish school system this follows the division: 9 years, 11 years, 12 years, +2 years at university, +3 years at university). The other group of independent variables is connected to different forms of media use and attitudes towards media and journalism. The respondents were asked about their local and national news use (days per week), use of social media such as Facebook, Twitter, Instagram, YouTube (days per week), and their preferred channel of informationfrom news media, social media, interpersonal communication or authoritiesin three different hypothetical scenariosregarding riots and civil unrest, storm warnings, and the head of a governmental body using tax money for private consumptionon a 7-graded scale (1 = not at all probable, 7 = very probable). The scenarios are different in character as riots and civil unrest is uncommon (although not unheard of) the other to occur more frequently in Sweden. Moreover, storms have natural causes while corruption has a clear human origin. Together, the scenarios offer a chance to see if the respondents think they would act differently depending on the type of event they were facing.
Attitudes towards media were measured in several ways (with one exception, the 7point Likert scale was from 1do not agree at all to, 7fully agree). First, there was a straightforward claim suggesting that the news media can be trusted, and then a question (e.g., non-Likert scale) about the extent of their trust in Swedish journalists. Second, the respondents were asked to take a position about whether journalists live up to their professional standards, as this is an indication of media skepticism (Tsfati and Cappella 2003). They were asked to rate whether: (1) Journalists are good at informing the public about what is going on in society, (2) Journalists are good at scrutinizing those with power in society, (3) Journalists are good at reporting about all parts of society, (4) Journalists in general are competent to do their job, (5) Journalists in general are good at providing full coverage of events. Third, the respondents were also asked to rate the social media in general (7-point Likert scale going from 1do not agree at all to, 7fully agree) and whether they agreed or not that social media: (1) Gives me information quickly about what is going on, (2) Gives me a poor overview of different opinions on an issue, (3) Gives me a feeling of community with others who are like me, (4) Does not give me a credible view of what is going on, (5) Offers amusement and distraction when I want it.
This should provide an extensive assessment of the relationships usually explored in credibility studies (e.g., age, gender, education, trust and media use) as well as opening up related but new inquiries (e.g., journalistic performance, preferred channel of information, use of and attitudes towards social media), as detailed in the theory section.

Data Treatment
A factor analysis was conducted to see which of the dependent variables were interrelated (Table 1). Three dimensions were drawn from the original 12 dependent variables, suggesting that the previous understanding of disclosure and participatory transparency could be refined further.
Participatory transparency formed a dimension of its own, while what has previously been defined as disclosure transparency came out as two dimensions in the PCA. The PCA suggests that there is still a distinct disclosure transparency dimension that is very clearly connected to specific news stories (e.g., admitting there was an error in a news story, explaining why a news story was published, etc.). The second dimension, stemming from the previously identified disclosure transparency, contains variables that are not related to the news story per se, but can be argued as important for the context of the news story. Both journalist opinions and hyperlinks are tangential to the news story but could potentially have an effect on opinions. Inspired by Hermida's (2010) concept of ambient journalism (i.e., dispersed information and interactions that form an awareness system around journalistic content), it is proposed that this is a kind of ambient transparency, via peripheral information, that might be relevant to better understanding the news story but not actually part of the news story. A more elaborate account of this concept can be found in the discussion. Cronbach's alpha was tested for all three dimensions (Table 2) and was over the recommended value (0.7). All 12 variables formed an index, with a Cronbach's alpha of .822. This shows that there are three distinct dimensions of journalistic transparency, but that they also fit together under a larger transparency in journalism umbrella. Some of the independent variables could also be converted into indices (see Table 2). All four regressions in the results section (Table 3) passed multicollinearity tests. Will use news media to keep me informed -If there was a storm warning in your area, how probable is it that you will use news media to be informed? -If there was riots and civil unrest somewhere in Sweden, how probable is it that you will use news media to be informed? -If a head of a governmental body is suspected to have used tax payers money to pay for private consumption, how probable is it that you will use news media to be informed? .825 Will use authorities to keep me informed Same as above but news media swapped for authorities, social media and interpersonal communication.

Results
Four regression models were run to investigate the RQs (see Table 3); one for each of the three dimensions detected in the PCA and one with all 12 transparency variables in one index. A first observation is that different forms of transparency appeal to different publics. There is no instance where there is significant relationship between an independent variable and all three dimensions of transparency.
RQ1 asked about age, gender, and education in relation to transparency in journalism. Education is a negative predictor towards participatory transparency; the higher the education the less appreciative people are of participatory transparency. Transparency is also gendered, as women are more positive about disclosure and men more positive about participatory transparency. In the regressions, age had a small effect on ambient transparency, where older people hold a more negative view.
RQ2 probed the relationship between trust in the media (RQ2a), assessment of the quality of journalistic performance (RQ2b), and attitudes towards transparency in journalism. These all explained some variance in disclosure transparency, and a positive evaluation of journalistic performance had a rather strong effect on ambient transparency. Participatory transparency was not affected by media trust and views on journalistic performance.
Attitudes towards social media (RQ3a) did not impinge on disclosure transparency at all, while the idea that social media was useful and could be trusted was a positive predictor of participatory transparency. The regressions were also a positive effect from thinking social media was useful as regards ambient transparency. Social media use (RQ3b) did have some effect on attitudes towards transparency. Increased Twitter use was negatively associated with participatory transparency, while increased YouTube use was positively linked with it. Increased YouTube use had a positive effect on disclosure transparency, while Instagram use had a negative effect.
RQ4 asked about the relationship between the preferred channel of information vis a vis transparency in journalism. The regression analysis found that the use of news media and authorities had a positive effect on disclosure transparency. A willingness to turn to channels run by authorities had a positive effect on ambient transparency. The use of social media to keep informed had a positive effect on participatory transparency.
Finally, RQ5 asked about the relationship between news consumption and attitudes towards transparency in journalism. The regression analysis showed that consuming the national public service broadcasting news report "Rapport" had a negative effect on disclosure transparency and there was a positive effect from using the national quality press "Dagens Nyheter" on ambient transparency. There was a positive effect on participatory transparency from using news from the national commercial broadcaster TV4. No other media use had an effect on any form of transparency. Overall, the results suggest that there are indeed distinct forms of transparency in journalism, and that these forms appeal to different publics.

Dispersing the Opacity of Transparency
A key rationale for this study was to see what form of transparency in journalism appealed to what form of public, and whether journalism studies and practice have to granulate the understanding of both transparency and public further. The overall impression is a resounding "Yes!" Journalistic transparency is distributed along at least three different dimensions according to the public. The results also show that there is dispersion in what kind of transparency appeals, or does not appeal, to different parts of the public. Talking about transparency and the public as one-dimensional concepts is thus insufficient. There is no way of pleasing everyone. For instance, turning on features of participatory transparency pleases males with lower education and positive attitudes to social media, but risks displeasing highly educated women.
However, there is something that is even more interesting and detrimental in relation to the ability of transparency to cure media distrust than the complexities of transparency and the conflicting effects that it produces. Above all, the results distinctly show that a favorable view of transparency in journalism is produced by an appreciation of the current quality of journalistic performance, and high trust in journalists and media, and news media and authorities as the preferred channel of information. Those most skeptical of journalism are also the least positive about journalistic transparency. Using transparency in news is thus largely preaching to the choir. It will take some higher forms of acrobatics to use transparency to convert media skeptics. The present study suggests that transparency in news has very limited reach as a cure for declining trustworthiness.
That said, participatory transparency is the only dimension where having a pre-formed positive view about a social institution does not matter. This form of transparency is about reducing the role of journalists in producing the news, however, and can also be interpreted as a form of media skepticism (see also Fletcher and Park 2017). This is also a potential opening for channels that are normally closed, as a study by Fletcher and Park (2017) showed that people with low trust are both more likely to use non-mainstream media and more likely to engage in participatory practices. Participatory transparency is thus an opportunity to interact with people who are skeptical towards journalism, but it will take commitment from journalists who are already spread thin, and the outcomes of these interactions are not predictable or without risk. As the regression analysis in this study demonstrate, gender (female) and education (higher) had a negative effect on participatory transparency and from previous research we know that adversarial comments negatively affect the credibility of journalism (Anderson et al. 2014;Karlsson, Clerwall, and Nord 2014). Using participatory transparency must thus be considered a balancing act, and the net effect on trustworthiness might be negative. Future research should evaluate whether various participatory features can sway those who are negative about journalism without side effects.

Reducing One Leap of Faith by Asking for Another?
Journalism is marked by uncertainty, and so it needs to be trusted (Kohring and Matthes 2007;Tsfati & Capella, 2005). One has to take a leap of faith. The discourse on how transparency in journalism can increase trust is centered about reducing uncertainty by showing how journalism is done (Kovach and Rosenstiel 2001): in short, decreasing the leap of faith. While acts of transparency improve the ability to monitor or control the content of a news article, those acts of transparency themselves remain non-transparent, and hence, trust comes into play again. The results from this study suggest that the media skeptic news consumer still finds themselves taking a big leap of faith to trust the transparency measures that are supposedly decreasing the leap of faith in trusting journalism in the first place. Skepticism towards journalism might also be directed towards journalistic transparency, as it basically stems from the same source. The problem, it seems, is less about the use or absence of transparency, the journalistic performance per se, and more about attitudes concerning journalism as a social institution, and indicatively, other social institutions.
Trusting others is about taking risks. Transparency is about minimizing that risk or giving the idea that the risk is less. That prompts the question of how much and what kind of transparency is needed in order to lower the risk so much that a less trusting person would be willing to accept it (c.f. Schoorman, Mayer, and Davis 2007). Another related factor affecting the evaluation of a current performance might be previous trust violations and the time and measures taken to repair them. Trust might also be unrelated to a rational evaluation of performance, but instead affected by emotional attachments and moods (Lewis and Weigert 1985;Schoorman, Mayer, and Davis 2007). A combination of qualitative studies (e.g., talking to various representatives from the public) and elaborate experimental settings exploring different forms of journalistic transparency could inform that line of query.

Ambient Transparency
Karlsson (2010) proposed two, but according to the factor analysis, there are three distinct transparency dimensions that are also connected, as the Cronbach's alpha test demonstrated. In particular, the suggested ambient transparency needs further elaboration. In his seminal article on ambient journalism Hermida (2010, 301) discussed ambient journalism as an awareness system where " … value is defined less by each individual fragment of information that may be insignificant on its own or of limited validity, but rather by the combined effect of the communication" [author's italics]. While Hermida extends this awareness system to include many forms of information, including contributions from users, I propose to limit ambient transparency to information provided by the news producers, as the PCA distinguished this dimension of transparency from that of the users (e.g., participatory transparency). As previous literature (Karlsson 2011;Kovach and Rosenstiel 2001;Singer 2007) has pointed out, transparency is at heart a process anchored in the news media and led by journalists. In order to improve trust in journalism it is journalistic performance that must be central, not any bit of information that might be relevant. Following this, ambient transparency can be understood as techniques or tools that are used/ added by news producers in the vicinity of (news) content making it possible for news consumers to evaluate and form new meanings of news stories, through the association of content with the provided context. Such tools include, but are not limited to; hyperlinks, journalists' personal opinions, and marking whether something is a news story or native advertising. Ambient transparency techniques add information around the edges of news stories but disregard the public inside the frame of the news stories, and do not explain the content per se, as unlike participatory and disclosure transparency.
The results from this study suggest that this form of transparency is primarily appreciated by those who think that journalism is performing wellit is added value to something already seen positively. It is possible that, those appreciating this form of transparency are also actively searching for information in established networks, as they also turn to authorities as acknowledged nodes in the public sphere (DN is the flagship paper in Sweden), and, a reasonable interpretation suggests, use social media networks to corroborate information from media and authorities. Conversely, ambient transparency does not seem to appeal to those who are guarded about the "legacy" information system and, thus, is not a likely tool to convert skeptics.

It's Complicated -Journalistic Transparency and Social Media
Increased Twitter use is negatively connected to participatory transparency and not associated with any of the other forms of transparency. On the one hand, this makes it hard to see Twitter as a critical carrier of the transparency ethic that some previous research has indicated (Lasorsa 2012;Lasorsa, Lewis, and Holton 2012;Revers 2014), and rather the opposite. There is, however, as the regression analysis show, a positive effect on disclosure and participatory transparency from increased YouTube use Conversely again, there is negative effect from increased Instagram use on disclosure transparency. Clearly, social media cannot be lumped together in one category regarding how it relates to journalistic transparency, and more research is needed to unpack these relationships. Overall, however, the results show that social media plays the second and third fiddle to other factors, and that, as a result, future research on transparency should treat them accordingly.

Contribution and an Alternative Research Agenda for Transparency and Trust in Journalism
This study contributes to the scholarly discussion about transparency in journalism in several ways. It has to some extent unpacked the complexity of transparency, showing that it is not a one-form-fits-all solution. It has in fact illustrated the complexity between different forms of transparency and different forms of publics. More specifically, it has shown that journalistic transparency is a poor remedy for distrust regarding the media, possibly because the transparency measures themselves must be trusted. It has pointed to the key importance of starting with the public to advance knowledge about the role that transparency can play in contemporary journalism. In doing so, it has suggested, through the example of ambient transparency, that there are additional dimensions of transparency to explore and discover.
Currently, much of the research into journalistic transparency and trust/trustworthiness is examining how journalism is open in its practice, and departs from the way in which scholars and journalists understand transparency. There is an alternative and, so far, untraveled path forward. Previous research has shown us that trust is connected to performance (Fletcher and Park 2017;Hall et al. 2001;Mishler and Rose 2001). This performance takes place in front of the public who can evaluate the quality of and approve/disapprove the performance or, in other words, the trustworthiness of the performance. We also know that factors such as ability (skills, competencies and characteristics), benevolence (the intention to do good to the trustor) and integrity (adhering to a set of mutually accepted principles) have been validated as important antecedents of trust in many research areas (Mayer, Davis, and Schoorman 1995;Schoorman, Mayer, and Davis 2007). If these preconditions are correct, then, in order to understand the potential role of transparency in journalism, I propose that the following needs to be done.
First, we need to further explore what the public identifies as, and expects of, specific journalistic performances of trust dimension factors, such as, (but not limited to) ability, benevolence and integrity. Having an open approach regarding how the public views this is extremely important, as their expectations are linked to vulnerability that, in turn, is linked to trust. Put differently, the public, or more aptly publics, will evaluate whether their "positive expectations of the intentions or behavior" (Rousseau et al. 1998, 395; see also Mayer, Davis, and Schoorman 1995) are delivered by the journalists or not. The outcome of that evaluation will decide whether they feel let down and vulnerable or not, which in turn, will affect their trust in journalism. These expectations may not be very different from the way that journalists and scholars have come to understand them, as indicated by previous research where the public seems much aligned with journalistic norms and practices (Karlsson and Clerwall 2019), but the grounds on which the public evaluates and trusts/distrusts news media cannot be found primarily in industry or academia. Relatedly, it is important to find out what kind (if any) of journalistic performance damaged the trust in the first place, in order to see what needs to be repaired and how (Schoorman, Mayer, and Davis 2007). For instance, Karlsson and Clerwall (2019) found that people were indignant about spelling mistakes and other linguistic flaws in the news. This is clearly a dimension of journalistic ability, and both quite easily fixed and easy to be transparent about, but not what is evaluated in academic research.
Second, we need to investigate more closely, or develop, the transparent and nontransparent variants of those specific journalistic performances of ability, benevolence and integrity. This study has uncovered ambient transparency that needs to be explored further. There might be further and other yet-to-be-discovered forms of transparency that different forms of publics might appreciate or take issue with.
Third and finally, we need to empirically investigate what counts as relevant, legitimate, and transparent performances of ability, benevolence and integrity and the extent to which that affects credibility, and varies across different contexts. It is evident from this study, for instance, that the current acts of journalistic transparency do not appeal uniformly to people. Likewise, since the present study was conducted in a high-trust society, transparency dimensions might have other outcomes in a low-trust society.
Only by doing the above will we, theoretically and empirically, have connected public trust, journalistic performance and transparency, and it may still be the case that factors other than transparency are much more important for the trustworthiness of, and trust in, journalism. It is also important to remember that journalism does not operate in a vacuum and that other actors might seek to influence how journalistic ability, benevolence and integrity is perceived through various means. Furthermore, the performance of journalism is evaluated not only by individuals but individuals situated in a social world. Taking a leap is very much a question of others also joining that leap (Lewis and Weigert 1985).
As previous research has shown, there are small positive effects of transparency on journalistic trustworthiness. There might be even bigger positive effects if we find out what citizens view as trustworthy journalistic performances, demonstrating ability, benevolence and integrity in a transparent way. Transparency might work in the long term, as journalistic norms and performances have changed before, but research needs to do a better job finding that out, particularly as transparency is central to the exploration of normative awareness in journalism studies (Carlson et al. 2018). Should it be the case that the field has overestimated the role of transparency, this must be addressed too, both in academia and industry.

Limitations and Future Research
This study has some limitations and caution needs to be applied. The response rate for the survey was rather low. The data in this study is from a high trust society with high news consumption and the data were collected a few years back. While this is certainly a limitation and fresh data from Sweden and elsewhere might alter the picture, it also offers a point of comparison for future research. The various forms of transparency may appeal to other people and transparency may serve a different role in another contexts. The regression analysis explains less than 20% of the variations and there are other factors to consider. Several factors should be included in future studies: how political inclination affect attitudes towards transparency, trust and attitudes towards social institutions in general, and affluence, amongst other things. There is also need for qualitative approaches, especially with regard to how the use of different social media platforms can produce such diverse effects. If there was a storm warning in your area, how probable is it that you will use authorities to be informed?
4.81 2.00 2091 If there was a storm warning in your area, how probable is it that you will use social media to be informed?
3.62 2.17 2091 If there was a storm warning in your area, how probable is it that you will use interpersonal communication to be informed? 5.24 1.60 2091 If there was riots and civil unrest somewhere in Sweden, how probable is it that you will use news media to be informed?
6.03 1.30 2091 If there was riots and civil unrest somewhere in Sweden, how probable is it that you will use authorities to be informed?
4.78 1.94 2091 If there was riots and civil unrest somewhere in Sweden, how probable is it that you will use social media to be informed?
3.83 2.19 2091 If there was riots and civil unrest somewhere in Sweden, how probable is it that you will use interpersonal communication to be informed? 5.05 1.72 2091 If a head of a governmental body is suspected to have used tax payer's money to pay for private consumption, how probable is it that you will use news media to be informed? 5.75 1.51 2091 If a head of a governmental body is suspected to have used tax payer's money to pay for private consumption, how probable is it that you will use authorities to be informed?
3.82 2.04 2091 If a head of a governmental body is suspected to have used tax payer's money to pay for private consumption, how probable is it that you will use social media to be informed?
3.33 2.06 2091 If a head of a governmental body is suspected to have used tax payer's money to pay for private consumption, how probable is it that you will use social media to be informed?