Dynamics of (dis)trust between the news media and their audience: The case of the April 2019 Israeli exit polls

This paper explores the dynamics of (dis)trust among experts, journalists, and audiences through the case study of an inaccurate exit poll aired on a leading Israeli television channel. It combines empirical data from the Israeli April 2019 elections with a conceptual view of exit polls as both sources of information and national rituals to address public discourse on the polls and its underlying suspicions. A multi-method approach yielded a corpus consisting of focus groups with citizens, in-depth semi-structured interviews with journalists, pollsters and experts, and qualitative textual analysis of news reports. Using inductive-qualitative analysis, we identified three types of public narratives, each casting blame for the erroneous exit poll projection on a different type of actor. The statistical and biased-media narratives tally with declining trust in the news media and assume misbehavior by pollsters and news creators respectively. The deception narrative, on the other hand, suggests that right-wing voters systematically sabotaged the exit poll projections. By extending trust beyond journalistic information, this narrative foregrounds the cultural meaning of election night rituals. Taken together, the narratives found in this study delineate (dis)trust as an interplay of active participants in the creation, reception, and interpretation of news. Our findings thus touch upon key attitudes towards both media and democracy and have implications for further studies on collective rituals and information evaluations in an era of eroding trust.


Theoretical framework (Dis)trust between news creators and audiences
Trust is a fundamental social emotion inherent in human interactions (Quandt, 2012;Uslaner, 2002). Trust is the expectation that specific social relations will be beneficial rather than harmful (Mayer et al., 1995), and is therefore widely considered a facilitator of social cooperation (Giddens, 1990). Yet, both trust and distrust are crucial for democracy; while the former empowers complex modern societies (Quandt, 2012;Uslaner, 2002;Van Dalen et al., 2011), the latter enables citizens to monitor their political institutions (Markova et al., 2008).
Trust plays a crucial role in the way audiences receive, evaluate, and select information (Kohring and Matthes, 2007;Tsfati and Cappella, 2003). While journalism studies have focused on the erosion of public trust in the news media (Hanitzsch et al., 2018;Quandt, 2012;Usher, 2018), (dis)trust can also be expressed in journalists' own attitudes. Several studies have demonstrated journalists' distrust in their information sources, in public institutions, and even in the news media itself (Aharoni and Tenenboim-Weinblatt, 2019;Tejkalová et al., 2017). Studies have also indicated journalists' limited confidence in their own audiences and their comprehension of news stories (Coleman et al., 2009). News (dis)trust involves complex recursive dynamics. Aware of the public's growing criticism, journalists adopt an objective style to gain trust (Tenenboim-Weinblatt and Baden, 2018). However, journalists might also reduce public trust: expressions of political cynicism might backfire, contributing to the audience's distrust not only in politicians (Hanitzsch and Berganza, 2012;Van Dalen et al. 2011) but, potentially, also in journalists themselves. The complexity of the journalist-audience relationship leads to the claim that news trustworthiness is not merely a matter of accuracy; it also comprises mutual expectations, imaginations, and values shared by both the news creators and the audiences (Ampofo et al., 2011). This paper moves this research trajectory forward through an examination of the interplay between journalists, experts, and audiences, all of whom are involved in the process of assigning meaning to the news. It emphasises (dis)trust in the public -a dimension of the journalist-audience relationship that has received little scholarly attention yet is of particular relevance in a 'posttrust' era.

Polls as objects of (dis)trust
Democratic elections stand at the centre of public trust in the political system (Hooghe and Stiers, 2016). In this context, exit polls, which offer the first tentative reports of election results, are an object of trustworthiness evaluations. Scholars have often criticised the timing, question phrasing, and simplistic representation of media polls that make them prone to information biases (Althaus, 1996;Best and Krueger, 2012;Herbst, 1995;Rosenstiel, 2005). Some have condemned the extensive coverage of polls and their transformation into news events in their own right (Lipari, 1999;Toff, 2018Toff, , 2019. Similarly, audiences have generally been found skeptical of opinion polls (Price and Stroud, 2005;Tsfati, 2001), unwilling to participate in them (De Heer and De Leeuw, 2002), and hostile towards their coverage (Chia and Chang, 2017) -especially when the results are inconsistent with their political beliefs (Herbst, 1995). Other studies have demonstrated the allegation that pollsters intentionally manipulate poll results -a criticism that is commonly voiced during election campaigning (Ampofo et al., 2011;Price and Stroud, 2005;Tsfati, 2001).
Audience skepticism of polls may be rooted in a sense of distrust in the news media or the political system, or in a general social distrust (Ampofo et al., 2011;Price and Stroud, 2005), and may be exacerbated by inaccurate projections or poll discrepancies (Ampofo et al., 2011;Toff, 2019). Negative perceptions of pre-election and exit polls have, in addition, been found positively related to higher levels of education and political involvement (Kuru et al., 2017;Lavrakas et al., 1991) as well as to the third-person effect (i.e. thinking that polls affect others but not oneself) (Price and Stroud, 2005).
Public trust in exit polls is usually measured in terms of their perceived accuracy (Lavrakas et al., 1991;Price and Stroud, 2005). In democratic societies, however, polls not only provide content but also act as moments of societal self-reflection wherein a community appraises its own values, expectations, and viewpoints (Herbst, 1995;Kertzer, 1988;Lipari, 1999;Price and Stroud, 2005;Toff, 2018). We therefore seek to unpack the suspicions surrounding the April 2019 Israeli exit polls by integrating questions of accuracy with the implications of these suspicions for the cultural meaning of polls.

Exit polls as mediatised national rituals
Far from being merely an instrumental source of knowledge, election night projections might carry ceremonial meanings. Accordingly, trust in exit polls not only serves an informational function but also permits citizens to partake in the public revelation of election results and to reaffirm democratic norms (Cottle, 2006;Ross and Joslyn, 1988). We therefore analyze the Israeli exit polls as a national media event. Originally coined by Dayan and Katz (1992), media events are rare, festive, and scripted ceremonies covered by live television broadcasting. Television exit polls, which are typically timed to coincide with the closing of polling stations, constitute a contest event (Dayan and Katz, 1992) over a country's premiership and occupy a salient space in election night coverage. These special broadcasts also meet other criteria of media events, as they attract extensive national attention 1 and sustain social solidarity (Cottle, 2006;Dayan and Katz, 1992;Sonnevend, 2018). Traditionally, the Israeli exit polls serve journalists and audiences alike in evaluating the outcome of the election; they are a crucial component in shaping people's expectations of the political reality they will wake up to in the morning and the future collective memory of this national event. In line with the ritual view of communication (Carey, 2008;Dayan and Katz, 1992;Kertzer, 1988), the television exit poll is a communal event in which journalists, pollsters, and the wider public assemble to interpret, imagine, and reflect upon the implications of the results for the future of the country.
As national spectacles of tradition, media events depend on the willing involvement of participants (Cottle, 2006) and thus rely on mutual trust between the public and the media. According to Scannell (2018: 155), witnessing mediatised public events at the very moment they unfold is a way of striving for truth. Nevertheless, mediatised rituals might convey biased information (Scannell, 2018) or provoke challenging non-hegemonic interpretations (Sonnevend, 2016(Sonnevend, , 2018 and are therefore surrounded by suspicions (Cottle, 2006). Katz and Dayan (2018) have recently stressed the interconnectedness of media events and public trust. The growing cynicism of the audience, they argued, is a substantive threat to the reception and entire existence of media events, which will, consequently, lose their veracity and ceremonial qualities and cease to exist in post-modern times (Katz and Dayan, 2018: 151).
In line with this pessimistic view, the case of the April 2019 exit polls raises questions about the integrity of this media ritual. The public debate on Channel 12's inaccurate projection and the suggested reasons for such poll disparities highlight the fragile foundation of media rituals on the presumption of voluntary public trust and participation. Essentially, it alludes to a breach in the relationship between the news media and their audience. Accordingly, this paper considers (dis)trust manifestations in the exit polls that relate to both their cognitive and ritualistic nature. It analyzes the public discourse around this rich case study and its implications for understanding the complex relationship between news creators and audiences in these times of uncertainty.

Method
This paper follows the conflicting stories about Channel 12's exit poll from experts, journalists, and members of the public and the suspicions these stories encapsulate. As media events, the television exit polls allow for the 'possibility of looking at different modes of media creation and reception' (Katz and Dayan, 2018: 147). Such a reception, however, might not be unequivocal as different audiences often have competing interpretations of media events (Sonnevend, 2016). We therefore examine manifestations of public suspicions by deploying, paradoxically, a hermeneutics of trust (Scannell, 2018: 156). Put simply, the aim of this study is not to formulate a decisive account of what actually led to the incorrect 2019 exit poll results; instead, we are interested in the narratives articulated around this event -narratives which are culturally relevant and consequential regardless of their accuracy. In order to achieve this goal, this paper applies a combination of methods from a large-scale study on the 2019-2020 Israeli elections. 2 The research corpus comprised in-depth semi-structured interviews with journalists, pollsters and experts, focus groups with members of the public, and qualitative textual analysis of media texts concerning the exit polls. In order to account for media producers' interpretations of the April exit polls, we analyzed 16 semi-structured in-depth interviews (INT1-INT16) with 11 journalists and five experts (including two pollsters) who were involved in the creation and interpretation of election projections. Six of these work for television channels (three for Channel 11 and three for Channel 12), six work for print/online news outlets (two for Yedioth Ahronoth/Ynet, two for Haaretz, one for Globes, and one for Israel Hayom), and the remaining four are creators of independent current affairs websites or podcasts. In terms of gender, the corpus was male-dominated and included only three women, which is in line with the general bias in the political journalism field.
In an attempt to examine the discursive negotiations of the public on what led to the exit poll discrepancies, we also drew on five focus groups of citizens (FG1-FG5) conducted in three waves (W1-W3). The first two meetings (W1, W2) took place prior to the April elections (2 months before and a few days before respectively), and the third meeting (W3) took place three weeks after election day and the airing of the television exit polls. Four of the groups comprised Jewish voters and were conducted in Hebrew (FG1, FG3, and FG4, participants aged 30-64, FG2, participants aged 19-28) and one group comprised Arab voters and was conducted in Arabic (FG5, participants aged 23-58). All groups were heterogeneous in terms of gender, education, and political orientation. Each commenced with 7-12 participants in W1, decreasing to 5-11 participants by W3 due to attrition. Our analysis focuses on participants' reactions to the April exit polls (W3), although we also consider expressions of polls distrust that were raised before the elections (W1, W2).
The analysis was supplemented by a qualitative textual analysis of media texts reflecting the public discourse on the exit polls. By using various search strings (such as 'exit poll', 'television poll', and the names of Channel 12's pollsters), we purposefully retrieved the main news articles on the television exit polls as well as public interviews with the television pollsters. The texts were sampled between April 9, 2019, the first 2019 round of elections and when the initial exit polls were broadcasted, and September 17, 2019, the second 2019 round of elections and when new exit polls were broadcasted. The April exit polls attracted extensive coverage during this time period. 3 Overall, the sample contained 42 texts (T1-T42) comprising interviews and reports.
The interviews and focus groups were recorded, transcribed, and anonymised. All research materials subsequently underwent a qualitative analysis consisting of several stages. First, after several rounds of in-depth reading, we extracted all references to the April 2019 television exit polls (e.g. distrust, enjoyment), including specific interpretations of Channel 12's inaccurate poll (e.g. political bias, poll deception). Next, we identified and arranged these public reactions by using an inductive approach, starting with an open descriptive coding and gradually building towards conceptual categories, based on a recursive process of personal interpretation, integration, and refinement (Lindlof and Taylor, 2011). The emerging patterns were assessed in relation to different theoretical aspects of polls (dis)trust, including cognitive and cultural dimensions (e.g. instrumental functions, ritualistic consumption). Finally, these identified themes were classified into three public narratives that emerged around the April 2019 Israeli television exit polls. All quotes included in this paper were translated into English by the authors.

Findings: Three narratives of suspicion
The discourse surrounding the exit polls unveils three types of narratives that explain the discrepancies between the Channel 12 exit poll and the final election results. Each of the narratives casts blame for the erroneous exit poll projection on a different type of actor: the statistical narrative discusses the pollsters' part in the inaccurate results; the biased media narrative relates more broadly to the role of news media in providing skewed information; and the deception narrative blames a segment of the public, suggesting that right-wing voters systematically lied to Channel 12's poll. While the first and second narratives are in line with recognised expressions of distrust in experts and journalists (Ampofo et al., 2011;Coleman et al., 2009;Price and Stroud, 2005), the third narrative signifies (dis)trust in the public -a relatively new aspect in the literature on media trust. We first briefly present the first and second narratives. We then go on to address in detail the third narrative which deals with the various perceived motivations attributed to rightwing voters by different actors.

The statistical narrative: Blaming the pollsters
The first narrative emerging after the release of the April exit polls placed the blame for Channel 12's erroneous results on the channel's chief pollsters, Dr. Mina Tzemach and Mr. Mano Geva. According to this viewpoint, Channel 12's pollsters failed to project the size of the party blocks due to an erroneous statistical model. A few actors even speculated that Channel 12's pollsters deliberately biased the television polls. Statistical explanations for Channel 12's inaccuracy were dominant in both our focus groups with citizens and interviews with journalists and pollsters. Examples of the statistical narrative were also prevalent in the media coverage of this incident.
References to public distrust of pollsters were widespread even before the April elections. Our focus group participants in particular singled out Tzemach's previous forecasting failures, mentioning that her polls 'are not exactly accurate' (FG1W2). Naturally, the quantity and intensity of such distrust increased after the broadcasting of Channel 12's inaccurate results. In the group meetings that followed the April elections, many participants mocked 12's poll, laughing, shaking their heads, and exclaiming out loud: 'What a mistake!' Zohar (F, 38) described this poll as 'a disgrace', and claimed that from now on Tzemach 'should be taken with more than a grain of salt' (FG3W3).
Many considered this poll's inaccuracy a miscalculation resulting from an insufficient sample size, an inappropriate method, or biased statistical assumptions. Others referred to the challenge of predicting election results in Israel's multi-party system and identified potential statistical errors that have led various journalists and members of the public to conclude that pre-election and exit polls are 'anachronistic' (INT 9) and have 'lost their credibility' (FG1W3). A number of actors speculated that Channel 12's pollsters 'intentionally misrepresented' the party blocks (FG3W3) in order to 'manipulate the polls' (INT16). Some even suggested that because Tzemach had conducted private polls for the Blue and White party, she was trying to 'favor them on television' (INT11).
The television pollsters were thus largely denounced by our interviewees and criticised in the news media for having 'too many unknown variables. . . with no attempt to provide proper public disclosure' (Nadav Shragai,Israel Hayom,11.04.19,T13).

The biased media narrative: Blaming the news creators
A parallel narrative points to the role of the media themselves in creating the gaps between the competing television exit polls. As communication scholar Gabi Weimann wrote in an opinion piece: 'The failure of the polls was caused not only by pollsters. . . it also involved journalists and political commentators who know the limitations of polls very well but insist on promoting them in every headline and news flash' (Ynet,14.04.19,T18). In this vein, the media were accused of presenting exit poll results imprecisely and overemphasizing inaccurate projections. In many of our interviews and focus groups, participants mentioned that journalists might have unintentionally distorted the polls as they competed for ratings. Some experts and citizens even speculated that news creators might have deliberately skewed the results. This narrative relates to the suspicions of both audiences and creators about the reliability of media products and the political interests that motivate them (Aharoni and Tenenboim-Weinblatt, 2019;Coleman et al., 2009;Hanitzsch et al., 2018). Expressions of the biased media narrative were less prevalent in the news coverage of the exit polls, where it was mostly found in opinion columns by experts.
Citizens, experts, and even journalists voiced their concerns regarding the contribution of the media to this specific case. Consistent with scholarly criticism of media polls (Althaus, 1996;Rosenstiel, 2005), they condemned both journalistic representations of statistical data and the excessive weight of polls in election coverage. For example, they pointed to the tendency of news media to frame mere statistical shifts as a case of altered public opinion (INT10) and to prioritise 'dramatic clear-cut projections' over 'hesitant and complex' ones (INT 13). One veteran pollster criticised journalists for formulating hypothetical and speculative poll questions, for example, asking about the suitability of certain politicians for premiership despite them not being nominated for this position (INT7). According to our interviewees, changes within the profession of journalism have also compromised the accuracy of election forecasting. While, in the past, the rhythm of the legacy media allowed pollsters to provide journalists with meticulous statistical reports, it is no longer possible to maintain such high work standards due to the demand of digital and television outlets for accelerated coverage (INT7, INT8).
Others we talked to were even more blunt in their criticism of the media, describing journalists' direct responsibility for the botched exit poll. Some accused Channel 12 of fabricating its poll to satisfy left-wing parties and voters (FG5W3), while others discussed the economic motivation for airing inaccurate statistical projections, which enables the media to 'talk about the polls and then talk about how wrong these polls were' (INT10). Ze'ev (M, 36), a focus group participant, interpreted the contradictory exit poll results as a media producer's decision to 'create some drama to get people to watch us' (FG1W3). Both journalists and experts warned that the economic incentive of polls endangers journalism, as it 'greatly influences the editors' (INT4) and prevents the media from presenting both the limitations of election night forecasts and their inability to account for the real election results.

The deception narrative: Blaming the public
Alongside the explanations that place the blame for the inaccurate exit poll on either the intentional or unintentional misbehavior of pollsters and journalists, a third explanation emerged focusing on public misbehavior. According to this narrative, the mistaken Channel 12 exit poll was a result of systematic deception by right-wing voters. Expressions of the deception narrative were prominent in all of our research materials, found in both the reactions of interviewees and focus group participants to the exit polls and in their journalistic depictions.
On election night, Zion Nanos, a Channel 12 journalist, reported from the Likud headquarters. He interviewed several activists who claimed to have lied in the channel's exit poll. In his report, broadcasted the evening after the election on Channel 12's newscast (10.04.19, T8), Nanos speculated that this manipulation by Likud activists 'explains our [Channel 12] exit poll errors.' The following day, Channel 12's pollsters Tzemach and Geva participated in a panel on the Morning News show entitled 'The Pollsters' Introspection' (Channel 12,11.04.19,T9). In this first interview of pollsters after the election, they reinforced the claim that deception by Likud activists was the primary cause for their erroneous projections. They reported identifying an inherent inconsistency between their poll results and the observed votes from the 60 voting stations sampled. They asserted that inserting the real results from these stations proved their statistical model to be correct, thus confirming the assumption that systematic lies had indeed taken place. The pollsters also supported Nanos's report with anecdotal evidence of messages and phone calls they received from Likud voters, apologizing for lying in the poll.
The deception narrative was also widely referred to in other media outlets. In a Haaretz article titled 'The Likudnik's Revenge', the incorrect Channel 12 exit poll was explained as an organised 'vote of protest' (Amnon Harari, 10.04, T12). Similarly, an article in TheMarker built on Nanos's report to claim that 'Likud members. . . lied in the exit poll' (Nati Toker, 11.04, T14). This narrative quickly became the subject of public controversy, raising doubts among journalists, experts, and members of the public. While some accepted the suggestion that Likud activists had lied, others deemed it 'ridiculous' and 'bizarre' and strongly contested it. Even those who did accept it could not agree on the motivation for such deception nor explain why only Channel 12 and not its competitors had experienced it. Three main kinds of motivations were attributed to the deception, each with important implications for public and media trust: (1) targeting the pollsters; (2) targeting Channel 12 or building on its popularity; and (3) targeting the political out-group. While experts and journalists mainly developed the first two motivations, the focus groups discussed all three, and especially the third one.
Explaining the deception as targeting the pollsters. First, the deception theory was explained as a vendetta against Channel 12's pollsters and, especially, the acclaimed Mina Tzemach. In the meeting after the April elections, some focus group participants suggested that Tzemach's previous exit poll failure might have inspired people to actively bias the 2019 exit poll. As Zohar put it, 'Everyone remembers that it was the exit poll of Mina Zemach in 2015 that was wrong. Then people might have wanted to. . . reproduce it' (FG3W3). Some experts and journalists also noted the possibility of intentional fraud against Tzemach, as one statistician explained: 'I believe that there are people in Israel. . . with a tradition of lying to Mina Tzemach. . . they give a wrong answer in order to embarrass [her]. . . This is enough to bias a sample' (INT13).
The perception that 'people always lie to Tzemach' (FG1W3) was also evident in the pre-election entertainment discourse. For example, in an episode of the satire show Erez Nehederet (A Wonderful Country) created especially for election day and broadcasted before the exit poll, Tzemach was described as 'the national pollster, the woman that all Israelis lie to' (09.04.19, T1). Although many public actors refuted this 'hidden, anti-Mina Tzemach agenda' (FG3W3), Tzemach herself referred to 'the deception in the last election' when she resigned 3 months later from her 46-year-long career as a television pollster (Maariv online, 04.07.19, T22). Tzemach's last exit poll in April 2019 was regarded as an incident that 'cast a shadow on a glorious career' (Arik Bender, Maariv, 10.04.19, T15).
Explaining the deception as targeting Channel 12. A second, more common explanation suggested that the misleading votes in the exit poll were aimed not at the pollsters but, rather, at embarrassing Channel 12. This explanation was supported by a Likud activist's statements on a television panel hosted by Channel 12: 'We have nothing against the [exit] poll that [you] do. . . but, the media brainwashing that is done before every election, that you take one candidate. . . and magnify him. . . it's so repulsive!' (11.04.19, T9).
In our focus groups, too, the media were blamed for their left-leaning ideology. For example, Lavi (M, 50), a right-wing voter, expressed his distrust of the media from the very beginning, stating that they 'are very mobilised against Bibi [Netanyahu's nickname], [which] makes it impossible to believe what they are saying' (FG4W1). Channel 12, which is Israel's leading television channel, was, in particular, accused of extensive coverage of the incumbent prime minister's corruption investigations and, in the eyes of some of our participants, a more positive coverage of his challenger, Benny Gantz.
Right-wing voters' distrust of the media and, especially, Channel 12 was in line with the prime minister's public attacks on the channel. Throughout the campaign, Netanyahu used his social media accounts to accuse Channel 12 of publishing fake news against him and his family. Netanyahu's assault on Channel 12 for its coverage of the corruption investigations against him culminated during the second election campaign later that year, 4 when he openly called upon his supporters to boycott the channel. 5 This subsequent new peak in politicians' distrust of the media underscored Likud activists' perception of Channel 12 as hostile and may explain their desire to 'punish' it for its believed anti-Netanyahu agenda. As one Likud activist explained in Nanos's report: 'You [Channel 12] abuse us the entire year, we abuse you for one night which is worth everything!' (T8).
Contrary to these explanations, others asserted that Channel 12 is no more left wing or critical of Netanyahu than its competitors, which were evidently not targeted. For example, Prof. Camil Fuchs, chief pollster at the competing Channel 13, rejected the idea of an anti-media campaign in several of his public interviews: 'Are people really lying?. . . Neither Channel 11 nor Channel 13 saw it. So. . . the issue of lies is not a common phenomenon' (Haaretz podcast, 06.08.19, T25). The Channel 12 vendetta narrative was therefore deemed by several journalists and experts as 'only a partial explanation' (INT9), which may indicate why the perceived motivation of Likud activists to sabotage Channel 12 was not entirely accepted in the media coverage. In their news stories covering the April exit polls, journalists were hesitant to adopt this explanation and added their own reservations when they did refer it, using expressions such as 'assuming' and 'probably' and relying on quotes by the Channel 12 pollsters rather than their own descriptions. Some journalists also used this incident to highlight the success of pollsters from the competing channels, especially Channel 13's Fuchs. Given the declining public trust in the Israeli media, they asserted, the pollsters from all three channels could potentially have been the victims of an organised lie. Fuchs, with his 'decades of experience in conducting exit polls', was described as a pollster who, in contrast to Tzemach, was able to 'predict this sabotage effect and "correct" it' (Amnon Harari,Haaretz,10.04.19,T7).
The fact that Channel 12 was the only media outlet that mentioned the deception narrative without refuting it fostered public suspicion of the channel. Some focus group participants emphasised 'holes' in the journalistic coverage of the incident, especially in Nanos's report which, as demonstrated above, was used by pollsters and journalists to establish the organised lies perspective. Right-wing voters, in particular, pointed out several possibilities, such as selecting certain interviews and editing them, which allow news creators to only voice the perspective that fits their agenda. The exit poll deception perspective, they claimed, was just a way of masking the true reason for the breach, namely, a statistical mistake made by the pollsters themselves (i.e. the statistical narrative):

Abraham:
The very fact that from such a large audience they only showed two people means that these were probably the only ones [who admitted lying]. . . There is some combination here between sampling errors and. . .

Gil:
(interrupting) [Those lying in the exit poll] are an insignificant minority!. . . But the media. . . chose to take them. Why? Because it's Channel 12, they won't come and say 'we failed', [they will say] 'we're not guilty! It's them!' (FG3W3) Neither the possibility of a vendetta against Mina Tzemach nor the possibility of a campaign against Channel 12 were regarded as fully convincing accounts in the public discourse. As we demonstrate below, a third explanation suggests that the exit poll deception attempted to build on the popularity of Channel 12 and its exit poll in order to influence a much wider population.
Explaining the deception as targeting the political out-group. The assertion that the real motivation for Likud activists to lie was to profoundly hurt left-wing voters caused the most emotional reactions in the focus group discussions. According to Vicky (F, 32), even if Likud activists had been lying in the exit poll, this would not impact the media as they 'will not stop conducting polls'. She therefore concluded that 'manipulating the polls is manipulating the people' (FG3W3). When elaborating on this explanation, Likud activists were once again accused of building on the previous failures of the television exit polls. This time, however, they were thought to be recreating not only the practice of false projections but also public affective reactions to these forecasts. In Nanos's report, this explanation was raised by Likud voters themselves, who referred to the effect of surprise achieved by lying in the exit poll: 'How else can I let you go to sleep laughing [and] wake up crying?' (T8).
'Going to sleep laughing and waking up crying' seems to be a leading narrative accompanying the relationship between left-wing parties and the Israeli television exit polls. In 1981, the Alignment party relied on the exit polls when it introduced its leader, Shimon Peres, as 'the new prime minister.' However, after the votes had been counted, Menachem Begin, the head of the Likud party, turned out to be leading by a small margin (Shamir and Arian, 1982). In 1996, both pre-election and exit polls predicted an advantage for Labor leader, Peres (Tsfati, 2001). The early celebration by Labor politicians and voters proved hasty, as the actual votes arriving the following morning indicated a small advantage to the Likud candidate, Benjamin Netanyahu. Similarly, on the 2009 election night, Tzipi Livni, the head of the Kadima party, celebrated her victory after getting just one mandate more than Netanyahu in the television exit polls. This turned out to be a premature celebration as the government was once again entrusted to Benjamin Netanyahu (Arian and Shamir, 2011). 6 As suggested by Channel 12's coverage of the April 2019 election night: 'This time, too, the centre-left did not learn its lesson' (T8). Based on the slightly favorable results of Channel 12's exit poll, Gantz immediately announced Blue and White's victory. After receiving the actual results, Gantz's 'victory speech' was greatly criticised, deemed by the media and our focus groups as 'embarrassing' (FG4W3) and 'pathetic' (T12). This political damage was offered by some actors as an alternative deception explanation, according to which Likud activists were perceived as aiming to build on Channel 12's high ratings in order to embarrass left-wing politicians and their voters. Indeed, some actors suggested that gaining favorable results in Channel 12's leading exit poll was the main incentive for Gantz's actions on election night. Channel 12's exit poll was, accordingly, described as an event that 'casts a shadow on [Gantz's] impressive personal political achievements' (T12), and Tzemach, who conducted the poll, was presented as the one who 'caused Gantz to fail' (Haaretz podcast,10.04.19,T11).
In addition to these political implications, an inaccurate exit poll may also have a deep emotional impact on voters. Such feelings were indeed prominent in left-wingers' reactions to the exit polls. As Ofer (M, 28), who voted Blue and White, described: 'I was so happy. . . but [in the morning] I woke up to a black day!' (FG2W3). A narrative of symbolic retaliation against left-wing voters thus emerged. Harel (M, 52), a left-wing supporter, tried to imagine the perspective of a Likud activist: 'Channel 12 is seen as the centre-left channel, the mainstream. This is where it will hurt the most' (FG1W3). The affective reactions to the dramatic shift from winning to losing, described by Vicky as 'false hope', were seen as a main incentive for right-wing voters to lie:

Abraham:
Why the hell would one lie in the exit poll? Hanni: What does it help you with? Vicky: It creates this false hope [among left-wingers]!. . . Because really, this false hope, this empty hope, works.

Gil:
Because it makes you happy!. . . you see the results, [then] you get up in the morning [and see that they were false]!
The attributed desire to make opponents unhappy tallies well with the deeply divided ideological camps and polarised political discourse in Israel (Arian and Shamir, 2011;Shamir et al., 2017). While right-wingers perceived this emotional rollercoaster as merely playful, left-wingers reacted with anger. They completely disapproved of this suggested act which, they claimed, 'tears the nation apart' and symbolises a breakdown of the 'rules of the political game' (FG3W3).
Indeed, in the focus groups where this narrative came up, the discussion quickly became a quarrel. In contrast to right-wingers' dismissal of this behavior, left-wing voters referenced some significant traumatic landmarks in Jewish and Israeli history. They were fearful that such a political event could escalate and warned of a recurrence of previous national disasters. Hanni (F,49) referred to the Holocaust, announcing that 'once a certain group is treated as an enemy. . . it's dangerous!' (FG3W3). Likewise, Gloria (F, 51) recalled the assassination of former Prime Minister Yitzhak Rabin as an indicator of the hatred among Israeli citizens:

Ze'ev:
People often feel that they don't get a real representation [of themselves] or their opinions in the media. . .

Dennis:
Hatred for the media is one thing but. . . hatred for media audiences is something entirely different. . .

Gloria:
What? Are you serious?!. . . A prime minister was murdered here because he was a leftist! What are you talking about?! (FG1W3).
Mina Tzemach also recognised the political out-group motivation, acknowledging that Channel 12 and their pollsters were not the only ones to suffer from such poll deception.
In several public interviews, she expressed her 'pain over the situation of Israeli society' (T9) and deep frustration with the increasing political polarisation.

Conclusion: The ecosystem of news (dis)trust and participation in democratic rituals
This paper follows the discourse surrounding the discrepancy between Channel 12's exit poll projections and the actual April 2019 election results. We identified three types of narratives attesting to this incident: the statistical narrative that blames the pollsters, the biased media narrative that blames media creators, and the deception narrative that blames the public. The latter entails three perceived motives, explaining that the systematic bias of exit polls targeted either the pollsters, the news media, or the political outgroup. The three narratives were prominent in all of our research materials; journalists, experts, and citizens were each named as potential culprits for biasing the exit poll results, while members of each group also expressed their own suspicions of all other actors. The motivations discussed in relation to the deception narrative were most prevalent among our focus groups, with citizens feeling that this case could have been a deliberate attack on both pollsters and journalists, and the general public. These findings illustrate the interplay between active participants in the creation, reception, and interpretation of television exit polls and their mutual feelings of (dis)trust. While blaming experts and journalists connects with familiar notions of eroding trust in the media and public institutions (Ampofo et al., 2011;Coleman et al., 2009), an intentional accusation of the public constitutes a distinct and new theoretical aspect of news distrust. First, it signifies a trust direction reversal, as it expresses news creators' distrust in the public, thus expanding the conventional view of (dis)trust as solely an audience issue (Hanitzsch and Berganza, 2012;Usher, 2018). Second, it goes beyond the common scholarly focus on the reliability of journalistic contents (Usher, 2018) to address essential faith (or, in this case, lack of it) in the ritual signified through their joint consumption. The deception narrative therefore endangers not only the cognitive value of exit polls but also their cultural role as 'the talk of the nation' (Best and Krueger, 2012: 1). Although rooted in a specific Israeli sociopolitical context, our findings nonetheless carry crosscultural relevance as they offer a new way to think about trust in relation to journalism: not only as audiences' content evaluations but as a recursive ecosystem of information, customs, and actors.
That said, it seems that the democratic ritual of the television exit polls has not entirely weakened its hold. A survey that we conducted after this incident showed that a notable, perhaps surprising, number of Israelis (74.4%) intended to follow the television exit polls in the second round of elections set for September 2019. 7 Among those who watched the April poll on Channel 12 (n = 215), 90% said they would 'probably' or 'definitely' follow Channel 12's exit polls again. And Channel 12 did, indeed, successfully maintain its ratings on the September election night. 8 This adherence is in line with the finding that distrust is not necessarily associated with a lack of news consumption or civic participation (Tsfati and Cappella, 2003;Uslaner, 2002). As one of the television pollsters we interviewed concluded, the exit polls are an object of 'entrenched national fascination' and people will continue watching them 'even if the results are imprecise' (INT8).
Although the level of participation did not seem to have changed, our findings portray a substantive qualitative transformation in the meaning of exit polls. Instead of providing the electorate with emotional support and reassurance (Cottle, 2006;Ross and Joslyn, 1988), the television exit polls aroused anger and frustration across the political map. Instead of bringing citizens together (Dayan and Katz, 1992;Kertzer, 1988), they nourished hostility and fueled ideological conflicts. Instead of reaffirming democratic core values and political emblems (Herbst, 1995;Hooghe and Stiers, 2016;Lipari, 1999), they were mobilised to blame pollsters, journalists, and members of the public and to undermine the symbolic foundations of the political system. What is troubling for democratic discourse about the suspicions surrounding the April television exit polls is therefore not simply that the reporting of election night can be manipulated or misused; in an era of deep polarisation and public attacks against the news media, it is the erosion of trust in democratic ritual, if not its deliberate undermining, that should worry us.
Although our findings illuminate the cognitive and cultural aspects of news (dis)trust, they are limited in several ways. While this study documents expressions of distrust in television exit polls, further research is required to quantify the impact of such media events on public trust in both polls and the media. Focusing on the Israeli elections moreover limits the generalizability of the results, and it is likely that distinct cultural settings will elicit additional interpretations. However, we believe that the narratives identified in this study can be applied to debates on inaccurate election projections beyond the Israeli context.
Far from targeting only pollsters and journalists, the alleged exit polls deception was perceived as a manipulation of the public as a whole. The story of the April 2019 Israeli exit polls is therefore a microcosm of the damaged relationship between news creators and their audiences. It shows that while doubting the media is essential for democratic debate (Markova et al., 2008;Usher, 2018), a media-audience dynamic in which finger pointing and blame casting are uppermost might be toxic. Amidst this crisis of trust, which presents important challenges to democracies around the world, conceptualizing (dis)trust as an interplay of different news players and different types of trust paves the way to better understanding the collective evaluations of information in the current political era.
2. The study is part of the PROFECI research project which examines the social dynamics of media projections (ERC Grant 802990). 3. In the week after the election more than 150 articles in 23 major Israeli news outlets mentioned the television exit polls. These articles are a part of a media corpus that we collected following the 2019-2020 Israeli elections. 4. The coalition negotiations that followed the April 2019 Israeli elections failed and did not result in a government formation. Therefore, two additional rounds of elections were, unprecedentedly, called less than a year after the April 2019 elections: on September 17, 2019, and March 2, 2020. 5. For more on Netanyahu's calls to boycott Channel 12, see 'Request rejected: Netanyahu calls on the public to boycott Channel 12', Globes, August 30, 2019, https://www.globes.co.il/ news/article.aspx?did=1001298936 (Hebrew). 6. Responses to the results of previous exit polls are also presented in a documentary called The Actual Results that was broadcasted on Channel 11, https://www.kan.org.il/ program/?catid=1500 (Hebrew). 7. The data is a part of a 10-wave panel survey that we conducted following the 2019-2020 Israeli elections. The survey was administered online in Hebrew and Arabic by iPanel to a representative sample of Israeli internet users of voting age. The data presented in this paper is taken from the fifth wave of the panel survey, which was answered by 771 respondents and took place between June 26 and July 3, 2019 -almost 3 months after the April elections. 8. In September 2019, Channel's 12 exit poll had a top rating of 24.1%, according to an article published in Globes on September 18, 2019, https://www.globes.co.il/news/article. aspx?did=1001301101 (Hebrew).