Abstract
Reliance on evidence is highly desired in disciplines such as science and law. However, the extent to which daily reporters use it to corroborate or refute sources’ say-so is disputed. To explore how evidence is built into stories in ways that are not entirely obvious from the manifest content, we studied the involvement of evidence in a sample of stories, published by leading print and online Israeli news outlets, using reconstruction interviews with the reporters who authored them. Findings indicate that reliance on evidence is an established news reporting routine found in 42 percent of the items. It is used significantly more often under epistemically-challenging circumstances (conflicts over facts, risky publications and unscheduled events) that attract extra reporting efforts (more sources per item, more verifications and longer reporting hours). To systematize reliance on evidence – as other disciplines strive to – news reporting must move further in their evidentiary genealogy, developing a unified system of guidelines on how all types of evidence should be admitted, evaluated and implemented.
Introduction
Reliance on evidence to corroborate or refute claims, hypotheses and beliefs is considered a desired practice in journalism (Carey, 1986; Coddington, 2018; Coddington and Molyneux, 2020; Godler and Reich, 2015; Martine and De Maeyer, 2018; Tuchman, 1972) as well as in law, sciences, philosophy and other disciplines (Douglas, 2012; Stegenga, 2012; Stein, 1996, 2005). However, the extent to which journalists actually use evidence remains unclear. While there is a consensus that evidence is used widely inside the important but tiny niche of investigative journalism (Ekström, 2002; Ettema and Glasser, 1984, 1998; Protess et al., 1987), the extent to which it is used in the broader realm of news reporting is disputed.
Scholars of journalistic knowledge tend to agree that news reporters rarely rely on evidence (Ekström, 2002; Ettema and Glasser, 1984, 1998; Godler and Reich, 2015). On the other hand, those who study news sources and verification insist that they do use evidence (Carlson, 2009; Diekerhof and Bakker, 2012; Tuchman, 1972; Wintterlin, 2020). Interestingly, what is considered evidence tends to differ between both camps.
While the latter use a more inclusive concept that tends to blur the boundaries between evidence, information and news sources, the former employ – as we do – a more exclusive concept of evidence that highlights its special epistemic role.
The purpose of the current study is to explore the use of evidence in daily reporting, to find out how evidence is built into stories in ways that are not entirely obvious from the manifest content. The paper examines the Israeli case, whose generalizability is discussed below, covering all types of evidence involved in the sample of items (N = 434).
There are four major reasons for studying reliance on evidence today.
First, even if news reporters were ‘evidence averse’ when Ettema and Glasser (1984, 1998) made their highly influential observations, today one may expect them to rely broadly on evidence, due to the growing stream of ‘evidentiary objects’ such as documents, photos and videos following the dematerialization of information, its digitization and growing mobility (Martine and De Maeyer, 2018). In this environment, producing and exchanging evidence becomes unprecedentedly swift and easy in the form of email attachments and instant messages or as social media posts.
Second, the necessity of evidence goes hand in hand with the growing overload of information, misinformation and disinformation. The greater the overload becomes, the more one needs justifications (such as evidence and reliable sources) for trusting one piece of information over another (Kelly, 2014; Williamson, 1997). Reliance on evidence is necessary to restore public trust in the media, enabling news institutions to shift gears from ‘trust me’ to ‘show me’ journalism (Kovach and Rosenstiel, 2010) which is trying to earn public trust rather than take it for granted. The growing importance of evidence can also be detected on the agenda of professional epistemologists, where, according to British philosopher Timothy Williamson, ‘questions of knowledge seem to have been marginalized by questions of justification’ (Williamson, 1997: 917).
The third reason to explore reporters’ reliance on evidence is the methodological and theoretical lacunae. Empirical studies of journalistic knowledge in general and evidence in particular are not only scarce (Anderson, 2018; Coddington and Molyneux, 2020; Westlund and Lewis, 2018), but also fragmentary. Studies tend to focus on particular ‘evidentiary objects’ such as eye-witnessing (Allan, 2013; Reich and Godler, 2017; Zelizer, 2007), documents (Anderson, 2010; Ettema and Glasser, 1998; Latour, 1986; Mor and Reich, 2017), hyperlinks (Martine and De Maeyer, 2018) databases (Loosen et al., 2017; Parasie, 2015), photos (Fosdick and Fahmy, 2007; Mäenpää, 2013; Taylor, 2000) and video (Bock and Schneider, 2017; Ristovska, 2018). This fragmentation may indicate that journalism is still in the early stages of its evidentiary evolution, seeking an overarching concept of evidence and ground rules for their systematic implementation.
Finally, reporters are shaping audiences’ standards of evidence and evidentiary literacy by reporting on other evidence users such as researchers, judges and forensic experts, however not always with enough awareness of their logic for admitting and evaluating evidence (Ho, 2015; Kelly, 2014; Stein, 1996, 2005).
The paper opens with a definition of evidence and presents relevant literature from philosophy and law that might inspire journalism studies with their more advanced, elaborate and systematic approach to evidence. The findings section, based on a combination of qualitative and quantitative reconstruction interviews, detects the frequencies and the circumstances under which evidence tends to emerge and suggests a typology of the roles of evidence behind the published news. We conclude with some recommendations for how the unhealthy combination of frequent but inconsistent reliance on evidence in news reporting can be mitigated.
Literature review
What is evidence? Our definition rests upon three attributes, which we import mainly from disciplines other than journalism studies.
Hence, we define evidence as any information (which may or may not be contained in material artefacts such as documents, photos and video) that points towards some other information, strengthening or weakening one’s confidence in that information.
According to the Oxford dictionary evidence means ‘the facts, signs, or objects that make you believe that something is true’, while according to the Cambridge dictionary, it is ‘one or more reasons for believing that something is or is not true’. Philosophers, too, emphasize the special epistemic status of evidence. According to Kelly (2008), evidence is needed ‘in cases in which the truth is not transparent’ (p. 948) ‘to close or narrow the non-trivial gap between the truths in question and what one would be in a position to know in the absence of evidence’ (p. 943). According to Bird (2018: 316), evidence ‘is something we want for a purpose [. . .] in order to make [. . .] epistemically worthwhile inferences’.
Anderson (2018) maintains that journalists are utilizing different forms of evidence ‘to render their knowledge claims more certain, contextual and explanatory’ (p. 1). According to his genealogical framework, new modes of evidence emerge in different historical key moments, following technological, cultural and occupational changes. Hence, changes of ‘evidentiary objects’ mark different ‘journalistic cultures of truth’ (p. 186) and practices of knowledge (Anderson, 2018). For example, data-driven journalism marks a shift from ‘record’ to ‘report’, from individualized narrative to social scientific insights, and from revelation of hidden facts to deeper comprehension of already known trends. On the other hand, the rise of interviews marks a ‘documentary disenchantment’ and a shift from ‘paper to people’ (p. 356).
Poor evidence of the use of evidence
The evidence on reporters’ limited use of evidence turns out to be meager not only in the core literature on journalism and epistemology, but also in broader circles of news sources and verification.
The idea that news reporters do not rely on evidence can be traced back to Ettema and Glasser (1984, 1998). In this seminal work, the authors tried to emphasize the uniqueness of heavy reliance on evidence among investigative reporters, contrasting them with daily reporters. However, Ettema and Glasser did not study daily reporters, rather inferred their aversion of evidence from prominent newsroom ethnographies (e.g. Fishman, 1980; Tuchman, 1978). These ethnographies highlighted the tendency of news beat reporters to accept regular source versions at face value, an idea that resonated in later studies (cf., Diekerhof and Bakker, 2012; Ekström, 2002; Godler and Reich, 2015; Shapiro et al., 2013).
According to Ettema and Glasser (1998), investigative reporters rely heavily on evidence to gain ‘adequate grounds for acceptance of a belief’ (p. 157), using a clear hierarchy of ‘context-specific’ evidence (p. 165). Daily reporters, on the other hand, ‘accept at face value’ their sources’ say-so – ‘not necessarily as true but at least as news’ (p. 159). Furthermore, according to Ettema and Glasser (1998) daily reporters are not even expected to collect evidence because they are bureaucratically justified by the official sources that provided it.
Even studies focused on epistemic aspects of news work overlooked the existence of evidence, or maintained that it was not used in regular reporting (Diekerhof and Bakker, 2012; Ekström et al., 2019). Godler and Reich (2015) suggested that ‘evidence par-excellence such as documents or photos is generally unavailable to the journalist’.
This can be explained by the barely observable nature of the phenomena. After all, the decision whether a piece of information could serve as evidence is made in passing inside the reporter’s mind. According to Carey (1986), journalists lack the know-how what to do with evidence. Even if they reach out for evidence in ‘the library, the computer, government documents, historical surveys’ they have no conventions for ‘sifting and judging’ it (p. 168).
Another reason for selective use of evidence is reporters’ endemic shortage of time and resources, which forces them to select where they invest their ‘epistemic efforts’ (Ekström et al., 2019). Like other responsible epistemic actors, according to ‘pragmatic encroachment’ theory, they tend to raise the threshold of proof under greater risks (Fantl and McGrath, 2009). This may happen whenever there are conflicts over facts (Reich and Barnoy, 2021), risks that may follow publication and unclear states of affairs that characterize unscheduled events (Molotch and Lester, 1974). Furthermore, due to the Pareto logic of news reporting (Barnoy and Reich, 2020), in which the majority of news gathering efforts are invested in a minority of stories, evidence may characterize items that involve greater ‘epistemic efforts’ (e.g. more sources, verifications and reporting time) or higher levels of newsworthiness.
Coddington (2018), one of the few who claim that reporters do rely on evidence, maintained that reporters ‘gather corroborating evidence to bolster the credibility of their own accounts’ (13). Using content analysis of online news in the USA, Coddington and Molyneux (2020) found that news texts rarely exhibit firsthand evidence, rather relying overwhelmingly on the assertions of others, with a growing share of thirdhand evidence mediated through other texts – social media posts, press releases, documents and media.
Studies on news sources and verification differ dramatically on the issue of what counts as evidence. According to the more inclusive approach, evidence includes any sort of information that stems from news sources or is about news sources. For example, according to Tuchman (1972), ‘[S]upporting evidence consists of locating and citing additional “facts” which are commonly accepted as “truth”’ (p. 667). Carlson (2009) argues that news sources ‘serve as an essential form of evidence’ (p. 527) reaffirming the content of their testimony. Diekerhof and Bakker (2012: 3) explained that in news reporting ‘sources serve as a form of evidence’ (see also Ericson, 1998). Similarly, according to Wintterlin (2020) reporters’ past experience with sources can serves as evidence that obviates verification (see also Godler and Reich, 2015). Scholars who maintain that the use of evidence in journalism is more scarce, indicate special circumstances under which they tend to emerge, such as eye-witness testimony (Romano 1986), databases (Howard 2014), visual and material evidence (Mortensen, 2015; Pantti and Sirén 2015) searches for causal explanations (Carey, 1986) and ‘forensic’ stories based on surveillance (Gates, 2020).
Lessons from philosophy, science and law
There are two main lessons for journalism in the philosophical, scientific and legal discourse on evidence, in addition to those mentioned above. First, the need to hedge the epistemic authority of evidence. Contrary to a naïve conception of evidence as simple and unequivocal proof (Dorst, 2020; Kelly, 2008), according to Williamson (1997: 723) one should proportion the belief in p [particular proposition] ‘like one proportions his or her voice to a size of the room’. Dorst (2020: 587) suggests that ‘evidence must be a modest guide – a guide that is not sure that it is a guide’.
The second lesson to journalism is the importance of systematizing the implementation of evidence. According to Stegenga (2012), one should avoid ‘ad hocery’ use of evidence (p. 155), rather evaluate it according to its ‘quality, relevance, salience and concordance [that are] among the most important features of evidence’ (p. 221). Douglas (2012) expects practitioners to examine the ‘internal consistency, empirical competency and predictive potential’ of evidence (p. 152), evaluating it according to communicability, completeness, rigor, transparency, scope and practicality. According to legal epistemologist, Stein (2005: 107), evidence law ‘needs to be tightened’, to systematize its implementation and limit courts’ leeway to rely on common sense and contestable values, making decisions in imperfect situations based on partial information.
Based on the literature review, we addressed three research questions:
Method
To address the research aims, we chose reconstruction interviews (Reich and Barnoy, 2020), in which a sample of reporters were asked to recreate the news-making process behind a sample of their recently published items, including reliance on evidence. Unlike content analyses, reconstruction interviews are not limited to explicit reliance on evidence; unlike observations, they enable probing whether evidence was involved in each and every item.
To operationalize our definition and address the bottom-up logic of practice theory (Witschge and Harbers, 2018), we avoided the term ‘evidence’, asking instead: In this item did you use means such as eye-witnessing, documents, photos, recordings or other means, that either strengthened or weakened your confidence in the information that you reported? We emphasized the category ‘other’ to maximize susceptibility to any kind of evidence that played a role in the studied sample.
Whenever a particular type of evidence was mentioned by the reporter, we used a series of follow-up questions about that evidence. However, to limit interview time, in items that involved more than one piece of evidence of the same type (e.g. two documents) – we gathered detailed data only about the most substantial one, according to the reporter’s judgement. The missing data could not distort our findings substantially since only 15 percent of the items (67 out of 434) contained more than one piece of evidence of the same sort.
Reporters’ replies were recorded in a structured matrix, covering the relevant three A’s offered by Lewis and Westlund (2015): actors (sources), actants (technologies) and activities (news gathering practices) involved in each item.
Interviews
Interviews were conducted face-to-face (except for two reporters who were unable to meet) one news item at a time, with the reporter and the interviewer sitting on opposite sides of a table: the reporter with a tablet containing his/her recent randomly chosen six to seven publications (covering 1153 news sources in 434 items) and the researcher with a tablet on which the questionnaires were completed, according to the reporter’s replies.
While the quantitative interviews were used to test the frequencies of reporters’ reliance on evidence and their associations with other aspects of news processes, the qualitative interviews recorded reporters’ reflections on the logic behind evaluations of evidence and their evidentiary and non-evidentiary roles in the published stories.
The second, qualitative wave included a series of 25 follow-up interviews with a sub-sample of reporters who were chosen purposively, in accordance with the standards of qualitative research, addressing three criteria: the interviewee showed adequate verbal and reflective skills in the first interview; published at least one item with a factual disagreement and was willing to participate in a second interview. These interviewees reconstructed 50 news items (2 items per reporter). The interview format was almost identical, mainly conducted face-to-face, in most cases during the month following the original interview (four interviews were conducted within 3 months after the original interview due to reporters’ unavailability).
Sampling procedure
To guarantee a representative sample of reporters and items in the quantitative reconstructions, we took the following steps:
(1) A month-long byline analysis of the leading news outlets 1 to create an updated list of reporters and news beats.
(2) To include all hard news beats (excluding sport, culture and life-style) we analyzed the relative proportion of the three beat clusters – political affairs (government, politics, national security, etc.), domestic affairs (crime, health, municipal, etc.) and financial affairs (macro and micro economics, central bank, etc.), in the reporting workforce.
(3) Sampling of ten reporters from each news organization according to the proportion of the three clusters.
TV and radio news were not included. First, since medium differences are in line with timeliness (daily vs instant reporting, both of which are represented here by print and online news) (Reich, 2015). Second, since broadcast items require longer interview time due to a slower process of presenting and scrutinizing audio and video.
Interviews in both stages lasted between 40 and 146 minutes and were conducted between October 2016 and August 2017. The response rate was 86.4 percent (70/81) in the quantitative stage and 76 percent (25/33) in the qualitative one.
Data analysis
The quantitative analysis started with frequencies of different types of evidence, testing the correlation between them and other key variables that appeared in the research questions. The qualitative analysis followed the seven-stage thematic analysis model of Kacen and Krumer-Nevo (2010): (1) A holistic reading of each qualitative interview transcript and of the answers to the follow-up questions. (2) Organization of data and reduction to initial categories, such as the roles of evidence. (3) Breakdown into specific themes, marking statements that could explain or contextualize the statistical findings. (4) Rebuilding the themes and adjusting the phrasing based on a comparison of the research’s original theoretical goals and its existing findings. (5) An additional holistic reading of the entire data in its original raw form. (6) Reexamination of themes to verify and adjust them. (7) The writing stage, which attempted to merge the qualitative and quantitative data.
Measurements
Almost all the questions in the quantitative stage were closed, with a spectrum of possible answers, and emerged in pilot interviews. Unexpected replies were registered as ‘other’. The main variables were operationalized as follows:
Results
Results show how news reporters use different types of evidence, whether they were explicitly mentioned in the item (Coddington and Molyneux, 2020) or not. We present the results according to the research questions.
Prevalence of evidence
Addressing RQ1 on the frequency of evidence, findings indicate that, contrary to accepted wisdom, reliance on evidence is a common practice in news reporting. Out of 434 items, 181 (42%) involved at least one type of evidence.
These findings indicate that reliance on evidence is not limited to the highly important, though peripheral, domain of investigative journalism, rather it is part and parcel of its core activity.
The fact that there were additional documents or eyewitness accounts that the journalists identified as informative rather than evidentiary indicates that journalists do distinguish between evidence and other types of information.
Which type of evidence is more prevalent?
Findings in Table 1 indicate that despite the capacity to transfer the most sophisticated digital data, traditional evidence still control? the mix of evidence, led by documents followed by eye-witness accounts. Video and still photos come next, with audio playing a negligible role. Databases played a marginal role as well – contrary to the broad research attention accorded it in recent years (Anderson, 2018; Loosen et al., 2017; Parasie, 2015). Only in two cases did journalists rely on databases: one was a public financial database, and the other – a reporters’ self-generated database.
Breakdown into evidence type.
At least 19 of the cases were based on reporters’ own eye-witnessing.
Refers to the number of documents, photos, videos, etc. in the entire sample.
Contextual factors
To explore whether evidence appears in epistemically-challenging items that draw extended reporting efforts addressed in RQ2, Table 2 juxtaposes items with and without evidence. Findings indicate that evidence tends to emerge significantly more often in epistemically-challenging circumstances accompanied by higher reporting efforts, and in items that are evaluated as more newsworthy.
Evidence and contextual factors.
N = 434. Significance measured with two-side chi-square test/ANOVA means comparison.
As evaluated by the reporter.
The epistemically-challenging circumstances are marked by higher perceived risks, more conflicts between sources over facts and less clear states of affairs that characterize unscheduled events (Molotch and Lester, 1974).
The extra reporting efforts are seen in the significantly higher number of news sources per item, more verifications and longer reporting hours. While these correlations cannot indicate that the extra work is invested in obtaining and processing specific evidence, they do indicate the Pareto principle, according to which a minority of items draw significantly more reporting energies and evidence. These items are also evaluated by the journalists as significantly more important and interesting.
Further analysis indicates also that items with evidence involve significantly higher levels of reporters’ initiative: 48 percent of the contacts in items with evidence, compared to only 27 percent in items without evidence. The difference between groups is significant (
The roles of evidence
To explore the role of evidence, addressed in RQ3, we turn to the qualitative interviews, where reporters were more nuanced regarding the logic behind evidence. Surprisingly, despite its epistemic nature, evidence often plays non-epistemic roles, in addition to their evidentiary ones. Analyzing their accounts, we suggest the following typology regarding the roles of evidence:
Conflict adjudication
The role of evidence culminates in cases of conflicts between news sources over the basic facts. In many of these stories, journalists tried to use evidence to adjudicate which party was closer to truth. However, as one of our former studies shows, adjudications appear in almost half (46%) of the conflictual stories, and they are often not shared with the public (Reich and Barnoy, 2021). For example, a customer claimed that the bank representative misled her while the bank denied it. She called the bank again, this time recording the conversation, sending the recording to a reporter. A shooter who killed a person claimed self-defense, as he was being attacked by the victim with a metal bar. Yet, both the official report and some video footage showed that no less than eight bullets were shot while the victim was still far enough away, indicating an alleged intent to kill.
Hypothesis testing
Another highly epistemic role is the use of evidence to test hypotheses, in accordance with philosophers’ view on the use of evidence in science (Achinstein, 1978). Within journalism studies, the extent to which journalists use hypotheses is highly disputed. While some scholars insist that they do (Hess, 1996; Stocking and LaMarca, 1990), others maintain that journalists avoid using any hypotheses whatsoever (Ekström, 2002; Siegelman, 1973); that their hypotheses are no more than prejudice (Ekström, 2002; Siegelman, 1973) or short term forecasts on the course of unfolding events (Golding and Elliott, 1979; Tuchman, 1978). Furthermore, not too rarely, journalists changed their minds based on new evidence, however mainly regarding incidental issues. For example, a reporter doubted whether an MK (parliamentary member) was indeed serious regarding submission of a particular issue as an official motion for the house agenda. Finding that the proposal had not been officially submitted, the reporter concluded that the MK was just fishing for headlines. In another case, the reporter suspected a corporation’s denials according to which it was fined by environment authorities after violating regulations. He questioned the environmental authorities who supplied the incriminating report.
Strategic backup
Rather than trying to validate their hypotheses, in these cases reporters looked for evidence to make sure they were ‘covered’ in case that they got it wrong. Similar to Ekström’s (2002) observation, journalists often made up their minds which party was closer to truth, without any evidence, while continuing the search for evidence as a ‘strategic ritual’, or a ‘safety measure’ to protect their choices after the fact. For example, to back up accusations against the Ministry of Defense for postponing the installation of CCD cameras in a Jewish settlement which had meanwhile been attacked by terrorists, the reporter asked his sources to supply the unsigned contract with the CCD systems company.
Concrete examples
Evidence often supplies specific examples that sometimes make the difference between abstract accusations that end up in reporters’ trash bins and publishable news stories. The crucial role of examples concurs with journalists’ tendency to be ‘suspicious of abstractions’ and ‘sensitive to the concrete’ (Darnton, 1975: 184). One example of this type of story was evidence that enabled the association of democratic schools to move from a non-substantiated slogan (‘the Ministry of Education is against democratic schools’) to a concrete newsworthy story (‘the Ministry of Education filed four closing orders against democratic schools’).
Storytelling aids
Journalists often harnessed evidence as a storytelling prop. For example, a convicts’ testimonial regarding his abuse by prison guards was incorporated into a storyline, despite its limited probative value, since it added narrative that was in tandem with the broader story on abuse against prisoners. It is important to note that these objects still played an evidentiary role, however mainly for the reporters themselves. On the other hand, for the public, it was a story component to make the story more alive and interesting.
The whole story
Finally, not too rarely, evidence such as comprehensive reports, research studies or official proceedings transcend into the entire story, playing the roles of most of the above. One example was a government decision proposal that ‘summarized the entire story’, according to the reporter, covering its reasoning, initiators, purposes and manner of implementation. Another example was a research report on the ways in which the Israeli government breaches the jurisdiction of its parliament. It is important to note first that these stories wouldn’t have been published without the evidentiary element that emerged wrapped up with many other elements. Second, that even when further information was collected, it only played a complementary role.
This typology indicates that, as the legal epistemologist Stein (2005) suggested, evidence does not emerge as a pure and labelled epistemic entity. Rather it is part of a multipurpose, versatile, dynamic flow of information. It is up to the journalist to designate whether a particular piece of information will be used as evidence and its particular evidentiary role.
Particular type of evidence
While the suggested typology represents all kinds of evidence, it is necessary to explore further whether particular types of evidence, for example, documents or video, play particular roles in the process of news-making, due to their different materiality, timeliness, availability, believability and potential to involve particular sources and contexts more than others.
Obviously, characterizing particular types of evidence requires more caution, since data here involves fewer cases and less richness compared to evidence in general. And yet, some unique features did emerge regarding the following type of evidence.
Documents were not only the most prevalent type of evidence, but also the most authoritative one in our interviewees’ eyes. Although constituting a traditional mode of evidence, and despite the growing capacity to doctor digital documents, they are still considered the ‘gold standard’ of proof in news reporting. One of their main assets is the high versatility, covering the entire typology suggested earlier, including conflict adjudications, supporting hypotheses, providing concrete details, serving as backup for publication, sustaining narrative and supplying ‘the whole story’ as well as for triggering stories and certifying information. For example, minutes of an urban planning committee that supported the claims of environmentalists or legal documents that exposed the intentions of company owners to terminate their partnership.
Eye-witnessing. Somewhat paradoxically, eye-witness accounts were quite widely used despite their limitations as (mostly) third-party testimonials. For example, one reporter who was not allowed to inspect an environmentally hazardous plant interviewed an expert who was allowed in. Contrary to documents, which emerge with a readymade certifying power as ‘immutable mobiles’ (Latour, 1986), the sincerity and believability of eye-witness accounts is, at best, questionable. They are more demanding due to their ephemeral and evasive nature, which requires recording and saving, and since their exchange requires sources’ and reporters’ auditory or physical co-presence. And yet, eye-witness accounts have some evidentiary power that is augmented by the corroboration of other accounts and by their capacity to enhance storytelling. They allow greater journalistic authorship, the use of interview techniques and the selection of suitable quotes. In the qualitative stage of the interviews, only once did a reporter mention his own role as an eyewitness. While having higher confidence in his own observation, the reporter expressed reservations about this type of evidence, being less transferrable to his editors and readers compared to third-party testimonials.
Video. Despite its richest set of modalities, video played a secondary role as evidence, since its use is curbed by timing. Practically, video emerged in slightly later stages of unfolding events, after the publication of initial online stories. Its content often served to change the course of the story, and resolved factual conflicts. One dramatic example is CCD footage that documented a violent confrontation between a mayor and a contractor who was eventually shot by the mayor.
Audio. Unlike video, which played an independent evidentiary role, showing that something indeed occurred or was said, audio appeared only in rare cases, serving as supplementary evidence. In these few cases of auditory evidence, it was used to support a source’s earlier testimony. One reporter, for example, used an audio recording of a telephone conversation between a bank teller and a client that proved the clients’ claims about discriminatory loan conditions.
Still photos. These were used as evidence either for figuring out spatial situations (what the murder scene looked like), or for adjudicating between approximations (how many participants attended a demonstration). While not supplying a final number, for example, in the latter example, photos served as indicators which party was closer to reality. Like auditory evidence, photos served sometimes to support or refute an earlier testimony and in other cases as independent testimony, for example, photos of scars on the body of a convict allegedly beaten by other convicts. Though not published, they helped convince the reporter of the convict’s case.
Discussion
Our findings show that reliance on evidence is an established routine in day-to-day news reporting. This means that even if reporters avoided evidence in previous decades, as prominent scholars indicated (Ettema and Glasser, 1984, 1985, 1998), they rely on them widely today in evidence-saturated news environments, where obtaining, producing, reproducing and exchanging evidence is unprecedentedly instant and easy. There are two indications why current news environments are indeed saturated with evidence. The first indication is that news reporters are too deadline-driven to go out and chase evidence proactively for almost half of their items. The second is that, as reporters showed us during the reconstruction interviews, their constant feeds of emails and instant messages are loaded with attachments of documents, photos, videos, etc.
Even in these feeds, however, evidence does not surface as a purely distilled epistemic element labelled as such, neither in the legal field (Stein 2005), nor in journalism. Rather, as indicated in its multipurpose functions shown above, evidence is part of a versatile, malleable, dynamic and chaotic flow of information. It is the journalists who select the relevant parts in this flow. If it has convinced them to either increase or decrease their belief regarding a putative state of affairs in particular items, they may admit them as evidence. In that sense, the methods used here utilize the optimal position of the reporters as the ultimate judges whether any snippet of information served as evidence in their recently published items.
Despite the capacity to exchange sophisticated audiovisual input, findings indicate that journalists adhere to traditional documents and eye-witnessing – two components in the ‘holy trinity of news objects’ (Anderson, 2018), the third being reporters’ own observations, which scored very low in this study. These preferences indicate that regardless of the historic shift ‘from paper to people’ (Montgomery 2007, p. 190) documents are still valued by reporters, probably for their versatile, traceable, informative and certifying power, while eyewitness accounts are valued for their experiential (Montgomery, 2007), malleable and mutually corroborative nature that adds detail, color and authenticity.
Findings indicate that reliance on evidence is part of a broader ‘economy of effort’ that rations the allocation of reporting energies. According to this logic, evidence tends to emerge in epistemically-challenging circumstances (factual conflicts, unscheduled events, risky publications) and around the top echelon of stories that attract the most journalistic attention (with more sources per items, more verifications and longer reporting hours).
Obviously, we need more systematic research efforts to highlight the broader logic of evidence in news reporting. And yet, the fact that reliance on evidence correlates significantly but weakly with the basic factors of news reporting (see Table 2), may serve as a serious indication that the use of evidence in news reporting is quite erratic and ‘ad hoc-ish’.
If further studies detect similar combinations of frequent but inconsistent reliance on evidence in other journalistic cultures, news reporting will require a comprehensive evidentiary reform to systematize its reliance on evidence, just as other disciplines such as science, law and philosophy strive to minimize such inconsistencies (Douglas, 2012; Stegenga, 2012; Stein, 2005). To help address Carey’s (1986) observation, according to which journalists lack the skills and criteria for ‘sifting and judging’ evidence (p. 168), this reform should include a clear set of applicable guidelines that systematize the admission, evaluation and implementation of all sorts of evidence in news reporting.
Without an integrative framework, journalism cannot calculate the ‘weight-of-evidence’ whenever science is contested (Dunwoody, 2005) or serve as mega-brokers of information strategically positioned to restore public trust in facts and evidence.
One of the most promising inspirations for a more integrative and systematic concept of evidence is evidence law, which, despite its obvious discrepancies, epitomizes a more advanced stage in the evidentiary genealogy. According to Stein, (2005: 35), legal adjudicators represent many of the epistemic realities of professional fact finders. They must reconstruct contested facts ‘on the basis of deficient evidence’, relying on accounts of ‘fallible and biased witnesses’; they are unable to consider ‘all evidence that they possibly could consider’; hence, they are ‘invoking generalizations and inferences that are nothing but rough approximations; and, finally, by subjecting the existing evidence to credibility tests that are never carried through to perfection’ (see also, Ho, 2015).
Obviously, news reporting is too immediate and informal to be governed by a full-blown cumbersome evidence law. Yet journalists can borrow from evidence law useful elements such as the basic vocabulary, including key terms such as admissibility, relevance, adequacy, sufficiency and probative values. Without such a vocabulary, no practitioner can think about evidence, let alone implement it systematically enough. Another systematizing element is the unified framework, which, like evidence law, strives to addresses all types of evidence and the ways they should be treated. If journalism scholars wish to help the journalistic community in this momentous challenge, they will have to shift gears from the fragmented discourse that addresses particular ‘evidentiary objects’ and their genealogies towards an all-inclusive concept, integrating different types of evidence that emerge in current news environments.
This paper has several shortcomings. The current study did not address the extent to which evidence served to corroborate ‘pivotal facts’ and ‘story linchpins’ (Ettema and Glasser, 1998: 167) or trivial points. On a more methodological level, post-hoc self-reports such as the ones used here may over-represent the actual reliance on evidence. In addition, our small qualitative sample limited our capacity to generalize about rarer types of evidence. And yet, alternative methods are far from being ideal for studying reliance on evidence. Observations and ordinary interviews can easily overlook evidence without purposive probing. On the other hand, content analysis is too dependent on explicit mentions of evidence, exposing only the tip of the evidentiary iceberg (cf. Coddington and Molyneux, 2020).
To what extent is the Israeli case generalizable to other news contexts? Obviously, the generalizability of any national case study is limited by a long series of local attributes that shape not only the reliance on evidence but also its availability and accessibility. These may include local standards of truth telling, fact checking, accountability, transparency and e-governance; levels of trust in the press and among political and economic institutions; journalistic adversarial and investigative role perceptions. While some of these factors, such as the more interventionist tendencies of Israeli journalists and the lower trust of Israelis in public institutions (Hanitzsch et al., 2011; Hermann et al., 2019) might lead to greater reliance on evidence, other factors such as the lower levels of transparency (OECD, 2020) can hinder their use. Hence, considering that Israeli journalists share basic norms and practices with other Western counterparts (Meyers and Davidson, 2014), there is no reason to expect dramatically different results in other Western news cultures.
Future studies should explore other news cultures and news genres that rely heavily on evidence (investigative journalism, data journalism, explanatory journalism), less epistemically savvy types of journalism (e.g. citizen journalists, user-generated content, bloggers) and more instant time frames (breaking news, push notifications and live blogging). Further studies may focus on sporadic types of evidence such as databases and audio, which were too rare in this study to allow the drawing of robust conclusions.
Footnotes
Acknowledgements
The authors thank professor Mark Coddington from Washington and Lee University, Dr. Yigal Godler and the anonymous reviewers for their helpful comments to earlier versions of the paper. Many thanks as well to Gilad Halpern from Haifa University for his assistance in the data collection.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
