Identifying Investigative Pieces

ABSTRACT This article aims to contribute to a better understanding of investigative journalism by proposing a multi-step method for identifying investigative pieces, despite the genre’s blurred definition. This method uses internal criteria observable in the texts themselves, regardless of their author, awards received, label, or discourses about them. Based on a literature review, an exploratory analysis and practical testing, we claim that examining articles’ underlying sources provides a set of criteria that allows researchers to select and exclude pieces in a flexible yet consistent way. When applied to a nine-month sample, the method allowed us to create a corpus showing great thematic, methodological and narrative diversity. Moving away from an elitist conception of investigative journalism, our corpus analysis shows that investigative journalism transcends the traditional boundaries of discursive genre. Furthermore, since the corpus includes both small and large investigations, it provides researchers with the first opportunity to observe empirically the varying degrees of investigative journalism.

In today's changing and challenging media ecosystem faced with audience erosion, economic pressures, declining public trust and professional identity crisis (Donsbach 2009), investigative journalism is seen as both an endangered (Miller and Dinan 2009, 252) and a promising form of journalism (Wahl-Jorgensen 2016).In recent years, investigative journalism has received increased attention from scholars as a subfield of journalism studies (Carson 2020), particularly because it is also undergoing important transformations.From individual authorship toward cooperation between journalists, medias and nations (Carson and Farhall 2018), from social sciences epistemologies toward data sciences epistemologies (Parasie 2013), and from legacy media to NGOs (Baggi 2011; Lewis 2011), the face of contemporary investigative journalism is continuously evolving.
However, despite its public visibility and highly valued reputation, investigative journalism remains a blurred concept (De Burgh 2008;Descamps 2017;Forbes 2005).According to Van Eijk (2005, 13), there is ongoing debate as to what can be considered an investigation: "The idea behind these discussions is that the epithet "investigative" may be considered a label of honor: It is a "higher form of journalism", which requires additional criteria above the ones for journalism in general".Indeed, a recurring theme in the literature is that investigative journalism is something "more" than regular journalism (Olsen 2018); the debate typically centres on what constitutes these "additional criteria".Researchers have suggested measuring this "more" in various ways: in terms of requirements and technique (Grevisse 2008, 114), amount of work required (Hunter 2011, 9-10), reflection (Houston 2009, p. vi) and "evidential standards" (MacFadyen 2008, 138).In journalism studies, investigative journalism is mainly defined by its input, that is by the process of the research, instead of by its output or the role of investigative journalists (Eijk 2005, 21-22;Gerli, Mazzoni, and Mincigrucci 2018, 25;Labarthe 2020, 38-39).As Kaplan (2008, 70) observes, "investigative journalists are very devoted to the process of investigative journalism (…) more than the actual results".This supports the idea that investigative journalism should be seen as a specific process that involves "more" than regular journalism.
Another recurring theme in the literature is that professionals often define investigative journalism not only by what it isa "prototype of quality reporting" (Olsen 2018, 240) but by what it is not: reporting that does not involve any research on the part of the journalist, which is often seen as the antithesis of "good" journalism (Conan et al. 2004;Lewis, Williams, and Franklin 2008, 17)."Investigative journalism should therefore be established as an antithesis to what is regarded as announcement and appointment journalism." 1  (Ingmar 2006, 10, our translation).In general, there is a consensus that investigative journalism is the opposite of purely passive journalism and something more than regular journalism.Mcquail (2013, 104), for instance, ranks journalists' main tasks on "a continuum of initiative and activity", ranging from "receiving and transmitting notices of events (a transmission function only)" to "actively investigating, exposing and advocating".
The question that remains, then, is: where does this "more" begin?Where do we draw the line between regular journalism and investigative journalism?Most existing definitions of investigative journalism are based on strong normative assumptions, particularly regarding "good" and "bad" practices and "high"and "low"quality journalism.Indeed, most definitions of investigative journalism are imbued with a certain amount of mythology and idealization (Bromley 2007;Labarthe 2020, 34-25;Ludwig 2011, 14).
Thus, investigative journalism is surrounded by symbolic stakes, as it constitutes a "model of professional excellence" for many practitioners (Marchetti 2002;Stringer 2018), and works as "a symbolic resource enabling them to reaffirm their professionalism" (Descamps 2017, 221;Wagemans, Witschge, and Deuze 2016, 171-173).In this sense, investigative journalism constitutes a normative set of practice with a strong potential for legitimization and demarcation, both within the profession and vis-à-vis competitors outside the profession (Bjerknes 2022;Stringer 2018).

A Gradual Conception
However, normative definitions of investigative journalism are often at odd with the actual practices (Zelizer 1993;Neveu 2013, 111).Indeed, Cancela, Gerber, and Dubied (2021) have shown that, when elaborating on concrete stories, investigative journalists do not share a discrete definition of investigative journalism, but rather rely on a gradual, multilevel definition based mainly on personal commitment.As the authors state (887): […] the definition unfolds on a continuum stretching from basic reporting all the way to fullfledged investigative endeavor.The interviewees used words like "levels", "degrees", "gradation" to talk about what counts as investigative reporting.They also mentioned various investigative pieces as "models" or "best examples", some of which were "little stories" and others "big ones".Descamps (2017, 298) observes that professionals often rank their own investigations on a scale from "small" to "large".Similarly, Ingmar (2006, 23-26) has noted that investigative journalists in Germany locate investigative journalism within a larger concept called Recherche-Journalismus, which includes "verificatory research" (the lowest tier) "thematic research" (the middle tier) and "revelatory research" (the highest tier, the equivalent of investigative journalism).In France, professionals distinguish between journalisme d'investigation, referring to the American tradition of investigative journalism, and journalisme d'enquête (Bastié, Tremolet, and Devecchio 2014), referring to their own practice.Many journalists in these countries prefer not to call themselves "investigative journalists", instead referring to themselves as journalistes d'enquête or Recherche-Journalisten.Indeed, "the term 'investigation', because of its connotations, involves not only an observation but an assessment" 2 (Descamps 2017, 251, our translation).Journalists' discourses thus almost systematically refer to investigative commitment as ranging from a vaguely defined minimum to an idealized maximum.
Given this lack of consensus on the definition of investigative journalism, how is anyone to identify investigative pieces?As Hamilton (2016, 35) notes, "the lack of a consistently applied definition of investigative reporting makes it hard to locate this work".Textbooks provide little guidance in this regard, either suggesting that journalistic investigation does not easily fit the concept of "genre" or dancing around describing it.Agnes (2015, states that "its complex writing borrows from the other genres" 3 and that "investigation takes many forms, and it is hard to account for its diversity".4  Martin-Lagardette (2005, 106) argues that an investigation can be composed of several genres, while Charaudeau (2011, 187, endnote 25) that investigation "is not a genre, strictly speaking".5 Of course, the very notion of genre is also a blurry one: Grevisse (2008, 98) states that journalistic genres do not always refer to a specific kind of content but can "sometimes refer to approaches (investigation), sometimes to types of enunciation (editorial), sometimes to specific forms (direct style interview, portrait interview, etc … .)".
Based on this literature review, investigative journalism would seem to be a particular type of journalism for which identifying the product requires identifying its process (or approach).As we have seen, its process differs from that of regular journalism mainly in the level of commitment demanded of the journalist.And this commitment, in turn, is composed of several criteria which can each be present in varying degrees (Cancela, Gerber, and Dubied 2021).Identifying investigative pieces thus involves balancing two contradictory needs: on the one hand, the need to set a threshold for what constitutes investigative work; on the other, the need to account for varying degrees of investigative commitment, since any selection method must be able to integrate the "many ways to do investigative journalism" (Matheson 2009, 91).

Looking for Internal Criteria
To date, only a few studies have conducted analyses of investigative pieces.The existing ones focus on ideal examples: pieces that have been recognized by professionals as "exemplary" and of a particularly high standard.As Ettema and Glasser (1988, 12) have pointed out, the pieces they analyzed are "the best stories of the best journalists, at least as identified by the criteria, albeit imperfect, of journalism awards and peer nominations".Second, most of these studies base their corpus selection on criteria external to the texts, such as the institutions that produce them (Olsen 2018;Valeeva 2017), the media outlet that published them (Valeeva 2017), the awards they won (Ettema and Glasser 1988;Hamilton 2016;Wahl-Jorgensen 2013a), or their citation as examples by professionals in the field (Eijk 2005;Valeeva 2017).Others do not explain their selection criteria, focusing on "well-known" examples of investigative journalism without questioning why they are categorized as such (Protess 1991).
The problem with these different methods, as Florian Stalph (2018) observed with regard to data journalism, is that they focus on best practices and pieces considered by professionals as exemplary.The risk is to form an incomplete picture of investigative journalism that contributes to a mythical, idealized vision of the genre rather than an informed one.As Hamilton (2016, 44) notes, "by definition, journalism award winners are outliers.It can be misleading to make claims about the general state of journalism by looking only at those narratives that win awards."In addition, these methods risk classifying pieces as investigative when they are not, simply because the pieces meet certain external criteria.To minimize such biases, we designed our selection method based on internal criteria observable in the texts themselves.
For the purposes of this study, we have considered the label "investigation" in news content as an external criterion, as several scholars have determined that pieces labelled "investigation" were actually other forms of journalism (Abdenour 2015(Abdenour , 72-73, 2018(Abdenour , p. 1070;;Labarthe 2018, 119) or that the "investigation" label is "overexploited" (Bromley 2007;De Burgh 2008).Moreover, Lugrin (2000, 64, our translation) argues that the peritextual elementshere, the use of the investigation labelcan be used as indicators, but are not stable discriminatory criteria. 6Another way of identifying investigative work is therefore necessary.

Focus on the Newsgathering Process
Only a few studies have applied "internal" criteria -that is, criteria that relate to the content of the news -to separate investigative stories from ordinary news (Abdenour 2015(Abdenour , 2018;;Carson 2014Carson , 2020;;Cordell 2020;Forbes 2005).They all use a mix of criteria related to journalistic roles, topics, and process.
Firstly, they see investigative reporting as being limited to stories "in the public interest", which for them means holding the powerful to account and therefore playing a watchdog role.However, as van Eijk (2005, 21) points out, a story can be an example of watchdog reporting without necessarily being investigative reporting (i.e., applying an investigative process) and that, on the contrary, not all investigative reporting necessarily ends with an exposé.Furthermore, the "public interest" notion is often used to separate the "elite" press from the "tabloid" press or "hard news" from "soft news" (Wasserman 2020;Costera Meijer 2020).Studies refer to it to establish which topics are "valuable" or "prestigious" and which are not (Harcup 2015, 93-94;Shapiro 2014, 561).Therefore, we argue that watchdog stories can be seen as more in line with the ideal of investigative journalism as it is embedded in Anglo-American journalistic culture (Norris 2017, 220), but not as the only way to practice investigative journalism.
Following van Eijk (2005, 21-22), we thus argue that a non-elitist conception of investigative reporting can be limited to analyzing the newsgathering method.Thus, Rosenstiel (2007, 194) also link the investigative nature of a stories to the "journalistic enterprise".Cordell (2020, 124) also mostly focuses on the process, as his analysis "examined what indicators or characteristics of investigative journalism techniques were used to construct the story in question".Carson (2014, 732-733) and Forbes (2005, 12) do also essentially use criteria linked to the journalistic process.
However, their selection criteria are still theoretical and definitional.For instance, Carson (2014, 733) states that the story should be "an example of active journalism" and should "investigate", Forbes (2005, 6) that it should be "in-depth" and "proactive".They however do not explain which indicators within the news content allow them to establish that a story meets these characteristics.Our paper thus aims to take this additional step.

Methodology
To improve upon previous classification attempts, we drew on inductive approaches for (Strauss and Corbin 1998): we conducted an exploratory analysis of a randomly selected sample of articles, developed an identification method based on this analysis, and then tested this method using a larger, pre-selected sample.

Functional Definition
Following Carson (2014Carson ( , 2020) ) and Cordell (2020), we first conducted a comprehensive review of existing definitions of investigative journalism.This review included European literature, which tends to have a slightly different perspective on this subject than American literature (Eijk 2005;Labarthe 2018).We then identified the different characteristics specific to investigative journalism and compiled them to arrive at a functional definition: We will recognize a useful definition if it parsimoniously lists the special and required characteristics of an activity that may be recognized as constituting [investigative] "journalism" and not something else.But, while a definition is always exclusive, it is also as inclusive as possible.A desirable definition of [investigative] journalism will therefore embrace [investigative] journalistic work in a wide variety of forms, subject areas, and cultural contexts (Shapiro 2014, 559).
Our functional definition integrates various elements of the existing definitions consulted. 7It lists six properties, all related to the research process: 1. In-depth research 2. conducted independently 3. at the journalist's initiative, 4. focusing on information, links or interpretations that are difficult to access 5. and resulting in a rigorous demonstration of facts 6. for which the journalist assumes responsibility.
According to this definition, the "more" of investigative journalism is expressed as the presence of all six properties, each of which fall under the central category (Strauss and Corbin 1998, 146-148) of the journalist's commitment.Indeed, our functional definition implies that investigative journalism demands more commitment at each step of the journalistic process.This is in line with Cancela et al.'s (2021, 3) analysis of interviews of Swiss journalists, who consider "personal commitment" and the "extent of efforts provided during the investigative process" as the decisive criterion for whether journalistic work can be considered investigative or not.Ludwig (2011, 9-10) also proposes a continuum of journalistic commitment and suggests a correlation between this commitment and the type of sources used.For him, the more widely accessible the sources are, the less effort a journalist needs to make.Several other studies have suggested an inverse correlation between journalistic commitment and sources' influence on news contenti.e., the greater the commitment of the journalist, the less he or she will rely on any one source and vice versa (Berkowitz 2009, 105-106;Carlson 2009;Charaudeau 2011, 176).In this perspective, assessing the influence of sources on a story should be a reliable way to measure journalistic commitment.

Exploratory Analysis
To explore whether we could empirically assess the influence of sources on news content (and thereby indirectly evaluate journalistic commitment), we conducted an exploratory semi-inductive analysis of recent newspaper editions.This analysis covered all articles published during one randomly chosen week (28 May-3 June, 2018) by six media outlets in French-speaking Switzerland: 24 Heures, Tribune de Genève (two local newspapers that share most of their content), Le Temps, Arcinfo, Le Matin Dimanche and L'Illustré.This list was designed to ensure a wide diversity of press in French-speaking Switzerland by including newspapers with a variety of readerships (upper class, generalist, or popular) that are published both daily and weekly, provide national and regional coverage, and have large and small circulations. 8We focused on newspapers because print outlets are known as a major source of investigative journalism (Carson 2020, 151-153;Hamilton 2016, 280).
We began by intuitively grouping the articles into "investigative" and "non-investigative" categories.In certain cases, the decision was relatively obvious.For instance, the byline "comm" (short for communiqué, or press release) explicitly indicates that the news content is a press release.Conversely, some pieces could be easily identified as best examples of investigative journalism.For instance, the article entitled "Le Système Constantin" (Constantin's system) carried the "investigation" label and won the 2018 Swiss Press Award 9 , which commended the author on her "several weeks of investigation". 10We therefore considered articles with the "comm" byline as prototypes of passive journalism, and "Le Système Constantin" as a prototype of investigative journalismprototypes being those elements that are "categorized most easily" (Dubied 2004, 96, our translation).All other pieces could then be described as closer to one prototype or the other.
We then conducted a comparative analysis to identify articles sharing common characteristics."Thus, when we classify like with like and separate out that which we perceive as dissimilar, we are responding to characteristics, or properties inherent in the objects that strike us as relevant" (Strauss and Corbin 1998, 114).This progressive classification allowed us to discern six criteria for investigative pieces, each of which were empirical evidence ("traces") of the six properties in the functional definition given above.

Findings: A Multi-Step Method
From this point, we were able to build our multi-step method for identifying investigative pieces, as shown below [Figure 1].Each step reflects choices made intuitively during our exploratory analysis, which were then documented and refined by consulting the academic literature and through practical testing.
Step 1. Independence: Looking at the Byline According to the literature, investigative journalists establish both their agenda and the form and content of their story independently of any external influence: Whereas news deals very rapidly with received information, usually accepting what is defined for it by authority (ministries, police, fire service, universities, established spokesmen) as events appropriate for transformation into news, investigative journalism selects its own information and prioritizes it in a different way (De Burgh 2008, 15).Shapiro has argued (2013, 561-562) that independence is actually a characteristic of journalism in general.However, he defines independence in its literal sense as "connoting an arms-length interest in publication itself versus direct benefit from the consequences of what is published" and "a disconnection from the risks and benefits of propagation".In this view, even a press release can be considered journalistic even though the journalist does not shape the content in any way.For the purposes of this article, we use the term "independence" in a restrictive sense, to mean that journalists write the news content on their own.
To measure the proportion of independent journalism published in the media, Lewis, Williams, and Franklin (2008, 14) looked at the number of stories that did not replicate either PR, agencies or other media.Herman and Lugrin (2000, 8) also considered news agency reports as externally produced pieces, because they "are certainly endorsed by the newspaper but not reworked by it.As a result, they are not specific to a single newspaper."In this study, the authors used the byline to identify these types of stories.Following them, we identified copies of press releases or news agency reports by their byline (ex: "comm"; "red/com"; "ATS"; "AFP") and classified them as non-investigative.Opinion columns and readers' letters were also identified by their bylines and placed in the non-investigative category.
Step 2. Thoroughness: Large Quantity of Unrelated Sources The Flemish Network of Investigative Journalists (VVOJ) defines investigative journalism as a form of "in-depth" journalism (Eijk 2005, 22).Aucoin describes it as a "time-consuming 'digging'" (2005,91), Ingmar (2006, 31) as "intensive" research, and Lustenberger as "thorough" research (2010, 93).These descriptions imply that investigative journalists seek to accumulate as much information as possible from different sources, up to the point of "saturation" (Grevisse 2008, 115).To judge whether news reports reflect a willingness to "get as close to the truth as possible", Olsen (2018, 271) argues that "one of the most concrete markers (…) is the large number of sources used in the stories".Conversely, O' Neill and O'Connor (2008, 287) consider the use of only one or two sources as a sign that information has been processed only superficially.Using number of sources as a criterion allowed us to rule out articles that simply relay reports provided "turnkey" by external sources who carried out their own investigations, as professionals do not consider such pieces to be investigative.(Bastié, Tremolet, and Devecchio 2014;MacFadyen 2008, 139).We therefore excluded from our corpus all pieces that used fewer than three sources.
Conducting in-depth research also involves consulting sources that are independent from one other.According to Moloch & Lester (1974, 106-107), this requires researchers to consider whether the "promoter" of an event is the same as the "effector" of that event.
In some cases, the promoter and the effector may be two different persons or institutions but be so close to each other that they can be considered as one and the same body: "It is clear, for example, that if Richard Nixon's Press Secretary promotes the President's trip to China or Russia, the effector (Nixon) and the promoter (Press Secretary) can be taken as identical for all intents and purposes."One way to determine whether sources are independent from one another is to examine whether they share direct interests.On this basis, we excluded from the corpus all pieces that used only clearly related sources.
Step 3. Initiative: Outgoing vs. Incoming Sources Most existing definitions of investigative journalism specify that investigative stories are undertaken at the journalist's initiative (Mcquail 2013).However, the idea to delve into a problem often comes from an external source, usually an informant (Hamilton 2016, 136;Labarthe 2018, 113).For example, the "Contaminated Blood Affair" in France began when a hemophiliac contacted Anne-Marie Casteret, a journalist at L'événement du Jeudi (Hunter 2011, 19).During the discovery phase of a story, the information received by the journalist is often "non-publishable, being fragmentary, initial, undeveloped, misleading or inaccurate" (Reich 2006, 506).The journalist's role is then to decide whether to investigate this information further and if so, to what extent.According to Reich (2006, 506-507), journalists will typically evaluate their own initiative after the information-gathering stage is complete.
To assess journalists' and sources ' initiative, Reich (2006, 503) distinguishes between "reporter-initiated contacts" and "source-initiated contacts".Grevisse (2008, 40-41, our translation) calls them "outgoing" and "incoming" sources.Outgoing sources are sources that the journalist chooses to seek out and are assigned greater symbolic value by journalists.In our corpus, sentences such as "when contacted by telephone, X indicates that (…)", suggest that the journalist has initiated contact with the source.Conversely, indications such as "during a press conference" indicate that the sources have taken the initiative.We could usually infer whether sources were incoming and outgoing from signals in the text, and we classified articles based mainly on incoming sources as non-investigative.
Step 4. Responsibility: Going Beyond Juxtaposed Viewpoints Another essential difference between investigative journalism and regular journalism lies in the enunciative responsibility of the reporter (Moirand 2006).According to Ettema and Glasser (2006, 138), investigative reporters assume responsibility for "for the quality of the facts reported".They normally also bear responsibility for the consequences of their publications (De Burgh 2008;Grevisse 2008;Hunter 2011).This means going beyond simply presenting a series of opposing viewpointsa form that regularly appears in news media (Desjardins 2005, 282-283;Ericson 1998, 87) or a series of reactions to an event.As Harcup (2015, 97) states: Investigative reporting typically abandons the journalistic convention of allegation-anddenial, or attributed opinions, in favor of an attempt "to establish facts which, if possible, decide the issue one way or the other" (Spark 1999).
We therefore excluded from the investigative category any article relying on a compilation of opposing viewpoints (Carson 2020, 78), as well as those relying on a compilation of reactions or comments.
Step 5. Difficulty in Procuring Information: Hard-to-Access vs. Accessible Sources Unlike basic reporting, which tends to treat quotations from authorized sources as facts (Carlson 2009;Ericson 1998), investigative journalism takes a critical stance towards information from official sources (Bastié, Tremolet, and Devecchio 2014;Eijk 2005, 22;Lustenberger 2010).This implies that investigative journalists are less interested in information made available voluntarily than in information not intended to become public.Several authors speak of information "kept secret", which implies that actors may deliberately seek to hide certain facts (De Burgh 2008;Ettema and Glasser 2006;Houston 2009).
However, this interpretation is often contested for being too restrictive (Eijk 2005;Ludwig 2011, 9).De Fleur (1997, 18) argues that investigative journalists are looking for facts that are "not easily discovered".For Abdenour (2015, 59), hidden information denotes information that is either: "(A) deliberately suppressed; (B) concealed by law (…); or (C) public but compiled by the reporter in a way that exposed a widespread pattern".Ingmar (2006, 29, our translation) states that all journalistic investigation entails facing "resistance" and requires the journalist to make a "special effort" to obtain the desired information.He argues that "the mere compilation of previously unpublished facts is therefore not sufficient for being qualified as investigative journalism.In addition, there must be particular difficulty for obtaining the information". 11In this perspective, exclusivity of information is not a sufficient criterion.However, information does not necessarily need to have been kept secret either.Rather, it should be hard-to-reach, which implies that the process of obtaining it should be in some way arduous.
To obtain hard-to-reach information, journalists must use or speak with hard-to-reach sources.We called these "hard-to-access" sources, as opposed to "accessible" sources, 12 i.e., available and comprehensible without any special effort.In the corpus, several articles mention that people refused to speak to the journalist, or only agreed to speak anonymously.Others mention having to process large databases, or many documents.In the end, we excluded from the corpus all pieces based solely on accessible sources, i.e., human sources that cooperated voluntarily or documentary sources that did not require any processing.
Step 6. Evidence: Documentary and Other Sources Investigative journalism requires journalists to prove their conclusions.As Hugo De Burgh (2008,20) states, investigation is "dispassionately evidential".In this view, investigative journalists meticulously accumulate pieces of evidence and assess their respective weight (Ettema and Glasser 2006).Generally speaking, journalists assign more weight to documentary evidence than to other types of evidence (Hunter 2011, 8;MacFadyen 2008, 143;Shapiro et al. 2013, 667).As van Eijk states: Documents are the ultimate facts.They speak for themselves, literally, without the journalist interposed between them and the audience.They are the textual equivalent of photographs, which in the eyes of many have a truthfulness of their own.(2005, p. 239) Journalists use documents to confirm or disprove different versions of events and thus avoid setting one person's word against another 's (Ericson 1998, 83).They also use certain types of documents, such as laws, guidelines or statistics, to prove that a transgression has occurred (Glasser and Ettema 1989, 10).We therefore classified any remaining article that made use of documents as investigative.However, we did not make this an eliminatory criterion, since documents are not always available (MacFadyen 2008, 138).Articles that did not quote any documents were classified as investigative if they provided other evidence for their conclusions, such as a greater number of human sources (Hunter 2011, 14) or a confirmation of the journalist's assertions by independent experts (De Burgh 2008, 16).
To sum up, we classified as non-investigative any article that: (1) was written by a person or institution external to the newspaper; (2) quoted only a single or two source (s); ( 3) quoted mainly related or ingoing sources; (4) merely juxtaposed viewpoints or reactions; ( 5) used only accessible sources; or (6) did not present documentary or other evidence.Conversely, we classified as investigative any article written by a staff journalist that quoted three or more hard-to-access, mostly outgoing and unrelated sources, that favor one thesis (over others) and provided evidence to support its conclusions.

Practical Testing
To test our method, we applied it to a sample of 819 articles published by the same six media outlets between 1 January and 30 September 2018 and pre-selected by us as good candidates for investigative pieces. 13After applying the criteria to each article [Figure 2], we arrived at a final corpus of 186 articles that could reasonably be considered investigative.
When practically tested on the sample, the selection method worked essentially in an exclusive way, with each step narrowing the sample further.Applying the internal criteria to the pre-selected sample allowed us to exclude a whole series of very common articles, such as news agency reports, editorials, political debates, or analyses.
The 186 articles in the corpus reflect a wide diversity of forms, topics, narrative styles, investigative methods, enunciators, and media outlets.Our functional definition and the multi-step method derived from it therefore appeared to be effective, since they were sufficiently broad yet also sufficiently precise to build a meaningful corpus (Shapiro 2014, 559).

Discussion and Conclusion
Our initial analysis of the 186 articles in our corpus reveals that when investigative journalism is defined as a specific process instead of by normative claims regarding what it should be about, 14 it is not limited to topics traditionally associated with investigative reporting, such as law, crime or fraud (Hamilton 2016, 55,77-81).Alongside articles denouncing, for example, a major drug fraud (Le Matin Dimanche, January 14, 2018) or the conflicts of interest of a para-public institution (24 Heures, May 19, 2018), our corpus includes, among others, an article on a car accident, which reconstructs and narrates the course of events to identify the cause of the vehicle explosion (Arcinfo, January 18, 2018), another portrays the murderer of a young woman (L'Illustré, September 19, 2018), another reveals and describes the poor financial situation of a hockey club (Tribune de Genève, January 5, 2018), and another highlights and explains the ever-growing distress of teachers in dealing with troublesome students in public schools (Le Temps, April 9, 2018).Indeed, as Descamps observed (2017, 274), "desacralized" definitions of investigative journalism, such as those commonly held in Europe (Eijk 2005) and specifically Switzerland (Cancela, Gerber, and Dubied 2021), allow for a theoretically infinite number of topicsincluding local topics affecting ordinary citizens, such as health, taxes, sport, accidents, or even everyday events.Moreover, the articles in our corpus do not all denounce someone/something; some narrate, explain, or describe facts or situations.

A Blurred Discursive Genre
Similarly, the results of our preliminary analysis suggest that the articles in the corpus are too diverse to be classified in the same discursive genre.Most traditional classification methods organize journalistic genres on a continuum between "distance/information" and "involvement/commentary", with investigative journalism located at the midpoint between these two poles (Dubied 2004, 50).However, if we were to place the articles in our corpus on such a continuum, they would be all over the map: some would be closer to the more "factual" genres, others to the more "subjective".Indeed, the results of our corpus analysis support the idea that investigative journalism transcends the traditional journalistic genres (Agnès 2015, 207-295), as it contains reports, portraits, analyses, and more, that reflect an investigative process. 15 From Small to Big Investigations In addition, the articles included in the corpus all reflect a higher personal commitment than those excluded from it, by virtue of the fact that they met all six of our criteria for investigative journalism.However, the corpus showed great variation with regard to the degree to which each of the six criteria were present.For example, all the articles used three or more sources, but some used four, while others used sixty-two.Similarly, some sources were harder to access than others: some journalists used public (open source) documents but analyzed them in an innovative way, while others used anonymous sources, confidential documents, or data collected by hand.The degree of initiative taken by the journalist also varied: stories ranged from follow-up investigations to leaked stories, tip-off stories and project stories (Eijk 2005, 241-243).The evidence used to support journalists' conclusions ranged from irrefutable evidence to different clues leading to the same conclusion.With regard to responsibility, journalists also adopted various positions: some denounced transgressions themselves, others used experts' words to do so, and still others only implied a denunciation, leaving the interpretation up to the reader.
Moreover, these criteria were not interdependent: a story might show a high degree of one criterion (e.g., initiative) but a low degree of another (e.g., evidence).We therefore suggest that the commitment reflected in investigative pieces should not be assessed on a single continuum but rather on a gradual, multi-part scale, as illustrated below [Figure 3].
Our concept of investigative journalism as a specific process that can materialize in a theoretically infinite number of forms and degrees is in line with previous research (Cancela, Gerber, and Dubied 2021).However, our corpus offers the first opportunity to observe how different degrees of journalistic commitment materialize in investigative pieces.

In Favor of Internal Criteria
To test the validity of our selection method, we compared our corpus with theoretical ones, which would have been composed through external criteria.On the one hand, our aim was to check that our method did not exclude the best examples, which would have been just as misleading as focusing only on them.As van Eijk noted (2005, 16), the "best examples" do not generate a definition of investigative journalism, but any sound definition should at least include them.In the same vein, we considered that any sound selection method should be able to include them.
We considered the two main awards institutions of French-speaking Switzerland for the year 2018: the "Newspaper category" of the Swiss Press Award 16 and Le Prix Dumur 17 , which awards the whole work of one journalist.All the awarded pieces were present in our corpus, which means that they all meet our selection criteria.Secondly, we considered the articles quoted as "good examples" by interviewed journalists (Author, YYYY).Among the 104 examples quoted, five concerned our period of study and the newspapers under scrutiny.Only one of them was not included in our corpus, but only because it was too short (less than 6000 characters).Furthermore, none of the "bad examples" made it in our corpus.Some studies about investigative journalism focus on media outlets deemed as producing investigative journalism (Cordell 2020;Valeeva 2017).In Switzerland, there is no such media outlet, at least among the newspapers.Nevertheless, we could have focused our attention on reference national newspapers and exclude local newspapers, since the former are known to do more investigations than the latter (Kaciaf 2018).However, we found investigative pieces in every newspaper under consideration, which shows that local and popular newspapers or magazines are also producing investigative pieces.
Finally, most of the articles included in the investigative category were written by a bunch of journalists, who are identified in the French-speaking journalistic field as "investigative journalists" (Author, YYYY).Most of them are part of formal or informal investigative units.An alternative method would have been to select the articles produced only by this bunch of investigative journalists.However, the authorship does not mean that the articles are investigative for sureespecially as most of these professionals also cover topics on a more daily basis.Moreover, our corpus also contains several "one-shot" investigations (N = 53), whose authors appear only once in our corpus, are not well-known investigators and are used to covering daily, local news.These "one-shot" investigations reveal that the practice of investigative journalism is not limited to established investigators: ordinary journalists can also carry it out from time to time.
In sum, our corpus shows that our selection method also includes the few identified best examples of the best journalists, while also including a wide range of unexpected non-standard articles.We thus assume that a method based on internal criteria might offer a more accurate picture of the state of contemporary investigative journalism than a one based on external criteria.

De-Mythicizing Investigative Journalism
Our method is promising in that it allows for a coherent description of contemporary investigative journalism in all its possible variations.It is based on our claim that examining underlying sources is an effective means of identifying pieces that fall into the category of investigative journalism.In this view, underlying sources are reliable indicators of journalistic commitment, a criterion which the literature commonly uses to evaluate investigative journalism.Following previous studies, our method inversely correlates journalistic commitment with sources' influence, and this study's practical tests have demonstrated that it is indeed possible to assess investigative commitment by analyzing an article's quoted sources.
Furthermore, our method avoids perpetuating a mythicized ideal of investigative journalism, as reflected in films (Harcup 2015, 98;Matheson 2009), professionals' discourse (Conan et al. 2004;Zelizer 1993), and even academic literature, which has thus far focused on "best examples" (Ettema and Glasser 1988;Valeeva 2017;Wahl-Jorgensen 2013a).To increase our understanding of the practice and nature of investigative journalism, as well as the transformations that lie ahead, it is essential to disentangle our methods from the powerful myths that surround this form of journalism.This article is a first step in that direction.
However, we think that our proposed method is of particular interest not only for Swiss researchers, but also for any scholar interested in investigative journalism.First, the Swiss information media landscape shares important characteristics with that of other Northern European countries and can thus be considered representative of that region (Hallin and Mancini 2004).Swiss journalists' work conditions and perception of their role are also similar to those of their peers in the global North (Bonfadelli et al. 2011;Bonin et al. 2017;Hanitzsch 2011).Second, our method provides important insights into what investigative journalism encompasses when it is defined as a specific process and not by publisher, topic, or awards received.
We however acknowledge that such position is not all-encompassing and that, like all other classification attempts, our remains debatable, as their depend on the point of view or the perspective adopted, and the cultures observed.We are aware that some might consider them too inclusive, especially journalists or researchers who strictly associate investigative journalism with the role of watchdog (De Burgh 2008, 19-20;Abdenour 2018Abdenour , 1059Abdenour -1060)).On the contrary, some might still consider the criteria set as too exclusive, for instance in countries where the process of investigative reporting is extremely difficult to implement (Norris 2017, 219-220;Márquez-Ramírez 2020).
The Transparency of the Process in the Product Finally, our paper opens the door for a discussion about investigative journalism as a discursive genre.The selection criteria for our method are based on explicit or subtle "traces" (Ricoeur 1985, 123-144) of the investigative process that appear in the text of news articles.We decided to take this approach despite Reich's assertion that academics should not examine the product in order to infer the process because this would be predicated on two unfounded assumptions: "that news processes are encoded into the product and that researchers are able to decode them reliably" (Reich 2006, 501).Although Reich's concerns are legitimate, in our view there is no way to avoid inferring the process from the product when identifying investigative pieces.We have argued here that investigative journalism is best defined as a process; the logical corollary is that the best way to identify works of investigative journalism is to examine the process that led to the product.Of course, it is impossible to reconstruct the full details of a journalist's process from news content alone.However, our method does not involve reconstructing the process, but only detecting traces of it in news content.
Indeed, we would suggest that it is precisely those traces that constitute the particularity of investigative journalism as a discursive genre.In other words, the discursive elements common to all the articles in our corpus are those traces of a higher personal commitment and, thus, of an investigative process.This is an important argument in favor of transparency of the research process in news contenta phenomenon that is a clear contemporary trend in journalism (Hellmueller, Vos, and Poepsel 2013;Karlsson 2011).It also has important implications for audience perception and reception: if readers cannot see any traces of an investigative process in a story, then they cannot be sure that it is a genuine investigation.As Randall (2016, 233) puts it: "The reader should never have to ask: 'How does the paper know this?'"

Further Research
Our further research will focus on analyzing the articles in our corpus in greater detail, looking at themes, topics, narrative styles, authors, and the persons or institutions targeted by the investigations.Our aim will be to provide a first picture of the breadth of contemporary Swiss investigative journalism through a detailed description of its various forms of expression (including the most recent and creative examples).
This article has highlighted the fundamental importance of the journalist-source relationship in investigative journalism, and this would also be a promising angle for future research.In the context of our corpus, a deeper analysis of the sources used in each article could provide an insight into the ecosystem (Nielsen 2015) of investigative journalism in Switzerland.
Moreover, it has already been shown that who or what is quoted is as important as how that source is quoted (Bednarek 2016).This involves the positioning of the journalist in relation to what is written/said/shown and, more globally, the reshaping (Wahl-Jorgensen 2013b, 308) of reality or events through narrative.Here, too, investigative journalism is of particular interest, since scholars (Marchetti 1997, 49;Wuergler 2021, 1-2) consider it to be the form in which the journalist's influence on the configuration of the news content is the greatest.
Finally, our selection method led us to exclude from the corpus 39 articles (among the 819 in the initial sample) labelled "investigations".Conversely, we included in the corpus many articles not labelled "investigations".This discrepancy between our selection criteria and the use of the "investigation" label by the media would deserve more attention.Indeed, it would be interesting to understand what criteria seem to govern the editorial choice to qualify an article as an "investigation" or not.In any case, at this stage, the use of the label seems to be rather "underexploited" than "overexploited".

Notes
missing investigations that did not contain any of these keywords, we also gathered all the articles of more than 10,000 characters, published during the same timeframe, without looking for special keywords.14. (e.g., 'public interest' concept; 'powerful people/institutions' (Harcup 2015, 95).15.For more details on the discursive analysis of the journalistic investigations included in our corpus, see Authors 2023.16. https://swisspressaward.ch/de/swiss-press-award-2019/ 17. https://www.prixjeandumur.ch/communique.php?a=2018

Figure 1 .
Figure 1.Multi-step method for identifying investigative pieces.

Figure 2 .
Figure 2. Procedure for building our corpus.