Induction in Journalism

28-05-2008

BROUSTAU, N., LE CAM, F., (2006), « Enlightening Induction in Journalism Studies. A Perspective for Researchers and Research », Annual conference, Canadian Association of Communication, Université York, Toronto, 1er juin-3 juin, (www document), http://florlecam.com/slj//lectures/induction-in-journalism/, Mis en ligne mai 2008, Dernière consultation le ……

***

Enlightening Induction in Journalism Studies
A Perspective for Researchers and Research

Nadège Broustau, Ph.D., Laval University, Québec, Canada. Nadege.broustau.1@ulaval.ca
Florence Le Cam, Ph.D., professor assistant, University of Rennes 1, France.
flecam@univ-rennes1.fr

Abstract :
As researchers in journalism studies, we are always confronted with difficulties of empirical studies. How can we better know what is happening for and from the actors ? What can we say about their discourses and practices ? The challenge for the analysis is to take distance from the sequence of daily events and to produce a scientific knowledge about media and journalism. For this purpose, we have both used an inductive perspective as an epistemological position as well as a methodological concern. Our text presents this approach from its controversy in philosophy of science, to its explanation in methodological handbooks. To make clear the way induction can really be an original and rigorous approach, we illustrate it with the results of our two Ph. D theses on media and journalism.

Text :

Our Ph. D. theses on media and journalism (1) have confronted us with the target of empirical research and the way journalism studies can not only be a description of what is happening in the field, but can really and progressively contribute to build the sense of contemporary changes. We have both used an inductive perspective as an epistemological position as well as a methodological concern. It means that we have chosen to immerse ourselves in our fields during long periods of time, without presuming of our results. This approach is often seen as an easy way taken by young researchers to begin a study. But what we want to explain is that the inductive perspective is more than an approach to discover the field, it is also an original and rigorous way to explain the process of production of science.
What our works make clear is that using induction in journalism studies helps moving off-center the look at media and journalism and taking into account new tendencies in media history, professional sociology and argumentative media analysis. While journalism is facing particular challenges, especially with blogs, converged journalism, citizen journalism, using induction enables researchers to analyze and to justify their thoughts (epistemology) or choices (methodology) but also to explain the process of production of the knowledge they can bring to their colleagues or the actors of the field.

In the common sense, “induction refers to any kind of inference in which we move from a finite set of observations about an ‘object’ or a ‘concept’ to a conclusion that is a general description of the object or the concept” (McCreath, 1999 : 23). This is in contrast with deduction. However, the search for a definition of induction in the literature on methodology brings a wide range of meanings that misleads its understanding. Two tendencies have emerged from our literature review : induction has first been the major issue of a debate within the philosophy of sciences at the turn of the 20th century (in particular Lachelier, 1872 ; Mill, 1866), then it has been assimilated by manifold discourses and reworked by specialists in methodology like Huberman and Miles (1994), Lindlof (1995), Mucchielli (1996), Lessard-Hébert and Goyette (1997), Guibert and Jumel (1997) and Grawitz (c2001). Induction is sometimes conceived as one of the features of qualitative approaches (Taylor et Bogdan, 1984), but it is often perceived either in opposition to hypothetico-deduction, or as one of the steps of the scientific knowledge process following deduction and abduction (Peirce, 1867 ; Pellisier-Tanon, 2001). We tackle here the volatile character of induction, but this review gives the opportunity to recall old and new debates on the notion.
Using our theses as support, this article is the first stage of a research project (2) concerning induction in journalism studies. We try to link different disciplinary perspectives to illustrate the relevance of the inductive perspective and its use in two research studies. To begin with, we will describe the historical evolution of the notion of induction through a review of literature of some philosophers of science. Induction is then at the center of an old controversy about the production of knowledge. Secondly, we will focus on the appropriation of induction in texts on methodology in social sciences and in communication studies, to see how the notion is presented and used by contemporary researchers. Finally, we will present some significant results of our theses to illustrate the inductive perspective and the way it brings to our field of study, journalism studies, opportunities to show scientific relevance, to make the reproduction of the study realistic, to assure the results of the research and to propose fertile theoretical notions.

1. Short history of ‘induction’

The notion of induction has been a central issue in the work of many philosophers since the 16 and 17th centuries and it has sometimes been called the ‘scandal of philosophy’ (Holland, dir., 1986). The classical definition of induction refers to the passage from the individual or the species to the gender. For the purpose of this article, only those philosophers who have made significant contributions to the evolution of the concept will be presented : Bacon, Hume, Mill, Popper and Peirce. The literature on this topic is vast and cannot be summarized in few paragraphs. We will focus here on two main debates that have emerged : the first one refers to the experiment and to ‘causal relation’, and the second advancement refers to the status of hypotheses in the inductive logic.

1.1 Experiment and causal relation

The Novum Organum, a text written by Bacon in the 17th century, has become a major text for theorists interested in induction. In this text, induction is defined as a careful and meticulous process, based on the experiment, and as something that rises in degrees towards generalities. Induction can lead to the truth and to the knowledge of nature, through the elimination of errors (Blanché, 1975 : 91). Bacon was the first to “tabulate the fundamental logical ideas of experimental enquiry” (Cohen, 1970 : 125).

This work will inspire those who, more than one century later, will question induction and the trust that we put in it. The 18th century is the field of major changes about the conception of the world which has two fundamental consequences (Blanché, 1975 : 93-94) :
1- the desacralization of the expression : ‘Laws of nature’. These laws referred to the decrees taken by God to manage the universe and became questioned in the 18th century.
2- the criticism of the ‘causal relation’. This relation referred to the fact that the letter A is always followed by the letter B. During the 18th century, a question emerges : how can we be sure that the relation will always be like this ?

This context makes the bases of induction a major issue. In the seminal text, Treatise of Human Nature, Hume never uses the word ‘induction’. But he was working the conclusions provided by experimentation… He first discussed the causal relation, arguing that we cannot know a thing without having experimented it before. The role of experimentation is to create, by repetition (3), what we expect to be an effect after a cause (4). Hume wanted to reveal the rules that could let us evaluate causes and effects (Cohen, 1970 : 125). So the main philosophical problem was : how to guarantee the effect after the cause in the future, even if we have proved it from the past ? This became a modern problem of philosophy (Blanché, 1975 : 96). Hume’s response is to return to Providence to solve this uncertainty. In his opinion, induction is then a forecast, and not anymore a generalization. Induction can help to imagine effects in the future (Blanché, 1975 : 100).

Many years later, John Stuart Mill starts to work on the subject and returns to the classical definition of induction as a generalization of experiment. At the beginning, he chooses to understand the notion as the means of generalizing cases from particular cases. For example, we have seen that many crows are black ; we can then conclude that in general crows are black. Mill wants to make the experiment the central point of his theory. For him, induction must determine effects from causes and it must distinguish the causes from each particular effect (Blanché, 1975 : 102). He explains that deduction (as a perspective progressing from general evidence and constructed hypotheses to particular conclusion) depends on induction : “Induction is the true reasoning that makes deduction possible” (Boss, 1990 : 55). Even in mathematics, numbers have an experimental base : to learn numbers, we count things (Boss, 1990 : 59). Mill wants to build a theory of induction that would create of methodology of experimental search.

The major point of the controversy about induction focuses on the ‘generalization method”. In The Logic of Scientific discovery, Popper has defended the position that classical deductive reasoning is necessary for sound inference and that induction is impossible and does not exist (Greenland, 1998 : 543). The logical-positivist movement in philosophy has combined the use of the accurate methods with a strict empiricism to be applied to deductive methods (Holland, dir., 1986 : 6) (5). Even the proponents of induction recognize the importance of deduction to scientific progress. Mill, for example, defends the deductive logic by integrating induction in the process. To him, the necessary method in science implies three stages : a direct induction, a deductive reasoning and a checking process (Boss, 1990 : 71). In fact, induction is not always opposed to deduction in the philosophers’ point of view, but they often treat the question separately.

Status of hypotheses

Since Mill, understanding a process like induction is a central issue of philosophy, psychology, mathematics, statistics, biology or artificial intelligence… This issue has often focused on the role of hypotheses in the scientific process. Popper explains that we never use arguments “based on observed repetition of instances that does not also involve a hypothesis” (Greenland, 1998 : 545). For him, the method must test hypotheses by attempting to refute them (Aliseda, 2003 : 342). The inductive perspective is based upon the discovery of things, the logic of heuristics.
Yet what seems to have never been examined is the difference : “between the concept of induction as a method of discovering acceptable solutions to scientific problems and the concept of induction – or rather, of inductibility – as the converse of a logical or quasi-logical, relation of support between propositions” (Cohen, 1970 : 128). Nevertheless, induction is often perceived as ‘unexpected’ for the researcher. In observing, he might not even think of hypotheses revealed by observation : but data themselves generate hypotheses (Greenland, 1998 : 545).
Peirce raised a central question about hypotheses by asking : ‘How do people avoid generating innumerable fruitless hypotheses in their search for useful generalizations ?’ (Holland, dir., 1986 : 4). Peirce responds by referring to the idea that induction is based on innate knowledge, and has developed a third possibility between induction and deduction : the abduction. Abduction is “thinking from evidence to explanation, a type of reasoning characteristic of many different situations with incomplete information” (Aliseda, 2003 : 340). In this third possibility, the scientific process is built from a “recursive circle” between abduction, induction and deduction that never ends (Pellissier-Tanon, 2001).

Most research on induction that we have met in this short history has made clear one thing, as Holland wrote : The study of induction is the study of how knowledge is modified through its use (Holland, dir., 1986 : 5). The scientific practice is then a process in itself.

2. Induction in methodology

From epistemology to methodology, induction is unequally described and reasserted, above all for didactical reasons. In order to bring out the current definitions of induction given in methods guide books, we first explored qualitative methods research in communication and media and we moved to more general guides in social sciences.

We have found five central themes that have emerged from our review :
1- the epistemological dimension of induction
2- induction as one of the qualitative elements
3- induction as a specific element of the interpretive paradigm
4- blurring the difference between induction and deduction
5- the absence of induction

2.1 From epistemology to methodology : a first step in general handbooks

The epistemological dimension of induction is asserted by some researchers like Grawitz (1991) in order to enlighten the different research procedures. Grawitz’s definition of induction uses platonician versus aristotelician thinking so as to present deduction versus induction, as a result of the confrontation between rationalists and empiricists. Grawitz’s definition is interesting because it lets us perceive the movement from contingency to ontology in induction while the deduction starts from the ontological status of things.
Quivy and Campenhoudt (1988) also propose to tackle induction from an epistemological point of view. Unlike Grawitz, they develop their definition of induction as a rigorous logic process that leads to the building of operational definitions.

Induction as one of the qualitative elements

Wimmer and Dominick (2006) use induction to distinguish qualitative and quantitative methods. Indeed, contrasting with the quantitative approach, “qualitative researchers use an inductive method : Data are collected relevant to some topic and are grouped into appropriate and meaningful categories ; explanations emerge from the data themselves” (Wimmer and Dominick, 2006 : 116).

Other authors, like Taylor and Bogdan (1984 : 5) or Patton (1990), target induction as the first and major feature of qualitative methods. Mucchielli also quotes Taylor and Bogdan to define induction. Apart from the great value of the definition of these authors of qualitative methods in general, these quotations end up in a tautological way : it is qualitative because it is inductive, it is inductive because it is qualitative.

Induction as a specific element of the interpretive paradigm

Lindlof and Taylor (2002) present the interpretive paradigm as the ontological and epistemological base of qualitative methods. They point out that, in this perspective, “theory should be developed inductively through the iterative testing of tentative explanations against the experience of ongoing interaction with group members” (2002:11).

Induction thus qualifies all the stages of the research :
1- as a qualitative way to approach data (in an inductive mode of analysis) (6)
2- as a way to analyze and interpret data (7)
3- as a way to produce knowledge (8)
Lindlof (1995:56-57) enhances as well the inscription of qualitative methods in the interpretive paradigm and, since this inscription, the inductive essence of qualitative methods. This specificity is due to the search of understanding over prediction or control, a search where explanations emerge from data themselves.

That is what Kirby, Greaves and Read (2006) illustrate in their definition of inductive labelling. To them, “inductive labelling means that each label grows out of the material in each bit, rather than being applied to a bit from some preconceived notion of what groups of data will appear.” (2006 : 240)

2.2 « Blurring the Difference » (9) between induction and deduction

As we saw it in the general methods guides in social sciences and in the philosophical texts, induction is often defined in opposition to deduction. For instance, Horning Priest (1996) defines inductive logic versus deductive logic. Inductive logic « involves reasoning from a specific case to a general theoretical conclusion », and then, as Mill enhanced that induction precedes deduction, « the new theory could be tested « deductively ». On the contrary deductive logic « reasons from the general to the specific », begins with a theory that can « predict » and then tests it (Priest, 1996 : 8-9).

Lessard-Hébert, Goyette and Boutin (1990 : 95-97) also stress the inner dialogue between induction and deduction. They quote Erickson (1986), for whom even if the qualitative researcher doesn’t enter a field with “any specific predetermined categories of observation […], he enters the field with a conceptual frame and some interests of research. Searching in the field, induction and deduction are constantly dialoguing” (10). There, the theorization process depends both on inductive and deductive logic.

In the same perspective, Schwandt (2007) enhances that inductive analysis is “by no means unique to qualitative work – much of probability theory and statistics is part of this logic of induction.[..] Typically, qualitative analyses employ some combination of inductive and deductive analyses. The claim that qualitative studies are inductive may actually be a way of saying that they reject the hypothetico-deductive method of the natural sciences.” (2007 : 147)

The absence of induction

Finally, it was surprising to find that some qualitative guide books in communication didn’t tackle induction at all. It was the case for example in Asa Berger (2000), Silverman (2004) and Huberman and Miles (1984).

The link between qualitative methods and induction is then not as obvious or inextricable as it seemed.

3. Discussion for journalism studies

As students we often placed our work between qualitative and quantitative approaches, which frequently means between inductive or deductive logics. Inductive process seems a position adopted by many students and researchers to assume the empiricist way they have taken. But, what we have seen in this study is much more complex than these discursive justifications. A deep analysis has to be made not only to justify the researcher’s thoughts (epistemology) or choices (methodology) but also to explain the process of production of knowledge (11).

3.1 Analyzing professional identity

In one of our theses related to the Quebec group of journalists and its confrontation with Internet (Le Cam, 2005), we assumed to say that it presupposes inductive, qualitative and iterative research methods (12). For the purpose of the study, we had to stay a long time in the field, meeting journalists, observing newsrooms, trying to understand what was happening to the journalistic practices. We had chosen an inductive methodology : starting from observation and experiment to understand the problem and to create hypothetical explications.

As the field of Internet is extremely unstable, we had to adjust our methodological choices. We thought at that time that it was the heart of the inductive position to be so close to your field that you could adjust yourself to its changes. Then, while weblogs appeared in the field of the online information diffusion, we decided to analyze them as a part of our study, even if we knew that the webloggers were not part of the Quebec group of journalists, but often outside its traditional boundaries. We assumed that, by looking at our subject from a different point of view, we reinforced the inductive logic. The analysis of the discourses and the practices of webloggers made then evident certain discursive strategies and the mechanism of the constitution of the identity of our professional group of journalists. The research was not only a question of adjustment. The observation had to use the researcher’s intuition step by step. It had to follow directions not even thought of at the beginning.

The field made extremely relevant the fact that what was happening on the web and to the field of information raises a challenge to professional journalists as a group in Quebec. By placing in question the very terms “journalist” and “journalism”, and by undermining its image, the Internet weakens the professional journalistic group identity and its historical specificity. This idea became the ‘direction’ of the research ; it became the general hypothesis of the research. Then, while we assumed an inductive perspective, we ‘fall’ in a sort of hypothetico-deductive process. But the logic was still the same. The research needed an iterative method, a position of the researcher between the experimentation and the validation, and the validation and the experimentation, etc. of the research at each step of the field, by trial and error. We were using a logic of abduction without knowing it.

3.2 Analyzing a media argumentative representation

The other thesis aimed at the study of the American media argumentative representation of the Elian Gonzalez story in 1999 and 2000 (Broustau, 2007). This international and controversial story led to a public debate between the American government and the Cuban government, between federal and local authorities, and between different ethnic communities in Florida. From this debate emerged ideological polarized positions. The research analyzed the argumentative coverage of the Elian Gonzalez story in three American newspapers (13).

Beyond the qualitative nature of the study, our inductive position first occurred during the data collection for the corpus. The research first intended to study the whole argumentative coverage, so the corpus first entailed all the articles tackling the Elian Gonzalez case. But as we were testing the argumentative analysis technique, it became obvious that not all the articles were relevant to the media argumentative representation we were looking for. Our attention stretched upon opinion articles as a genre and after putting this genre to the test, upon this genre as a significant field to study argumentation in its persuasive dimension. As a result, we chose to focus only on the editorials and the columns. This simultaneous data collection/corpus construction and data analysis was the first sign of inductive analysis, as this proceeding was a first bifurcation in the subject studied.

A second bifurcation reinforced and confirmed what was in fact our Grounded Theory position. We refer here to Paillé’s adaptation of Glazer and Strauss’ Grounded Theory (1994). This bifurcation happened during the data analysis. As an example, we use here the categories found to characterize the columns. After the analysis of the polarized arguments, began a work of categorization. One categorization dealt with what has emerged specifically from the columns. A first big emerging difference existed between texts based on fiction (or projection) and texts based on personal experience. But the category “fiction” was too literary and did not convey enough the argumentative dimension of our study. Therefore, we tested different categories that could correspond to the tone of the texts, the particular way of writing a column through fiction : we tested caricature, irony, satire, images d’Épinal, on several scattered texts of the corpus. But we always found negative cases. So, we reconsidered our categorization to take into account the persuasive purpose of argumentation. Rereading the texts based on fiction, we found that the logical foundation of them was to push the reasoning to its limits : that was the absurd reasoning. Then we reworked the categories to entail this particular way of writing a column through fiction in an absurd way in order to persuade or to argue. And the final categories were “demonstration based on absurd logic/reasoning” versus “demonstration based on personal experience”. Observation of the data (the arguments in the texts) that led to a constant movement of categorization through the test of negative cases was a specific process of induction.

A reflexive practice on research done during and after the thesis points out three important points :
1- Inductive logic is fundamental to discovering a field of study. It leaves surprises come to the researcher. If we had chosen a deductive methodology, the weblogs and other alternative sites would not have probably emerged. It helps you to analyze your object from different perspectives. That leads to rediscover the parameters of the context.
2- This rediscovery appears during the bifurcations of the initial subject of research. In the inductive perspective, the initial frame of research is made to be “overmoved”. It refers to what has been called the iterative process. It means that you proceed by trial and error.
3- Inductive methodology must be distinguished from the process of production of knowledge. The choices you make while you are investigating your field are research hypotheses. It can be intuition at the beginning but you have to make choices, and theses choices must be explained as hypothesis within the course and the continuation of the study (14). It means that you need to build – and to be conscious of – your hypotheses to pursue your research.

In the process of production of knowledge, you finally always mix inductive and deductive logics because you take choices, advance some ideas, return on the field, as if you had adopted an “hypothetico-inductive” method.

The challenge is to explicit this dialogue between inductive and deductive moments in order to ensure the validity of the research. Here is what is at stake :
1- the characterization of hypotheses in an inductive process
2- the fact that the researcher has to take on her responsibilities for her interpretive work.

Conclusion and further challenges

The inductive perspective can help researchers to adopt a reflexive position : it implies to take into account and to assume the status of researcher, that is to say the duty of writing a scientific report of a research in order to make the reproduction of the study realistic. The researcher has the duty to be understood and to show how he has proceeded in his work. She explains the choices, the trial and errors she makes and every modification she imposes to the knowledge she wants to build. She is then in a constructivist methodological position which respects the influence of the researchers on the data. An inductive perspective also encourages to move off-center the look at journalism and “traditional” medias. It makes the discovery and the surprises from the field clearer. It allows researchers to adjust their methodology, their tools and to take new ways of research. It lets them follow the movement of news and media : to take into account the potential innovation of news media and new events. By following the movement, we mean to encourage researchers in journalism studies to integrate sociological, philosophical, historical, economical and linguistic perspective, as also encouraged by Zelizer (2004). This last point intends to imply studies of audiences, content, actors and context.

This is especially true now that big changes have occurred and are still in progress in the media milieu : the social role of the media and the social perception of journalists and journalism have been transformed by the development of internet and by the convergence of economical property and of journalistic practices in the media. These changes raise philosophical issues, based on ethical questions about the nature of journalism and of information.

Journalism is part of everyday life and everybody seems to have something to say about it. So it can be hard for researchers in journalism studies to pull them out of the field and out of the journalistic discourses. Adopting an inductive perspective helps researchers to understand the ‘everyday life’ of journalists and media and to avoid applying popular beliefs in their researches by pushing preconceived ideas aside and by letting the meaning emerge from the data. It is part of a sociological comprehensive perspective which gives to actors the competence to explain their own practices.

Notes :

(1). Broustau, N., (2007), La trajectoire argumentative des représentations médiatiques dans les textes d’opinion en presse écrite : le cas Elián González dans le Miami Herald, le Washington Post et le New York Times, Québec, Université Laval, Ph.D. Thesis, August 2007, 483 p.
Le Cam, F., (2005), L’identité du groupe des journalistes du Québec au défi d’Internet, Québec, Université Laval et Rennes, Université de Rennes 1, Ph.D. Thesis, June 2005, 555p.
(2). The initial project was presented at the Canadian Communication Association in June 2006. The title of the conference presentation was « A few questions about induction in journalism studies. Debates on the notion and their perspectives for researchers and research », Annual conference, York University, Toronto, June, 1st-3.
(3). ‘Hume’s preoccupation with the causal analogue of enumerative induction is the direct ancestor of a similar tendency in certain twentieth-century psychologists’ theories of learning. To explain learned behaviour by reference to a repeated pattern of stimulation and rewarded response, for example, is to confine one’s explanatory model to processes of enumerative induction” (Cohen, 1970 : 126).
(4). “Hume’s preoccupation with the causal processes analogous to enumerative induction was the archetype both for the obsession with stimulus-response learning that has obstructed psycho-linguistic investigations and also for the obsession with the probability-calculus that has issued in philosophical theories of induction like those of Keynes, Ramsey, Nicod and Carnap” (Cohen, 1970 : 124).
(5). See : Swann, Andrew, J., (1988), “Popper on induction”, British Journal for the Philosophy of science, vol. 39, n3, pp. 367-373.
(6)Qualitative research being primarily inductive, emergent and somewhat unruly (2002:66). Moreover, following the Verstehen of Dilthey and the “nouvelle science” of Giambattista Vico, induction is, with history, the necessary way to study cultural forms so as to understand human nature and its creations (2002:31).
(7). For example, in interpersonal communication research : explanations of personal identities and social realities are « developed through the inductive analysis of data generated in intimate observation » (2002:20). Inductive approach is also shown to be paramount in feminist theory and methods (2002:57), in ethnography of communication (2002:44)
(8). Because, for instance, « a strong current of inductive thinking stimulates the development of categories… » (2002:15)
(9). The phrase is from Horning Priest, 1996 : 9.
(10). Lessard-Hébert, Goyette and Boutin (1990:96)
(11). For further examples, see the review Recherches qualitatives, vol. 26, (2), 2006, and especially Guignon and Morrissette (2006), Blais and Martineau (2006).
(12).Theses methods had to be applied to various aspects of the problem : an analysis of the history of journalism in Quebec through the study of professional and trade-union documentation ; a study of the impact of the new technology on the journalistic milieu, specifically the study and analysis of the practices developed by online journalists on the sites of various traditional media ; and finally, the specific active study of the practices of select online publications by means of a personal weblog.
(13).The Miami Herald, The Washington Post and The New Yok Times, from November 1999 to June 2000. The analysis of the arguments was based on Toulmin’s Uses of arguments (1959) reframed by Liakopoulos (2000).
(14). For example, Erickson (1986 : 139) speaks of a deliberative process during the data collection. To him, “there are no pure inductions” (quoted in Lessard-Hébert, Goyette and Boutin, 1990:97).

Bibliography

Aliseda, A., (2004), “Logics in Scientific Discovery”, Foundations of Science, vol. 9, (3), pp. 339-363.

Asa Berger, A., (2000), Media and communication research methods : an introduction to qualitative and quantitative approaches, Thousand Oaks, CA, Sage, 295p.

Bertrand, I. and Hughes, P., (2005), Media research methods : audiences, institutions, texts, New York, Palgrave MacMillan, 286p.

Blais, M., Martineau, S., (2006), “L’analyse inductive générale : description d’une démarche visant à donner un sens à des données brutes’”, Recherches qualitatives, vol.26, (2), pp.1-18.

Blanché, R., (1975), L’induction scientifique et les lois naturelles, Paris, Presses Universitaires de France, 170p.

Boss, G., (1990), John Stuart Mill. Induction et utilité, Paris, Presses Universitaires de France, coll. Philosophies, 121p.

Broustau, N., (2007), La trajectoire argumentative des représentations médiatiques dans les textes d’opinion en presse écrite : le cas Elián González dans le Miami Herald, le Washington Post et le New York Times, Québec, Université Laval, Ph.D. Thesis, August 2007, 483 p.

Cohen, L.J., (1970), The implications of induction, London, Methuen, 248p.

Guignon, S., Morrissette, J., (2006), “Quand les acteurs mettent en mots leur expérience”, Recherches qualitatives, vol.26, (2), pp.19-38.

Grawitz, M., (2000), Méthodes des sciences sociales, Paris, Dalloz, 1019p.

Grawitz, M., (1991), Lexique des sciences sociales, Paris, Dalloz, 376p.

Greenland, S., (1998), « Induction versus Popper : substance versus semantics », International Epidemiological Association, vol. 27, pp. 543-548.

Holland, J.H., (dir.), (1986), Induction : processes of influence, learning and discovery, Cambridge, Mass., MIT Press, 385p.

Horning Priest, S., (1996), Doing media research : an introduction, Thousand Oaks : Sage Publications, 265p.

Huberman, and Miles, (1984), Qualitative data analysis, Beverly Hills, CA : Sage, 263p.

Kirby, S., Greaves, L. and Reid, C., (2006) Experience Research Social Change – Methods Beyond the Mainstream, Peterborough, Ont. : Garamond Press, second edition, 280p.

Largeault, J., (1979), Hasards, probabilités, inductions. Petits écrits de circonstance, Toulouse, Université de Toulouse-Le Mirail, 202p.

Le Cam, F., (2005), L’identité du groupe des journalistes du Québec au défi d’Internet, Québec, Université Laval et Rennes, Université de Rennes 1, Ph.D. Thesis, June 2005, 555p.

Lessard-Hébert, M., Goyette, G., Boutin, G., (1990), Recherche qualitative : fondements et pratiques, Montréal, Agence d’ARC, 180p.

Lindlof, T. R. and Taylor, B. C., (2002), Qualitative communication research methods, Thousand Oaks, CA, Sage, 357p.

Lindlof, T. R., (1995), Qualitative communication research methods, Thousand Oaks, CA, Sage, 314p.

McCreath, E. (1999), Induction in First Order Logic from Noisy Training examples and Fixed Example Set Sizes, Ph.D., The University of New South Wales, October.

Mill, J. S., (1843), Système de logique déductive et inductive. Exposé des principes de la preuve et des méthodes de recherche scientifique, Livre III : De l’induction, édition électronique, Chicoutimi, Bibliothèque Paul-Émile Boulet de l’Université du Québec à Chicoutimi.

Mucchielli, A., (1996), Dictionnaire des méthodes qualitatives en sciences humaines et sociales, Paris, Armand Colin, 275p.

Paillé, P. (1994), « L’analyse qualitative par théorisation ancrée », Cahiers de recherche sociologique, no 23, 34p. (pp. 147-181)

Pellissier-Tanon, A., (2001), « L’induction, au cœur du dilemme des savoirs en sciences de gestion », Stratégies, Actualité et futurs de la recherche, Paris, Vuibert, FNEGE, août, pp.55-66.

Quivy, R. and Campenhoudt, L. van, (1988), Manuel de recherche en sciences sociales, Paris, Dunod, 271p.

Schwandt, T., (2007), The SAGE dictionary of qualitative inquiry, Los Angeles : Sage Publications, 322p.

Silverman, D., (2004), Qualitative research : theory, method and practice, London, Sage, 377p.

Strauss, A. and Corbin, J., (1990). Basics of qualitative research : grounded theory procedures and techniques, Thousand Oaks, CA, Sage, 270p.

Taylor and Bogdan, (1984), Introduction to qualitative research methods : the search for meaning, New York, Wiley-Interscience, 302p.

Wimmer, R. D. et Dominick, J. R., (2006), Mass media research : an introduction, Belmont, CA, Thomson et Wadsworth, 468p.

Zelizer, B., (2004), Taking journalism seriously. News and the Academy, Thousand Oaks, Sage publications, 286p.