Section IV
Conducting Research in the Interpretive Paradigm
In this section, we transition from metric research to interpretive research. We do so by looking at the requirements for studies that investigate texts, first from the metric framework and then from the interpretive framework. The purpose of this transition is to show the basic commonalities that are part of both of these frameworks. From there, Geoff Baym leads us through the intricacies of the critical interpretive analysis. Finally, we return to the action of ethnographic studies.
As I did in the introduction to Section III, let me remind you of the assumptions that form the borders of the epistemological domain you will reach at the end of this section. We will be walking away from causality, certainty, and closure and into agency, multiplicity, and indeterminacy; away from objectivity, measurement, and the individual and into epistemological standpoints, interpretive frameworks, and the community; away from well-conventionalized method and into emergent forms; away from the dominant and into the disruptive. It will be a slow but steady transition, and in the end, we will be in a very different place.
Interpretive research focuses on the meaningfulness that arises not from some intrinsic set of characteristics, but from cultural, societal, and membership practices. Meaning is a joint production of human actors. The purpose of interpretive methods is to document the practices and to craft the narrative of the multiplicity of this socially constructed meaningfulness. The research narrative itself is multiple. Instead of reaching closure, there is always another story to be told that will enlarge our understanding of the social world in which our lives emerge. The analyst is center stage, not as the skill technician of metric research and not as the discoverer or great explorer, but as a cultural agent, an instrument of interpretation. I hope you enjoy the tour.
CHAPTER 12
Coding Text
CHAPTER PREVIEW
Whatâs It All About?
Coding is a process that seeks to reach below the surface manifestations of symbolic and discursive texts for the purpose of identifying the underlying characteristics, structures, social meanings, and cultural work that such texts have, produce, or enact. Coding is done in both metric and interpretive research. In metric research, one typically has a template that guides the investigation; in interpretive research, that template emerges from the engagement of the texts.
What Are the Major Topics?
Texts are not the production of some innocent author. Rather, they appear at the intersection of the author, the industry, the media, the audience, the act of interpretation, and the demands of the text itself.
Metric coding starts with an investigation of theory and prior research; continues to the development of the sample of text, the classifications that will be used to code the texts, the instructions for the independent coders, the quantification of those codes, and their statistical analysis; and ends in their contribution to theory.
Interpretive coding starts with the selection of texts; continues to the close reading of those texts, during which codes will begin to emerge, and the iterative processes of coding and recoding; and ends in their contribution to theory development.
Whether metric or interpretive, all coding begins with a thorough understanding of the facts to texts. Gathering these facts can be greatly aided by computer analysis.
Texts can be coded in word processing programs, spreadsheets, database applications, and dedicated programs.
What Special Terms Are Used?
Database program
Digitized text
Live text
Meta-communication
Reflexivity
Window of text
INTRODUCTION TO CODING TEXTS
A text is any symbolic or discursive form. A symbolic form might be a framework of dance, a set of photographs, or a visual narrative as well as a genre of music or natural sound recordings. Discourse is any extended language and symbolic use that is under some common governance. A classroom lecture, a magazine article, a strategic communication campaign, a newspaper story, and even a tweet (although not very extended) are all discursive forms, because there are implicit rules that govern topic, construction, word choice, and performance that are recognizable in their performance and in their violation. The analysis of texts is the bridge between metric and hermeneutic epistemologies and methodologies. The analysis must be both factual and interpretive in order to be complete. Theorists and methodologists, naturally, disagree on the relative balance between these two. One of the reasons for this disagreement is that the balance changes according to the purposes of the analysis.
For an example of how this balance works, we could think of a study that examines the effect of contemporary newspaper layout design on the amount of news content. Many newspapers have gone from a 54-inch broadsheet to a more âreader-friendlyâ 48-inch width.1 Modern designs make a heavy use of âwhite spaceâ; use large-font callouts, ârefers,â and promos; and put a strong emphasis on visual elements (number and size). All of these elements would appear to reduce the amount of unique story content. In taking this approach, newspapers appear to be following the trend of many manufacturers of reducing the standard âcontainer sizeâ (a half-gallon of ice cream is now 1.75 quarts) in order to charge the same for less product. For newspapers, reducing the ânews holeâ (or making a graphic element larger to fill it) means that less story product is needed with the attendant savings in reporting costs.
A study of this sort is heavy on the factual side of the balance beam. We have limited interest in the meaning or meaningfulness of the content; we are mostly interested in the factual characteristics of that content. Certainly, we have to put into place some definitions and reliable procedures to identify white space, nonstory print components, visual elements, and the dependent variable of unique story content. These requirements do not appear to be high in difficulty, however.
They also donât tell us very much, simply increasing the precision by which we can describe what is a fairly obvious set of consequencesânewspapers print all the news that fits. Letâs change the question to investigate the quality of the news that is printed. The loss of capacity may not be as important if the quality of the reportage has increased in equivalent measure. A very recent survey taken by the Associated Press Managing Editors (APME) indicates that âseventy-one percent of the survey participants said cutbacks have âsomewhat affectedâ or âgreatly affectedâ the quality of their newspapersâ coverage.â2 More than 100,000 jobs have been lost in the industry over an 18-month period.
This clearly is a much more difficult question than that of measuring column inches of white space. I will need some set of criteria that define the quality of a story. Those criteria would appear to be both intrinsic to the construction and writing of the story itself and extrinsic to the event that is being reported. To build the first set of criteria, I would probably go to the style manuals that provide the conventions of good reportingâthe inverted pyramid; who, what, when, where, why, and how; local angle; multiple sides; good sources; and other norms. (Notice that these criteria have no connection to the quality judgments of an actual reader. The analyst cannot directly conclude about readers or audiences from content data.)
The second set of criteriaâthe extrinsic setâis much more difficult to specify. I would need to have some standard of the event or circumstances of the story. That standard does not exist. Consequently, I might use a set of experts to evaluate the story for both its factual content and its utility for understanding the implications of the event (why it is news).
Quite clearly, none of these criteria is simply factual. The conventions of âgoodâ journalistic writing are in part arbitrary, in part practical (the inverted pyramid is said to have developed to preserve the meat of the story should most of it be lost in telegraphic transmission), and none guarantee good writing, which itself is a judgment. The value of the criteria is that they are not idiosyncratic to the analyst and can be justified to others. A modernist scholar would call them objective; a postmodernist would call them culturally embedded. All of them appear in an act of judgment exercised by the analysts. Our confidence in those judgments vary. The presence or absence of white space is nearly unequivocal; whether the who, what, when, where, why, and how of something has been adequately presented is not.
Metric scholars manage these equivocal conditions by developing a set of rulesâa coding manualâthat directs the activities of the coders toward the theoretically determined properties of interest. The coding manual itself is developed in iterative pretesting, in which codes are added, divided, eliminated, or combined as needed until the code set creates a good fit between theory that directs the study and the properties of the content. Metric coding reports on the degree of agreement among coders as evidence of its nonidiosyncrasy or objectivity with 100% being desired and 60% to 70% considered publishable. Disagreements are resolved by adjusting the coding rules or, if irresolvable in that manner, by a third party, usually the principal investigator.
Hermeneutic scholars (and hybrids who do coding) typically use a grounded approach that eschews theoretical direction but builds the coding scheme in interaction with the texts. When a unit of text doesnât appear to fit any available code, a new code is provided. Subsequent analysis might divide, eliminate, or combine codes. When multiple coders are used, disagreements are expected, considered a valuable insight, and usually negotiated to a new code conclusion (if the disagreement is intractable, the senior author wins). In the end, the primary differences between the two approaches are the starting pointâmetric in theory and interpretive in the discourseâand the final form of the argument. In the middle of things, there are not many differences between the best coding practices of each approach, though those differences are hotly defended.
SYMBOLIC AND DISCURSIVE PROTOCOLS
Symbolic and discursive protocols most often address questions of the properties of content, but they can also participate in the analysis of processes, consequences, and quality. Our introductory example of the newspaper is primarily a study of properties, but if conducted as part of a pre- and postlayout changes protocol, it becomes a consequences analysis. And, clearly, the second half of the example is intended to speak to the quality of current reporting. Finally, the whole of the analysis might be a good explanation for what appears to be the death spiral of newspapers as we currently know them and begins to speak to economic and information theory.3
Four Intentions
Symbolic and discursive protocols can also be understood by the intentionality or intentionalities they serve. There are four sets of these intentions: the intention of the author (industry), the intention of the text, the intention of the auditor (audience), and the intention of the interpretation (auditor and analyst). These four sets work both extrinsically on the object of study and intrinsically on the research argument (research is discourse too).
For example, as I write this section, my intention (authorial intention to the extent that I can know it) is to enlarge the discussion of metric approaches to content analysis in order that a more inclusive view can be obtained. My âhandlersâ at Sage Publications (industry intentions) will advise and edit based on their own set of goals for the writing. The fact that I am writing a textbook imposes pedagogical requirements on the writing (intentions of the text) but also allows for a more personal âteachingâ style.
For most textbooks, the intention of the auditor reflects the primacy of the instructor as she or he adopts the book. It is these adoptions that create the audience. Both author and industry intentions have to account for this set of auditor intentions. Both author and industry presume that the instructor wishes to produce a competent course. The question for all threeâauthor, industry, instructor/auditorâis the extent to which this section of writing helps achieve that goal.4
And what of the intentions of the audience created in the course adoptions? Is there a model reader here? Can we assume any global characteristics of the members of the reading class? I think it is fair to presume that most are not here because of their choice, or, at least, the choice was not of the book but of the course. Are the majority motivated to learn, get credit, or do something of both? If it is just for credit, then my burden is less, because in creating a list of intentions, I have created an excellent short-answer question for an exam for which the reader can get credit. If it is to learn, my burden is greater, because I have to be sure that I demonstrate the value of this heuristic for analyzing discourse. Weâll see.
Finally, we arrive at the intention of the interpretation. Consideration of this intention is probably the most contentious and has the least presence in the literature. It is contentious because of the lack of agreement as to where this event (the interpretation) occurs. For some, it is behind the eyes in some cognitive process that has behavioral consequences. For others, it is in front of the eyes in some actionâthe interpretation is not fully formed until one does something that entails the text. It has the least presence in the literature because of the methodological obstacles to its investigation.
How that behavior or action entailment is conceptualized is also problematic. Is it behavior that can be independently addressed without reference to the context of its performance or to the regularized practices of the individual? Or is it action that is embedded in an ecology of performance and further informed by the sociocultural protocols of competent enactment?
Finally, there is the question of the relationship between the interpretation as performed by an auditor (audience) and the interpretation as performed by an analyst (discipline). Can the analyst speak for the auditor? In metric research, do the marked properties of the content reliably direct the interpretation of the audience? In hermeneutic studies, does the interpretation generated in analysis resonate in the multiple contexts of reception?
The four (plus the parenthetical three) intentionalities provide us with the justification for careful analysis of content. That analysis allows us to draw conclusions about the approaches of an author, the standards and conventions ...