Previous Section: Overview & Introduction

Critical Interpretive Research: An Introduction

Blocks 1 and 2: Description

The first two Blocks dealt with approaches to description in the empirical-analytic and interpretive paradigms respectively. Each paradigm has its own sense of what description requires. Empirical-analytic inquiry aims to describe phenomena by using measurement instruments. Interpretive inquiry, in contrast, aims to describe phenomena through the accounts of the people involved with them.

    The classwork for these and several subsequent Blocks, both empirical-analytic and interpretive, involved materials drawn from the evaluation of a curriculum reform project that had been designed to foster critical thinking and achievement through classroom discussion. In the empirical-analytic Blocks these materials were also used for the Block Projects; in the interpretive Block Projects students collected their own materials.

Block 1: Empirical-Analytic Description -- Psychometrics (5 weeks).

The first empirical-analytic block introduced psychometric approaches for conceptualizing, quantifying, and interpreting educational achievement variables, for investigating their reliability and validity, and for reporting the results to educational stakeholders in light of different purposes for assessment. The Block Project was to write a report describing scores on an interpretive writing task, and to assess the reliability and validity of these scores as an outcome measure for the discussion program described above.

Block 2: Interpretive Description -- First-Person Accounts (4 weeks).

It is important to recognize that the phenomena that empirical-analytic inquiry seeks to describe and measure have already been grasped and interpreted by the people who encounter them in everyday practice (Taylor, 1971/1979; Giddens, 1983; Thompson, 1990). Interpretive inquiry, in contrast with empirical-analytical inquiry, aims (as a first cut) to describe how social phenomena are experienced by the people involved with them. Empirical-analytic research assumes that rating scales of opinions and attitudes can, if necessary, be "added" to supplement the primary, objective description of a social phenomenon. But to an interpretive researcher, this addition of a "subjective" component to an "objective" description remains suspect and incomplete. An interpretive researcher wants to know how people grasp, understand, and interpret events and artifacts.  We can say that "in social inquiry the object of our investigation is itself a pre-interpreted domain" (Thompson, 1990, p. 275), so that the "object domain" of investigation is at the same time also a "subject domain" (Thompson), or an "intersubjective domain". (Recommended reading for the Block: Mishler, 1986; Gordon, 1980.)

Detailed Syllabus for Block 2
A way to obtain first-person accounts
How to investigate the way a person grasps and understands an event, a situation, or an artifact? One way is to obtain a first-person account from them. This can be done by conducting an interview

A interview can be treated as a way of soliciting factual information, or even a stimulus-response sequence of questions and answers. But to an interpretive researcher, an interview is an opportunity to solicit, as skillfully as possible, extended discourse in which the interviewee provides an account of some topic. 

It follows that as researcher we need to learn both how to conduct an interview, and how to analyse the material thus obtained.  In Block 2 these were considered  in reverse order, first introducing the analysis of an interview transcript, then turning to the question of how best to conduct an interview. (The rationale for this was that one needs to know what one is looking for before one can learn how to seek it.)

Argument & Narrative analysis Some researchers analyse interview transcripts by coding and categorizing. But this type of analysis can easily lose sight of the structure of the account. The extended discourse solicited in an interview is structured as narrative and argument. These two are the main modes of any discourse (Kinneavy, 1971); indeed they can be viewed as  two primary modes of humanthought (Bruner, 1986). When a person answers an interviewer's questions they generally use both kinds of discourse structure, both telling stories and making claims. If we want to describe (and then go on to interpret and explain) how someone understands a situation, event, or artifact, we need to examine the narratives and arguments with which their account is organized. 

Narrative, the first kind of structure to discourse, organizes events in time within an orientation or plot (cf. Carter, 1993). As Connelly and Clandinin (1990) put it: "The main claim for the use of narrative in educational research is that humans are storytelling organisms who, individually and socially, lead storied lives. The study of narrative, therefore, is the study of the ways humans experience the world." There are a variety of approaches to narrative analysis (cf. Prince, 1987), but whichever one chooses, one needs to read an interview transcript both for its plot (what is being talked about, and for its narrating (how it is talked about).

Plot will include elements such as the characters involved, the steps that constitute the action the characters take, the goal being sought, the setting, any trouble that arises for the characters (every good story involves trouble!), the instruments used to overcome the trouble, and the outcome of it all.
The narrating of the story can include a variety of different narrative devices that give emphasis (flash-backs and flash-forwards, ellipsis, repetition, comments and evaluation interposed by the narrator), and the presentation of a point of view from which events are recounted.

Together, attention to plot and narrating can help us grasp the theme of the account -- the point the narrative is being used to make. Put in a nutshell, our analysis aims to articulate how the narrating presents a plot in order to propose a theme.

Argument, the second kind of structure to discourse, organizes not with a plot, but by making assertions or claims and by providing justifications in their support. As is the case with narrative, there is no single approach to the analysis of argument. Stephen Toulmin (1958) did ground-breaking work on the structure of argumentation, while more recent approaches to argument analysis have attended to its interactional and pragmatic character.  Whichever approach to analysis one chooses, it must involve identifying the elements of an argument -- the claims being made, the grounds upon which these claims rest, the warrants put forward for the claims, any qualifiers, possible rebuttals, and backing -- and articulate the chaining of claims in the discourse (cf. Toulmin, Rieke & Janik, 1974).
How to Interview? How should one solicit the discourse that contains these narratives and arguments? How to conduct a good interview? Does one emulate Terri Gross (of PBS' Fresh Air) or Larry King (of CNN's Larry King Live)? Again, there is no single answer, but it must be emphasized that an interview is best understood as a conversation, not a question-answer exchange. For example, as Eliot Mishler (1986, p. 82) puts it, "a 'story' is a joint production" that serves what Michael Halliday (1973) has called the interpersonal function of discourse. 

A good interview requires good conversational and interpersonal sensitivity. Traditional survey researchers have developed numerous strategies, techniques and tactics for conducting semi-structured interviews (cf. Gordon, 1980) and along with these a vocabulary which we can coopt in order to talk about the aims of an interpretive approach to interviewing (e.g. McCracken, 1988). 

* Skillful interviewers need to attend to the scope of questions. Narrow scope questions can be answered with a single word, such as "Yes" or "No"; broad scope questions -- "How come?" -- call forth extended responses. 

* We need to pay attention to the sequence in which questions are posed. 

* We need to know how to ask probe questions that will encourage elaboration. Often the best probe is silence; novice interviewers tend to speak too much and too soon.

* We need to be sensitive to when and how to initiate a new topic and when and how to relinquish topic control. Novice interviewers generally change topic too soon. 

* We need to make informed decisions when selecting interviewees, including sampling considerations. Often sampling will not be random but instead "representative," "convenience," or "special". 

* And we need to know how to properly obtain informed consent. 

Readings for this Block: Bruner, 1986; Toulmin, Rieke & Janik, 1974; Carter, 1993; McCracken, 1988, pp. 9-34.

  Week 3 introduced some general considerations, while students were busy conducting their own interviews for homework and their Block Project. The transcript of an interview is a text. The report of an interview analysis is a text, too.  In a whole variety of ways, practitioners of interpretive inquiry must become familiar with collecting, manipulating, and interpreting textual material, as well as writing texts of their own.

The talk that is elicited in an interview is fleeting and evanescent. It is here, then gone. If it is to be studied systematically it must be "fixed" in some way. Tape-recording and transcribing are the familiar ways in which interviewers do this. What is not so obvious are the subtle but important ways that recording and transcribing transform speech.  (We shall return to this point later, drawing on the work of Paul Ricoeur, 1971/1979.)

When an interview is taped-recorded and then transcribed, decisions must be made about how to format the transcript, and what conventions to use to indicate pauses, hesitations, repetitions, gestures, and paralinguistic phenomena. These choices will reflect theoretical assumptions. In a whole variety of ways a transcript is never simply a verbatim record. Transcription is a selective process (Ochs, 1979);

And when the researcher writes a report, important choices must be made about the appropriate level of detail or abstraction, the use of "emic" or "etic" terms (Geertz, 1976/1979), the choice of first or third person viewpoint, and so on. We looked briefly at some of the options for genre and format when writing an interpretive research report. (Cf. McCracken, 1988, pp. 34-48; Ochs, 1979; Richardson-Koehler, 1987; Smith, 1987; Urban, 1990; with Mishler, 1986 recommended.)

Its influence on interpretive research

In the final week of this Block, Week 4, we considered how the European philosophical movement known as phenomenology has provided one of the roots of the interpretive paradigm (cf. Bernstein, 1976, pp. 115ff; Giddens, 1976; Rabinow & Sullivan, 1979; Dallmayr & McCarthy, 1977). 

Phenomenology was originally an attempt to step back from and suspend the naive realism of our ordinary way of experiencing, to reflect on the active and organized character of experience as a complex whole. This attempt now seems misguided -- we can't strip away the language, the culture, the habits of everyday life, and if we could, what would be left worth knowing? But that kind of "transcendental" phenomenology did highten our awareness of the ways human beings do, in an important sense, "constitute" our world. 

Conceptions of this "constitution" (cf. Bernstein, 1976, p. 160) have ranged considerably, from Edmund Husserl's notion of a "transcendental ego" that applies abstract, universal, mental structures to the basic data of experience, to Harold Garfinkel's very different notion of the human subject as a mundane, everyday "member" of a social group organizing the social world by practicing methods of skilled improvisation.  Here is a Summary Table of the conceptions of constitution. (cf. McCracken, 1988, pp. 48-66; Taylor, 1971/1979; Kvale, 1986.)

Next Section: Blocks 3 & 4

© Martin Packer, 1999

Home Page

Developmental Psychology

Cultural Psychology