Perspectives on the Clinical Interview as an Interactive Genre

• In Event: Integrating Issues of Knowledge and Interaction in Analyses of Cognition and Learning

Abstract

Clinical interviewing has a long history as a way to determine subjects’ ways of understanding and coming to understand (Gruber & Voneche, 1977). It remains commonly used by learning researchers (Clement, 2000; diSessa, 2007; Ginsberg, 1997).

Clinical interviewing is provocative and potentially fertile as a subject of study from the complementary perspectives of knowledge and interaction. It is provocative because it has been criticized as a research method from interactional and socio-cultural points of view (Bannon and Bødker, 1991). Criticisms include that it is ecologically invalid, that clinical interaction is very likely to introduce interactional artifacts in apparent “knowledge,” and even a skepticism concerning the validity of knowledge, itself, as stable and transportable to new contexts. It is fertile because, although the methodology is mainly of interest to knowledge analysis (KA) researchers, it seems likely that interaction analysis (IA) researchers are in the best position to understand its properties.

We report on the first of a series of exploratory studies of clinical interviewing as a genre of interaction aimed at determining the ways in which IA and KA problematics, methods and empirical strategies may be (a) deeply synergistic, (b) complementary, or (c) contradictory. Our data is a series of 7 one-hour video-taped clinical interviews conducted by one of our team, focused on intuitive knowledge of physics.

We focus on the strategy of revoicing (O’Connor & Michaels, 1993) as it functions in clinical interviews. Revoicing was identified first as an interactive strategy used by classroom teachers to promote academically productive interactions among students. Revoicing in science discussions has been shown to serve many functions, including highlighting productive student moves, “bending” students’ language toward normative science, and positioning students in particular roles in scientific inquiry. Our analysis involved developing a grounded coding scheme, which included the following slots: the intended or resultant function of the revoicing (12 codes); the form of the revoicing (3 major codes, 6 subcodes); interviewee response (9 codes); and interviewer follow-up (7 codes).

Result highlights:

• Intensive use: Nearly 30% of interviewer turns during the hour that was coded were classified as revoicings.
• Function: The functions of revoicing are distinct, sometimes dramatically distinct, from those previously identified for instructional revoicing. One distinctive function is merely to provide time for interviewees to reconsider prior contentions, often resulting in change. Difference in function compared to classrooms poses a number of questions, such as what processes establish the specializations of this strategy to align with distinct goal structure of different contexts.
• Diversity: Revoicing seems to come in very many, often very different, forms. This provides a challenge to defining revoicing as a coherent category.
• Interviewer knowledge: We see indications of articulate expertise of the interviewer with respect to the “design” of revoicings, similar to explicit strategizing by teachers, previously documented. Some aspects of interaction seem to be knowledge-based, which, in part, accounts for the “when” of revoicing, and specific shaping according to context.

Authors

• Andrea A. diSessa, University of California - Berkeley

• James G. Greeno, University of Pittsburgh

• Sarah Michaels, Clark University