and inquiry. The concepts of Climate and Engagement represent characteristics of the class as a whole, which in turn also have important influences on the thought and collaborative activities that students engage in. Finally, the concepts of Scientific Thinking and Inquiry capture the students' on-going intellectual accomplishments in their collaborative activity, which we view as the ultimate goal of the teacher.
While informal observations of the seminar discussions suggest that this is a plausible account of the causal relations among the six teaching concepts as viewed by the student teachers, we carried out a more formal analysis of the reasonableness of this model as representing the views of seminar participants as they evaluated videotapes of teaching. The statistical approach we chose was a path analysis of the correlations among the six scores assigned by individual graduate students to each of the 45 videotapes. This analysis is based on writing a recursive system of regression equations reflecting the possible causal pathways of influence among independent (Pedagogy, Management), intervening (Climate, Engagement), and dependent variables (Scientific Thinking and Inquiry), and then testing the significance of the path coefficients. Separate path analyses were carried out for each independent variable (Pedagogy and Management). Figure 9 shows the influences among variables in the analyses for which we found statistically significant path coefficients. For instance, there was evidence of a direct influence of Pedagogy on Thinking, but no evidence of a direct effect of Management on Thinking. And the effect of Climate on Thinking is only indirect, through its influence on Engagement. The path analysis model corresponds well with the implicit causal model we had posited linking the teaching concepts within the conceptual framework. It is of considerable interest that the apprentice teachers appear to have developed this model without its directly being presented to them in the seminar materials. And again, the data analysis is based upon individual scores generated by each student teacher and thus the correlations demonstrate that the causal model was shared by the participants in the seminar.
Our analyses have shown that the participants in the seminar developed a shared perspective on teaching that was facilitated by introducing an initial reflective framework and then using it as a basis for evaluating videotapes of teaching. This shared perspective was a reinterpretation by the participants of the set of teaching concepts that we introduced. Our remaining research question is whether the use of this conceptual frame was worthwhile in facilitating productive discussions of teaching.
Since we had videotaped all of the seminar meetings, we could address this through an analysis of the conversations about teaching that took place in the seminar. We selected six tapes of video club conversations about student teaching and four tapes of conversations held in evaluating non video club tapes. The recorded seminar conversations were then parsed into conversational units by the four graduate students in the research group, based upon where there were changes in the discussion topic. This yielded a total of 108 conversational units. Each of these units was then classified and rated using a scoring system developed by the research group.
Of interest here are two particular features: Framework Association, which deals with whether or not the language used in the conversations made explicit reference to the framework we had introduced, and Productiveness, which is a rating of whether or not the conversation was thought-provoking and reflective, with many members of the group actively participating. Interscorer reliabilities for these judgments were 90% and 94%. In our analyses we found that there was a significant influence of Framework Association on the Productiveness of a conversation. Of the conversations that made reference to the conceptual framework, 82% were rated as productive, compared with 56% for conversations that were not framework related. This effect of framework use was independent of whether the conversation was in the context of a video club or was an evaluative discussion of a tape that was being scored. It was also independent of whether or not a faculty member played a role in precipitating the conversation. It thus appears that the conversations that most successfully explored the processes of teaching involved an explicit use of the interpretive framework of teaching goals and process. Thus cognitive facilitation has helped to open up the complexities of interpreting classroom videotapes to a systematic exploration and analysis by this group of preservice teachers.
We have found that introducing a conceptual and linguistic basis for metacognitive reflection was effective in promoting the collaborative inspection of cognitive activity. We have demonstrated this in two contexts,