One goal of teacher education is to transfer the knowledge base and the skills that enable student teachers to enact lessons of high quality. Classroom-based internships that are assisted by cooperating teachers, mentors or coaches are of pivotal importance in this context. A model that provides directives on how to assist student teacher learning during interships is Content-Focused CoachingSM (Staub, 2001; West & Staub, 2003). A core element of this approach, as transferred to initial teacher education, is the highly collaborative nature of the interaction between student teacher and cooperating teacher, which includes content-specific cooperative lesson planning. In a quasi-experimental intervention study two groups were compared: In the intervention group 16 student teachers were coached by cooperating teachers who have been trained as Content-Focused Coaches, whereas in the control group 16 student teachers were assisted by experienced cooperating teachers, who have not been trained in Content-Focused Coaching. Student teachers were randomly assigned to a cooperating teacher of the intervention or the control group. We hypothesized that the kind of assistance and learning suggested by Content-Focused Coaching has a positive effect on the quality of teaching demonstrated by the student teachers. The cooperating teachers of both groups did not differ significantly on a number of relevant variables (e.g., number of years of teaching experience, number of years working as cooperating teacher, pedagogical content beliefs in mathematics, beliefs on assisting student teachers during internships) at the beginning of the intervention study. The quality of teaching aimed for is subject-specific as well as of interdisciplinary nature (Klieme, Lipowsky, Rakoczy & Ratzka, 2006).

To assess the effect of the intervention, a mathematics lesson of each student teacher was videotaped at the end of a seven-week internship. An expert in didactics of mathematics rated the videotaped lessons based on high-inference scales. The rater was not informed beforehand whether the lessons were taught under control or experimental conditions. The rating scales (Ditton & Merz, 2000) included 18 items of instructional behavior (e.g., articulateness, differentiation, classroom management), which had to be answered on a 10-point response scale ranging from 1 (very low) to 10 (very high value). Furthermore, a second expert rated ten videotaped lessons, chosen at random. Interrater reliability (Wirtz & Caspar, 2002) was satisfying for all 18 items (Spearman’s rho: .50 up to .76, p < .05). The cooperating teachers filled in the same rating form directly after the lesson was taught by the student teacher.

Analysis of variance based on the overall rating score of the expert (Cronbach’s α =.85) reveals a significant difference between the intervention and the control group (F(1,31) = 7.533, p < .05; MIG = 8.13; MCG = 7.12). Student teachers in the experimental group received a higher average mean score in quality of teaching than did their colleagues in the control group. At the level of individual rating items the values of five items differed significantly (Mann-Whitney U-Tests; (p <.05; MIG > MCG): interestingness, positive error culture, motivation and encouragement, use of structuring means and providing explanatory connections. The rating by the cooperating teachers (Cronbach’s α = .63) displayed no significant difference between the intervention and the control group, except for the items of articulateness (p < .05; MIG > MCG) and the use of content concepts (p < .05; MIG > MCG).