Bell, Lederman, and Abd-El-Khalick's (1998) major concern is that we based our main conclusion on an assumption rather than on what they felt was "substantial direct instruction" (Bell et al., 1998(Bell et al., , p. 1058)). In fact, there was little direct instruction about the nature of science. During two quarters of science teaching methods coursework, there was ฯฝ 3 h of direct instruction on the nature of science. The first methods course devoted the first 2-h class period to an explicit discussion of the nature of science (Palmquist & Finley, 1997, p. 597). In the second methods course, the nature of science was mentioned four times throughout the quarter (p. 598). According to notes from observations of the course, the discussions that went along with the "mentioning" lasted ฯฝ 1 h total. The preservice teachers were not assessed during either course on their knowledge of the nature of science. Thus, the direct instruction on the nature of science lasted ฯฝ 3 h out of 50 h of classroom instruction and a 40 h practicum in the public schools. Furthermore, the majority of this instruction came on the first day of the first course, 6 months before the final interviews. We do not believe that spending that little time on a topic, which was not assessed, can be considered substantial. In addition, when the methods instructors were asked which aspects of the nature of science came across explicitly in their teaching, they both emphasized that learning is driven by prior knowledge and learning is influenced by the social context in which it occurs. Both stressed that students learn science in much the same way as scientists do. Table 3 in Palmquist and Finley was an attempt to match those and similar notions to the contemporary model of the nature of science.
We based our conclusions on open-ended surveys, interviews about the surveys, interviews about teaching, observations of teaching, and an analysis of the lesson plans. Bell et al. were concerned about the method of our teaching analysis and the possible contamination of the analysis by our prior knowledge of the preservice teachers views of the nature of science before observing their teaching. All data analysis was done after the postmethods interview, significantly reducing observer bias. Once we started the analysis, every piece of data was analyzed in the same manner. It was read and compared to the models of the nature of science (p. 600).