Although misconceptions in science have been established in interview studies, their role during the learning process is poorly examined. In this paper, we use results from a classroom study to analyze to what extent nonscientific ideas in electrochemistry that students report in interviews enter into their learning in a more authentic setting. We audio-recorded talk between eight pairs of Swedish upper secondary students during a practical on electrochemical cells. Learning was operationalized on a discursive level as a description of what students do and say when taking part in an activity. This enabled an analysis of how encounters with misconceptions influenced the development of students' reasoning, compared to other encounters during the learning experience. Misconceptions did not constrain the development of students' reasoning. Rather, their reasoning developed in response to the contingencies of the specific situation. When misconceptions were encountered, they appeared as alternatives and questions not actively defended. Sometimes, encounters with these misconceptions were generative of the students' reasoning. The results indicate that demonstrating misconceptions in interviews is not enough to assume that they interfere with learning in other contexts. Educational implications and future lines of research based on these findings and on the methodology applied are discussed. © 2007 Wiley Periodicals, Inc. Sci Ed92:141–164, 2008