To prevent the pitfalls reported by Jakobsson et al. (2009) for other studies on misconceptions, students were given unlimited time to answer and the questionnaire was written in everyday language, avoiding specific scientific terms. Also in contrast to some practices reported by these authors, answers were not categorized strictly according to their exact wording (where or not they were strictly scientifically acceptable or not), but by taking each student’s full set of answers into account, in the very few cases that the answer contained any ambiguity. Indeed, the categorization of answers to the open-ended questions was made easy by the fact that already from the first semester of our study on, most of the answers expressed a limited set of clear prior ideas. We were thus also able to prevent any false mental-model inconsistency due to misinterpretations of the language used in the questions regarding different contexts or scenarios, as described by Alonzo and Steedle (2009).