How to reliably diagnose children’s concepts in learning science? Using the water cycle as an example

The text addresses the question of how methodology in research on children’s conceptions about the world possibly affects results. This methodological question has received little attention yet in educational and psychological literature. As part of a preliminary study focusing on the effectiveness of internally differentiated factual texts using the water cycle as an example, the authors examine(d) children’s conceptions of it. The data set includes 121 pre- and post-tests of nine- to ten-year-old children. Results show that free recall tasks tend to underestimate children’s performance, whereas cued recognition tasks tend to overestimate children’s performance. The findings demonstrate that it is worthwhile and important–in terms of reliability and validity of the data–to check the survey methods for their potential biases on the results and to plan for method triangulation from the beginning when surveying preconcepts.