This study investigates Turkish students' knowledge structure coherence in physics. In particular, this study investigates the conflicting findings reported in loannides and Vosniadou's [loannides and Vosniadou [2002] Cognitive Science Quarterly, 2, 5-61] and diSessa, Gillespie, and Esterly's [diSessa et al. [2004] Cognitive Science, 28, 843-900] studies about students' understandings of force. loannides and Vosniadou's study of four different age levels of students in Greece demonstrated broad consistency in students' understandings of force. diSessa and colleagues' quasi-replication in the United States demonstrated conflicting results supporting a more fragmented elemental perspective on students' knowledge structure coherence. The current study investigates these conflicting findings by studying students in a third country using the analytic methods from both studies to clarify the debate over knowledge structure coherence. The levels of consistency demonstrated by students in the Current study are somewhat higher than the levels reported by diSessa. Gillespie, and Esterly according to both coding schemes, but are closer overall to the levels reported by diSessa, Gillespie, and Esterly than to the levels reported by loannides and Vosniadou. In addition, closer inspection of students' explanations suggests that students' explanations may code as consistent according to the coding schemes for a particular force meaning category but not actually represent a coherent understanding of that force meaning. These results therefore more closely support fragmented elemental perspectives on knowledge structure coherence. The results, however, demonstrate important systematicities in students' thinking and support the possibility that differences between the student populations in the countries of the original studies contributed to the differences in findings of the original studies. (C) 2009 Wiley Periodicals. Inc. J Res Sci Teach 46: 570-596, 2009
