In order to improve the course design of a CS1 model-driven programming course we study potential indicators of success for such a course. We explain our specific interpretation of objects-first. Of eight potential indicators of success, we have found only two to be significant at a 95% confidence interval: math grade from high school and course work. The two significant indicators explain 24.2% of the variation of the exam grade. The result concerning math grade contradicts earlier findings. We discuss four aspects of our research: the explanation power of the potential success indicators, the impact of our findings on teaching, limits of what to conclude from the available data, and the variety of the notion "objects-first". Because of the variety of interpretations of "objects-first", the present research is necessary as a supplement to earlier research in order to make generalizable results on the success factors for objects-first programming.
2005 International Workshop on Computing Education Research, Icer '05 Proceedings (seattle, Wa, Usa, October 1-2, 2005), 2005, p. 155-163
International workshop on Computing education research, 2005