Effect of disciplinary content of a simulation on open-ended problem-solving strategies

There is a long-standing debate about so-called generic vs. domain- or context-specific problem-solving skills and strategies. To investigate this issue, we developed a pair of structurally and functionally similar computer simulations and used them in an experimental study with undergraduate students from all fields of study. One, described to users as a teaching tool based on current entomology research, featured three species of ants whose tasks could change as a result of encounters with each other. The other, which we call a non-disciplinary simulation, featured three types of abstract moving shapes whose color could change after a collision. Users received no information about the latter, not even that it was a simulation. Both simulations, which students used in succession, allowed users to change the number and types of objects, and to move them freely on the screen. We told students that the problem was to describe and explain what occurred on the screen, and asked them to explain, while they were working with the simulation, what they observed, what they did and why. We conducted a first, quantitative analysis of various 'surface' indicators that can be related to patterns or strategies: computer-generated traces of significant events, including user interaction; users' operations (meaningful sequences of actions); basic characteristics and factual content of students' verbalizations. We found significant differences between the two simulations for a number of indicators, which are being used as guides in the next part of the study: an in-depth analysis of students' strategies