Predictors of well-structured and ill-structured problem solving in an astronomy simulation

This study compared the problem-solving skills required for solving well-structured problems and ill-structured problems in the context of an open-ended, multimedia problem-solving environment in astronomy. Two sets of open-ended questions assessed students' abilities for solving well-structured and ill-structured problems. Generalized, rubric scoring systems were developed for assessing problem-solving skills. Instruments were also developed and administered to assess cognitive and affective predictors of problem-solving performance. By regressing the scores on the cognitive and affective predictors onto students' scores on the well-structured and ill-structured problems, we concluded that solving well-structured and ill-structured problems require different component skills. Domain knowledge and justification skills were significant predictors of well-structured problem-solving scores, whereas ill-structured problem-solving scores were significantly predicted by domain knowledge, justification skills, science attitudes, and regulation of cognition. Implications for problem solving in science education are presented.