A difficulty component model for developing computerized problem-solving assessments

Problem solving has become a central educational objective in developing many educational systems in recent years. The competency of problem solving includes many cognitive operations that are also important to future learning, effectively encountering and overcoming obstacles, and even adapting to new conditions, communities, or societies. In this study, a six-component model of problem-solving assessment is proposed; moreover, a computerized problem-solving assessment was developed according to the specifications of the content and cognitive structure. To verify the reliability of the assessment, 1724 7th and 8th graders in Taiwan participated in the test. The test results suggested that the proposed six-component model accounted for around 67 % of the variance in item difficulty. The component of disclosure of information revealed a relatively stronger impact on item difficulty. On the other hand, simulations could help students to solve the problems more easily. The effects of grade and gender were both statistically significant. Based upon the cognitive component model, the performances of students were classified into four levels, that is, below basic, basic, proficient, and advanced. In further studies, it is expected that the proposed model can provide useful guidelines for integrating problem-solving competency into regular classroom instruction and for improving students’ performance.