This Small Business Technology Transfer Phase 1 research project will develop and implement an on-line performance-based assessment system for quantifying scientific problem solving effectiveness where quantitative measures can be normalized across problem solving tasks allowing longitudinal comparisons to be made across individuals, classes, schools and science domains. The research project will derive a metric that combine estimates of the quality of a strategy that are derived from artificial neural network analysis with the strategic outcomes on problem solving tasks to develop a value-based metric of the problem solving process. This metric will also provide a strong measure of the value of the strategy employed to document the validity, utility, and generality of this assessment measure using an existing dataset of over 200,000 problem solving performances that span grade levels from middle school through the university.
Such an extensible formative, summative and programmatic assessment system of learning will have broad relevance for helping teachers to teach, students to learn, and administrators to make informed data-driven decisions through the continual, and real-time formative evaluation of students' problem solving progress, a dimension not frequently or rigorously assessed in today's classrooms, yet a critical component of 21st century skills. This system will impact all levels of science education and would allow cumulative comparisons of problem solving across science domains, classrooms, teachers and school systems thus helping to re-think the ways scientific problem solving is systemically assessed and how the impact of teaching these skills becomes quantified. The development of a commercialization pathway for this assessment tool will be facilitated by the involvement of the American Chemical Society Examinations Institute, the Higher Education Press in Beijing, and the Alternative Education Program of a local school district.
