AUTOMATIC ASSESSMENT OF CREATIVITY IN HEURISTIC PROBLEM-SOLVING BASED ON QUERY DIVERSITY

Abstract

Creative problem-solving emerges as one of the most relevant skill of the 21st century knowledge society. Fortunately, there are many creativity training programmes that have proven effective. However, most of these programmes require a previous measurement of creativity, which involves time-consuming tasks conducted by experienced reviewers, i.e. far from primary school classroom dynamics. In this study, we propose a model to predict the creative quality of students’ solutions based on the analysis of query patterns and the use of Wikipedia. This model has been able to predict the creative quality of solutions produced by 226 school students, aged 10 to 12 years old, reaching a sensitivity of 78.43%. The agreement among reviewers regarding students’ creative characteristics has also been evaluated using two rubrics. We hope this model can be used to foster prompt detection of non-creative solutions in order to enable intervention and improve the final result in terms of creativity.