Behavioral patterns associated with solving ill-defined complex problems from a multidimensional perspective: Perception, cognition, metacognition, and motivation

Abstract

Students in the 21st century are expected to possess the ability to solve ill-defined complex problems (ICPs). One challenge to understanding students’ ability to solve ICPs is the lack of methods for measuring noncognitive and metacognitive behaviors and relating those behaviors to cognitive behaviors with the goal of investigating differences in student performance across ability levels. Based on the principles of the synthetic intelligence (PSI) framework, this study utilized a computerized interactive assessment platform to design a multidimensional evaluation framework (including the four dimensions of perception, cognition, metacognition, and motivation) and analyzed log file data collected from 132 elementary students with regard to solving ICPs. The results revealed new problem-solving strategies among students in the high-achievement group, who spent more time constructing problem models. Due to their ability to exercise goal-oriented self-control, students in the high-achievement group were able to fully explore the information they needed to optimize their solutions. The results also revealed three types of behaviors that characterized differences in motivation, the most notable of which characterized students who succeeded after relentless attempts. This study also explains the interaction mechanism underlying mental processes based on the PSI framework. The findings suggested that educators can highlight differences between environmental stimuli and students’ internal assumptions, encourage students to adopt strategies that disambiguate the task goal and object, and strengthen their ability to search for relevant information to improve their performance in solving ICPs. The results also provide a new paradigm for assessing problem-solving capabilities based on the PSI framework.

Keywords

• 21st century skills/thinking skills
• behaviour pattern extraction and analysis
• educational data mining
• interdisciplinary projects