Qiwei Britt He

Abstract

This study introduces a generalized method for profiling individual respondents’ process data as unique images, enabling pattern extraction through image processing techniques. By applying pairwise comparisons of these images, we identify homogeneous behavioral clusters that reflect distinct problem solving strategies. A case study using process data from problem-solving tasks in international large-scale assessment PIAAC will be presented as an illustration. The method can also be extended to multidimensional image processing approaches by incorporating additional layers beyond action and time sequences, allowing multiple dimensions to represent interaction patterns and degrees of engagement.