Beyond Strategy Use: Examining Strategy Stability in Complex Problem Solving Through Logfile Data

Complex problem solving (CPS) is a key educational outcome in technology-rich environments. Previous logfile studies have demonstrated that effective exploration strategies are associated with higher CPS performance. However, little is known about whether students can sustain effective strategy use across successive problem-solving tasks. This issue is important because strategy stability may represent a distinct dimension of CPS skill that is not captured by static indicators of strategy use alone. This study examined students’ Vary-One-Thing-At-A-Time (VOTAT) strategy use across eight computer-based MicroDYN scenarios. A sample of 197 secondary school students from a through-train school in Macao completed the assessment. Exploration behaviours were coded as no VOTAT, partial VOTAT, or full VOTAT. Latent class analysis identified strategy classes, while transition-probability analysis examined the stability and change in strategy use across adjacent tasks. Three strategy classes emerged: proficient strategy users (65.5%), low strategy users (20.3%), and intermediate strategy users (14.2%). Proficient strategy users achieved the highest CPS performance and showed high stability in full VOTAT use across tasks. Low strategy users rarely transitioned to full VOTAT, whereas intermediate strategy users had difficulty sustaining full VOTAT once they adopted it. Educational-stage differences in class membership were not significant, although senior students outperformed junior students in knowledge application. The findings indicate that students differ in both their use of effective strategies and their ability to sustain effective strategy use across problem-solving situations. They also highlight the value of logfile-based transition analyses for understanding students’ problem-solving processes.