Mapping Socioeconomic Variation in Children’s Digital Programming Activity: From CodeMonkey Logs to Behavioral Profiles

How do different forms of digital and diagnostic evidence reveal socioeconomic differences in students’ computational thinking? This presentation examines this question through a study of 3rd- and 4th-grade students learning computational thinking in CodeMonkey, a digital programming environment. The study combines a pen-and-paper diagnostic CT assessment with log data, including first-try accuracy, attempts per challenge, progression, and solution time. Rather than treating these traces only as separate indicators, the analysis also identifies digital behavioral types through clustering. The findings show no meaningful socioeconomic differences in diagnostic CT scores and no consistent differences across isolated log-based indicators, except first-try accuracy. However, socioeconomic differences become visible when students’ log-based indicators are examined as multidimensional behavioral profiles. The presentation contributes to digital social-science research by showing how computational traces can reveal patterns of engagement that remain hidden in outcome-oriented diagnostic assessment, while requiring careful interpretation as behavioral, not direct conceptual, evidence.