An Experimental Study on the Association Between Affective States and Novice Programmers’ Performance

Problem-solving in programming requires not only cognitive but also affective engagement from students. Despite many novice programmers struggling with foundational programming concepts and problem-solving using these concepts, few studies explore how students’ emotional states impact their learning in introductory programming. This article investigates the association among affective states, programming concepts, and performance in novice programmers from a rural Brazilian school. Specifically, we explore how different affective states relate to success rates across core programming concepts. Unlike prior research often relying on IDE logs or broad academic indicators, this study employs a sensor-free, self-reported approach to assess students’ emotions in real time during programming tasks. Utilizing causal inference methods, we analyze both the direct and indirect effects of affective states on programming performance. Through two exploratory studies, our findings reveal that positive emotions such as enjoyment, motivation, and engagement are significantly associated with higher success rates, while negative states, like boredom, anxiety, and frustration, correlate with lower outcomes. These results enhance the understanding of the relationship among affective states, performance, and concepts in introductory programming tasks. Furthermore, this study offers valuable insights for improving educational practices and tools in programming education by emphasizing the critical role of addressing emotional dimensions in teaching and learning.