Cognitive load comparison of traditional and distributed pair programming on visual programming language

Research results from computer science education show that pair programming is an effective teaching strategy on learning performance for computer science education in K-12. However, pair programming is not popularly used in K-12. One important issue is because sometimes pair programming by K-12 students is not as effective as expected. This study examines whether pair programming is effective from cognitive load perspective in distributed situation. We compares the cognitive loads for the visual programming language, called StarLogo TNG, under three circumstances that is alone, pair programming, and distributed pair programming. The experimental results show that learners prefer traditional pair programming to alone. Moreover, the pair programming can reduce germane cognitive load significantly. However, there is no significant difference between the single and distributed pair programming. Meanwhile, there is no significant difference on learning performance between traditional and distributed pair programming. Those results indicate the learning support and knowledge sharing tools for distributed pair programming could be improved.