Collaborative learning is an activity in which two or more students learn something together. Many studies have found that collaborative learning improve students' memory retention and motivation to learn. Peer Instruction (PI) is one of the most successful evidence-based collaborative learning methods. This article investigates issues of student motivation in courses on data structures through a one-on-one PI collaborative learning activity. Students typically find the concepts of data structure vague and the related programs complex and difficult to comprehend. Pair programming was conducted to scaffold university students understanding data structures by commenting related programs. To investigate students' motivational status, the ARCS model was used to evaluate students' motivation in each of four dimensions: Attention, Relevance, Confidence, and Satisfaction. Moreover, three items about the retention of learning were added to measure how information is lost over time. Every "pair" (i.e., two students) was asked to comment on the programs provided by a specific data structures concept. The control groups were asked to elaborate on the data structures programs by adding algorithmic description comments to the beginning of programs. The experimental groups were asked to elaborate on the data structures programs by adding code description comments to the data structures programs line by line. The experimental results show that the confidence and retention of learning in experimental groups are significantly better than those of the control group. The findings suggest that the effects of pair programming activities can be leveraged based on an understanding of students' motivation for and retention of learning.