Analysis of learning behavior in a flipped programing classroom adopting problem-solving strategies

Programing is difficult for beginners because they need to learn the new language of computers. Developing software, especially complex software, is bound to result in problems, frustration, and the need to think in new ways. Identifying the learning behavior behind programing by way of empirical studies can help beginners learn more easily. In this study, a flipped classroom combined with the blended learning mode was adopted by switching the in-class instructional time and the out-of-class edX lecture time. The research laboratory developed a system that collected the teacher and students discussion activities when adopting problem-solving strategies in online discussion. The participants were 44 third-year college students from the computer science course of a university in northern Taiwan. The experiment was from 30th March to 4th May in 2015. A total of 120 initial comments were collected, along with 401 responding comments. These 521 comments were coded according to the problem-solving steps as posing questions, offering solutions, discussion, and sharing conclusions. The inter-rater reliability of the two coders reached a kappa value of 0.81. The results indicated that the flipped classroom combined with the problem-solving strategy was more effective than the previous problem-solving behavior sequence.