Investigation of the effects of hybrid programming platforms on programming performance and cognitive load theory

Programming education includes difficult and sophisticated progress due its nature. The approaches of programming education have been evolving in a continuous manner. This study aims to focus on the effects of hybrid programming platforms, which can be considered as one of the recent approaches, on both programming performance and cognitive load of learners. 49 sixth graders participated to the study. The intact groups were assigned to one of the conditions, which are hybrid-hybrid, hybrid-block, and hybrid-text. The study took place during 8 weeks. The groups were intact, but their registration to the school is based on their elementary school graduation scores and gender, therefore each group can be assumed homogenous. All groups were exposed the same tasks in hybrid platform during 4 weeks (Phase I), but one group was restricted to work with block-based environment, one group was restricted with text-based environment, and one group continued with the hybrid environment. The tasks were completed in Pencilcode.net platform offering a hybrid environment in addition to block-based and text-based options. Since this platform had all-in-one, it was preferred by researchers, so that the bias caused by technical issues were eliminated. Each task was assigned separately, and the screenshots were recorded digitally at the end of each week. The screenshots and cognitive load scales filled by students were analyzed. Moreover, the fieldnotes and interview were analyzed qualitatively. Although the results of ANCOVA showed a significant difference among Phase II performance scores when Phase I scores were controlled, the post-hoc results did not refer significant differences between groups. The results of ANCOVA for cognitive load scale showed significant differences among Phase II cognitive load scores when Phase I scores were controlled. The post-hoc comparisons indicated higher cognitive load experienced by either block-based groups than hybrid group.