Investigating expertise related gaze during code debugging and experts’ didactical gaze

With the growing demand for programming expertise on the labor market (Prabhakar, Litecky, & Arnett, 2005) the investigation and optimization of programming education become increasingly relevant. Modern eye-tracking technology has proven to be a fruitful tool for studying expertise development (Reingold, & Sheridan, 2011). The present study aims to contribute to expertise research by comparing eye-movement patterns of programming experts and novices. Furthermore, the study explores how experts change their gaze behavior when explaining a task didactically. Superimposing eye movements of a didactically behaving expert onto instructional videos (EMMEs) is assumed to foster learning (e.g., Jarodzka, Van Gog, Dorr, Scheiter, & Gerjets, 2013). The results of the investigation of didactic gaze patterns could help to answer the question how displaying didactic eye movements helps guiding a learner’s attention. We will conduct a study in spring 2018 with programming experts (N = 20) and novices (N = 20). Experts will first debug short program code snippets while their eye movements are recorded. Subsequently, they will receive the instruction to behave didactically. The group of novices will only debug the program code regularly. Based on previous expertise research it is hypothesized that in comparison to novices, non-didactically behaving experts run the code less often, show shorter average fixation durations in the code area, perform larger saccades in the code area, switch less between areas of interest, and read code more linearly. Furthermore, the eye movements of didactically and non-didactically behaving experts will be analyzed exploratively to answer the question how experts try to make themselves more understandable for novices.