Motivational and cognitive challenges faced when debugging block-based code

Brian Belland
Associate Professor
The Pennsylvania State University

NeedIf computer science is to be for all, it must also be for the youngest learners. Integrating computer science content in early childhood education (ECE) can help early learners gain problem solving and executive function skills, as well as prepare them to enter computer science pathways. To do so, preservice, early childhood teachers need to learn to debug block-based coding.Guiding questionsWhat are the cognitive and motivational challenges preservice ECE teachers face as they learn to debug block-based programming?How can scaffolding be designed and implemented to help preservice ECE teachers overcome these challenges?OutcomesWe used Bayesian regression, discriminant analysis, and generalized estimating equations (GEE) approaches to investigate cognitive and motivational challenges to debugging among preservice ECE teachers. Bayesian regression indicated that debugging process quality is a significant positive predictor of the quality of debugging solution for both low- and high-complexity debugging tasks, performance-avoid goal orientation is a significant negative predictor of low-complexity debugging outcome quality, and performance-approach goal orientation is a significant positive predictor of high complexity debugging outcome quality. Discriminant analysis indicated that sentiment analysis and word counts of debugging and field experienced teaching reflections were the strongest classifiers of lesson plan quality. The first GEE indicated that achievement emotions in STEM was a positive predictor and mathematics interest was a negative predictor of overall lesson design quality score. The second GEE indicated that our instruction on debugging led to a significant increase in views of coding. Furthermore, we found that designing a lesson and using it in field experience was not a significant predictor of views of coding.We used a clustering approach on a dataset originating from a meta-analysis of scaffolding (Authors, 2017) to investigate the optimal scaffolding approaches within computer science and teacher education. Only studies in computer science and teacher education contexts were considered. The three most important predictors of medium and large effect size were the context in which scaffolding was used, if and how scaffolding is customized over time, and the decision rules that govern scaffolding change.Broader impactsOur results have implications for broadening participation in CS among women and members of other underrepresented populations. First, the vast majority of ECE teachers and preservice teachers are women. In our studies along, 98% of participants were women, which is in line with national averages. Many of our results may be useful in broadening participation among women. In particular, they help to identify the most critical cognitive and motivational challenges faced by preservice ECE teachers as they learn to debug. Furthermore, once the preservice teachers impacted by the project enter the field, they can serve as positive role models to early learners, who will be able to see women who are engaging in computer science. This in turn may serve to reverse longstanding stereotypes of computer science as being for white men. This may also help children of color see a place in computer science for themselves, which is so important.


Anna Y. Zhang, The Pennsylvania State University, University Park, PA; Eunseo Lee, The Pennsylvania State University, University Park, PA; Emre Dinc, The Pennsylvania State University, University Park, PA; Afaf Baabdullah, The Pennsylvania State University, University Park, PA; ChanMin Kim, The Pennsylvania State University, University Park, PA