Confidence vs Insight: Big and Rich Data in Computing Education Research

There are now many large datasets available for programming education research. They tend to be very large-scale, but often lack context or detailed participant information. This ''big data'' is in contrast to the ''rich data'' that has generally been collected from smaller, qualitative studies, with detailed context and participant information. Big data is often criticised for its lack of context, while rich data is often criticised for its small sample size which makes generalisable conclusions dubious. In this position paper we examine the constraints, advantages, and disadvantages of each type of data, and discuss how they can provide differing information on phenomena in programming education research. We argue that both types of data are useful and that we should value the potential findings of each, as well as encourage their combination in order to provide a complete picture of how people learn to program.