Chapter 6

Handbook of Learning Analytics
First Edition

Going Beyond Better Data Prediction to
Create Explanatory Models of Educational Data

Ran Liu & Kenneth R. Koedinger


Abstract

In the statistical modelling of educational data, approaches vary depending on whether the goal is to build a predictive or an explanatory model. Predictive models aim to find a combination of features that best predict outcomes; they are typically assessed by their accuracy in predicting held-out data. Explanatory models seek to identify interpretable causal relationships between constructs that can be either observed or inferred from the data. The vast majority of educational data mining research has focused on achieving predictive accuracy, but we argue that the field could benefit from more focus on developing explanatory models. We review examples of educational data mining efforts that have produced explanatory models and led to improvements to learning outcomes and/or learning theory. We also summarize some of the common characteristics of explanatory models, such as having parameters that map to interpretable constructs, having fewer parameters overall, and involving human input early in the model development process.

Export Citation: Plain Text (APA)     BIBTeX     RIS

Supplementary Material
References (33)
About this Chapter
Founding Members
Previous Image
Next Image

info heading

info content