Theories All the Way Across: The Role of Theory in Learning Analytics and the Case for Unified Methods

We argue that learning analytics research—and really any empirical research—is composed of a primary modeling pathway: a series of steps that progressively reduce the extraneous information that confronts a researcher and thus clarifies important claims about, issues in, and properties of some activity in the world. Each of these steps is necessarily informed by some theoretical perspective (implicit or explicit) and is therefore contestable. Because none of these analytical steps can be validated from within, research is stronger when qualitative and quantitative methods are unified in the research process.