Learning analytics in support of qualitative analysis

The purpose of this manuscript is to describe a style of learning analytics research that makes closer contact with the qualitative methods of the learning sciences. We do so by drawing on two examples of existing research. The first is our prior work attempting to automate part of the Classroom Video Analysis assessment, which seeks to measure knowledge of teachers that is predictive of their performance. The second is from work which looks at how middle school students explain the Earth’s seasons. In that work, we attempted to use unsupervised methods to capture elements of what was previously a fully qualitative analysis. Our goal is to provide the reader with a sense for this style of research that brings qualitative analysis and analytic methods into closer contact. To accomplish this, we make use of a new system called Tactic, designed specifically to support this mode of research.