Statistical Models for Analyzing Learning Data
If the “Choosing a Statistical Test” tables feel too abstract or you prefer a narrative entry point into quantitative analysis, this is the resource for you. We find it particularly valuable because it was written specifically in the context of SoTL, making the examples relevant and accessible.
Oftentimes we find ourselves with a plethora of data on teaching and learning but lack the direction needed to find the best way to analyze these data. We collect and often grade many different assessments over the course of a semester, offering numerous and diverse outcomes. If we want to truly understand, utilize, and disseminate what we learn from these assessments, analyzing the data using an appropriate and rigorous method will help us do just that. Knowing that what we do in the classroom is having an impact on students’ learning above what might happen just by chance alone is an important motivator for using statistical analyses (Gurung and Schwartz 2009). Selecting the optimal analysis is often the most difficult part in the research process. The previous chapters outlined some design models to measure student success. The current chapter will build on that foundation, providing potential models for analyzing the learning data that have been collected. It should be said at the outset of this chapter that statistical analysis is only as good as the data collected. Using appropriate and rigorous data collection methods provides results with the most integrity (Wilson-Doenges and Gurung 2013).