Breakout Session I – Issues related to quantitative methods and data analysis in PER
Location: Hatch Ballroom C&D
David E. Meltzer, Iowa State University and
Richard R. Hake, Indiana University
A number of issues are always present, implicitly or explicitly, when using quantitative methods to carry out PER. A few of these include: identifying targeted outcome variables as well as possible confounding variables, distinguishing among the criteria measured by different diagnostic instruments, validity and reliability of diagnostic instruments, possible biases in sample selection, statistical significance vs. "practical" significance of observed effects, criteria for assessing magnitude of effect (e.g., normalized gain, effect size, etc.). We will present a short introduction based on our experiences in confronting these issues. We will then invite participants to provide written questions based on issues that have arisen in their own experiences, and these will be presented for open discussion among all participants.