Mixed Methods: Positivists are from Mars, Constructivists are from Venus

摘要

You can't work in Computer Science Education Research for very long without stumbling into a religious war over qualitative versus quantitative methods. Recently, I read a general education paper where the authors were brave (or foolish) enough to advocate "mixed methods". That is, they advocate the use of both quantitative and qualitative methods within a single study. When I started in CSEd, I wanted to do "good old fashioned" positivist, quantitative research. I thought naively that education research should be like physics, or medical research. I suspect this is how most of us start in CSEd research. I now prefer mixed methods. I am drawn to studies which combine "hard numbers" with qualitative data that "makes sense" of the numbers. For example, I was part of a project that studied novice programmers. We collected extensive quantitative data (statistics on "performance" and "doodles"). However, I could not have made any satisfying interpretation of those numbers had we not also collected qualitative ("think out loud") data from the novices.