When measuring the effect of a certain teaching method, it is necessary to make experimental group and control group which is controlled with factors other than the teaching method , and to investigate by Pre-Post design. However, at the educational situation, it is often difficult to make sufficient control, and there remains a possibility that false educational effects can be detected at the sake of reverse causality or pseudo correlation. So, in this paper, I will introduce statistical ways which can remove the influence of factors other than the teaching method. One way is to use control variables in multiple linear regression analysis as the third variable, and another way is to make triple cross tabulation with qualitative control variables. Using these ways, we can grasp the effect of a teaching method more precisely and useful knowledge will be accumulated.
View full abstract