If you are talking about doing empirical pedagogical research, your first step should be thinking about what sort of research project you can do in a feasible manner. For example, if you are a professor teaching a multi-section course, you can set up studies with students in 4 sections assigned to use one learning strategy and 4 sections assigned a different one and compare results. You need to be able to create two groups containing at least 40-50 people per group (more is better) to have useful statistical results.
If you do not have enough study subjects for statistical analysis (a basic rule of thumb is that N=50 is the minimum for significance), you will need to to some form of ethnography, watching and closely describing an individual learner or pedagogical method.
You may be able to do historical research as well, doing something like looking through archives of syllabi for a given course and analyzing how they have changed over a 10 or 20 year period.