When beginning your research, here are the questions you need to ask yourself [from Mostly Harmless Econometrics]:
- What is the causal relationship of interest? What specific mechanism will cause a change in the dependent variable of interest? Often one uses economic theory to predict these causal relationships.
- What experiment could be used to capture the causal effect of interest? Before you can decide on an identification technique, one must figure out what the ideal experiment would be. If you want to estimate the effect of physician payment on surgery rates, would you randomize patients to different physicians? Different physicians may select into different payment schemes. Would you randomize physician payment? In this case, different types of patients may select different doctors. What would be the ideal?
- What is your identification strategy? Many medical studies use randomized control trials, but there are very few RCTs investigating economic phenomenon. A researcher must decide how to eliminate problems of selection and endogeneity. Common strategies include OLS, difference-in-difference, instrumental variables, and others.
- What is your mode of statistical inference? This is the nitty-gritty stuff. How will you estimate your standard errors? What variables do you include in your regression? Is the sample representative? What is the correct group to study. In my paper on Marriage and Weight Gain, I limit the sample to individuals aged 18-55 since these are individuals most likely to be in the dating market.