The effect of curriculum change on nutrition behavior, problem with regression function

I’m looking into the FICTIONAL effect of a change in the national primary school curriculum on long-run nutrition behavior. This change caused nutrition education to get mandatory. I believe that, through this nutrition education, knowledge on nutrition improved. With the use of a difference-in-difference specification I’ll be testing for a causal link between nutrition education and long run nutrition behavior.
I, however, get stuck at the regression function. I’ve got no idea what variables to include. This is how it looks like until now:
Nutrition behavior t = Alfa i + B1 x Curriculum i + B2 x Parental income i x age i + gamma t + e it
My question: are the current variables correct (what issues will these variables bring me) and what variables should I include furthermore?


Active Member
In a way, in a sense, "fictional" data analysis is a contradiction in terms. The correct data analysis begins with the questions: "Which data are available?" "Which variables are available? "Can I collect even more variables by administering a controlled experiment?"... Unfortunately, if we are imagining things we can imagine whatever we like.

Your specific question can be partially addressed by comparing countries with different educational systems and extrapolating their differences into the desired universe.