I was asked to redesign a microeconometrics course for next academic year. It's aimed at advanced undergrads.
I was thinking along these lines:
intro, OLS, LDV
diff-in-diff, regression discontinuity
IV
panel data (is it too much to include dynamic panels?)
What else would you suggest? For example does anyone need Oaxaca decomp, Manski's bounds to throw in as minor topics or practicals.
Yes, the plan was to do an extra session on rainfall and menstrual cups etc and at least 3 session PLUS practicals on robust SEs. Robust SEs are robust man.
What about Oaxaca, manski and matching?? I haven't seen anyone do that s**t in any recent papers. Should I drop them?
Oh yeah and is it too much trying to do both Stata and R in the practicals?
I am going to hate this teaching already...
f**k it. Do something modern, like networks and
bozones.
I was asked to redesign a microeconometrics course for next academic year. It's aimed at advanced undergrads.
I was thinking along these lines:
intro, OLS, LDV
diff-in-diff, regression discontinuity
IV
panel data (is it too much to include dynamic panels?)
What else would you suggest? For example does anyone need Oaxaca decomp, Manski's bounds to throw in as minor topics or practicals.
If they've taken basic econometrics + stata, why bother with OLS etc at the start?
Depends how HRM you are, but I'd say the likes of dynamic panels or non-parametrics are too much. Better to send them out into the world solid on bread and butter stuff like fixed effects.