[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: gllamm vs. xtmixed vs. a2reg
I measured university-level obstacles to patenting activity (e.g. absence of a TTO, perceived bureaucracy of the university administrators, excessive teaching loadings, etc...) by asking to the faculty. Please note that no survey design exists: I targeted the whole faculty population, but of course there have been non-respondents.
1) Can I predict the university-level patent counts from the faculty's opinions (and obviously from some university-level characteristics, e.g. size)?
2) If yes, is there any specific Stata command to run?
3) If no command exists, which of the following estimation strategies do you prefer?
a) producing university-level averaged opinions and regressing university patent counts on them
b) regressing university patent counts on individual opinions and clustering on individuals
c) regressing university patent counts on individual opinions and weighting observations with the inverse of the number of the respondents for that university
d) regressing university patent counts on individual opinions and both clustering and weighting
Many thanks for your help,
* For searches and help try: