* Stata Do-File
* Tana Johnson and Johannes Urpelainen (2014)
* International Bureaucrats and the Formation of Intergovernmental Organizations: Institutional Design Discretion Sweetens the Pot
* International Organization 68 (1): 177-209
* Published online 24 January 2014


* key variables individually summarized, then correlated and cross-tabbed
* note: chi-squared test assumes no expected frequencies < 5, so also ran Fisher's exact test (which requires no such assumption)
summ IBdiscretion
summ expertise3catforum
corr expertise3catforum IBdiscretion
tab expertise3catforum IBdiscretion, chi2 exact

* difference in means, prior to matching, for treatment version "treated2only" 
* (more conservative)
gen treated2only = 1 if IBdiscretion == 2
replace treated2only = 0 if IBdiscretion == 0 | IBdiscretion == 1
ttest expertise3catforum, by(treated2only)

* difference in means, prior to matching, for treatment version "treated1or2"
gen treated1or2 = 1 if IBdiscretion == 1 | IBdiscretion == 2
replace treated1or2 = 0 if IBdiscretion == 0
ttest expertise3catforum, by(treated1or2)

* difference in means, with coarsened exact matching, for treatment version "treated2only"
* did not impose one-to-one (k2k) matching, in order to preserve as many observations as possible 
* (more conservative)
imb yrcreated issuearea, treatment(treated2only)
cem yrcreated issuearea, tr(treated2only)
ttest expertise3catforum if cem_matched==1, by(treated2only)

* difference in means, with coarsened exact matching, for treatment version "treated1or2"
* did not impose one-to-one (k2k) matching, in order to preserve as many observations as possible 
imb yrcreated issuearea, treatment(treated1or2)
cem yrcreated issuearea, tr(treated1or2)
ttest expertise3catforum if cem_matched==1, by(treated1or2)