|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Do I use survey? with suest? Correction
I had an extra line in the original code, and the interaction
variable in the last comment was misnamed. Sorry for the double-send.
Steve
On Dec 13, 2006, at 8:47 PM, Pura Rayco-Solon wrote:
We did a reseach using a controlled before-and-after
quasi-experimental design, but with a different set of
subjects at endline. How do I analyse this data?
Pura, I am going to assume that you have svyset your data properly to
reflect the sampling design. If you need help
with svyset, please ask. (The variable group (or province) should be
the stratum variable.)
In what follows, I assume that your intervention area and time
period variables are "group" and "period"
group 1 = control 2 = intervention
period 1= baseline 2 = endline
y response variable
The following code will then work for you.
Good luck!
Steve
/* Code for Pura's Question */
gen gp11= group==1 & period==1
gen gp12= group==1 & period==2
gen gp21= group==2 & period==1
gen gp22= group==2 & period==2
svy: reg y gp11 gp12 gp21 gp22, nocons
lincom (gp22-gp21) // End - Baseline Group 1
lincom (gp12-gp11) // End - Baseline Group 2
lincom (gp22-gp21)-(gp12-gp11) // Group difference
/* ALTERNATELY */
xi: svy: reg y i.group*i.period // __IgroXper_~2 interaction
line will give you the same result as the last lincom
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
