Dear stata users,
Greetings
My research involves assessing patients' satisfaction with the quality of
primary health care centres. in total there were 866 completed
questionnaires and the study involved 12 primary care centres from two
different sectors(Six for each sector). At this stage I've been running few
tests to compare the satisfaction mean for both sectors.
I'm using an alternative method for the normal t-test which as the following
clttest satisfaction, cluster(primary-care-centres) by(sector)
I'm also using an alternative test for Chi2 which is:
clchi2 gpa_satf agegroup, cluster( primary care-centres)
I'm also running xtreg adjusted for clustering:
Iis primary care
Xtreg satisfaction age, mle
The problem is when using robust regression I get completely different p
value and more surprisingly coefficients differ markedly, I'm not sure why
this is the case.
I'm wondering if anyone could help with these two issues:
1- Why the alternative t-test produces remarkably higher p value? And is
there an alternative test(s)?
2- Why different regression models produce different results?
Many thanks for you all
Ibraheem
*
* 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/
