multiple logistic regression by groups

[sujata]

Hi Friends,

I am fairly new to Stata and trying to learn the strings of statistical analysis using Stata more efficiently.

I am trying to do multiple logistic regressions by groups i.e by regions.
Logit per say wouldnt run if i use it by(groups)
I also tried clogit,

clogit tuberculosis smoker i.age_group , group(region)

But i got an error message
5872 (group size) take 295 (# positives) combinations results in numeric overflow; computations cannot proceed
r(1400);

I there a better way of doing this? Also instead of me copy pasting the results i would like to get the results in an excel or stata file. How can I?

Thanks in advance for your help!

Regards
Sujata

[Rich Goldstein]

as the FAQ says, please show us exactly what you typed (for your logistic); you can certainly do something like the following: by groups: logistic y x

[sujata]

Hi Rich,
Thanks for the response. Mine is a case control data and i want to find the odds ratios of having TB(outcome) if smoker (exposure) _ age adjusted by regions

My data looks like this

input respondent_id region tb smoker case_control age_group_cat age_group
NVL001 1 1 1 1 2 30-45
NVL012 2 0 0 1 2 30-45
NVL023 3 0 1 1 3 46-60
NVL034 2 1 1 1 1 <30
NVL045 1 0 0 0 1 <30
NVL056 1 1 1 1 3 46-60
NVL067 3 0 0 0 3 46-60
NVL078 2 1 1 1 2 30-45
NVL089 3 1 0 1 2 30-45
NVL0910 2 0 1 1 3 46-60
NVL1011 2 1 1 1 1 <30
NVL1112 1 0 0 0 1 <30
NVL1213 3 1 1 1 3 46-60
NVL1314 3 0 1 1 3 46-60
NVL1415 3 0 0 0 1 <30
NVL1516 2 0 1 0 2 30-45
NVL1617 1 0 0 0 2 30-45
NVL1718 2 0 1 0 3 46-60
NVL1819 1 0 1 0 2 30-45
NVL1920 3 0 1 0 3 46-60

I tried using- by regions, sort: logistic tb smoker i.age_group_cat, or
But the problem is they pop up at different tables. Since I have 235 such regions, I want the results to come together in one table or if possible output the tables into an excel sheet. Instead of me having to copy paste the tables one by one.

Thanks again
Sujata

[Rich Goldstein]

see the help for -statsby-