Hi all,

I'm a little stuck on a paper looking at differences in fasting glucose levels between different ethnic groups. We have collected data on ethnicity and glucose as well as several possible confounders (BMI, smoking status (expressed as 1/0), energy expenditure in avg week). Using various regression techniques and ttests i have determined that bmi is a considerable confounder with strong links to both fasting blood glucose level and the white european subset of my study. I would like to run a t-test along the lines of: ttest fast_glucose, by(eth) but then somehow make it regress for bmi.

Is there a command to do this in stata or am i using the wrong statistical analysis?

Any help appreciated,

David