Age Adjusted Means

[ryc]

Hi All,

I am relatively new to Stata and lack the knowledge on how to produce age-adjusted means and prevalence rates for use in a Table 1 of an upcoming manuscript. Based on what I do understand, I can produce age-adjusted means using the adjust post-estimation command following an anova or regress command. Although this seems correct, I'm having an issue understanding why group-specific results differ from the command which contains the by(group) option. Please view the following code for a better illustration of what I mean:

regress y age if group==0
adjust age, se
*xb = 3.99107

regress y age if group==1
adjust age, se
*xb = 4.8631

regress y age group
adjust ageebct, by(group) se
*xb0 = 4.00985
*xb1 = 4.89223

I am having trouble reconciling why the adjusted means of the first and second models differ from the stratified means of the thirdmodel. Also, I'm not even sure if I could get an age-adjusted mean of the full un-unstratified sample.

If someone can point me in the right direction or at least explain which of the above three models are proper, I would greatly appreciate it.

Thanks for your help.

Rick