calculating single pvalue when running multiple regression analysis

[Rowa Taha]

Hi,
I'm hoping someone is able to help, as I've seen similar queries posted. I'm running a multiple regression analysis using 3 variables, all of which are categorical (contain 2 categories each), adjusting for age as the randomisation was stratified by age and the outcome of interest is binary, whether they did or didn't attend a study follow-up point. The variables pertain to different SMS strategies used to try to optimise data completion and adherence to follow-up. Participants received a combination of SMS using a factorial design.

The code I am using is below, but I'd like to calculate a single pvalue for the overall model, adjusted for each of the others using a likelihood ration test. the code I am using is as follows.

logistic attended_6wks i.frequency i.response i.automation age

I can calculate p values for each variable using liklehood ratio test as follows for frequency for example,

logistic attended_6wks i.frequency i.response i.automation age
est store A
logistic attended_6wks i.response i.automation age
est store B
lrtest A B

But how do I calculate a single pvalue for the overall model, ie all all adjusted for each other?

Thank you so much for the help.

[George Ford]

The chi2 of the model (and its probability) is a test of all the regressors (is the model any better than one with just a constant term).

What, other than that, are you trying to sort out? Whether a smaller model is different from a bigger one?

[Rowa Taha]

Im trying to work out the cumulative effect, ie the the pvalues adjusted for the other interventions.