survey data -cronbach alpha

[Soumya Upadhyay]

I have a survey data. There are two variables that form a construct. To test if they hang together, I ran the cronbach alpha while using the svy: command.
For example, svy:alpha x1 x2, std item

But I get an error as below-
alpha is not supported by svy with vce(linearized).

First, is there a need to test the inter-reliability coefficient while using the svy command? Or is it okay if I did it without it. If I run chronbach alpha without svy, it runs fine. Thank you!

[Soumya Upadhyay]

I have a survey data. There are two variables that form a construct. To test if they hang together, I ran the cronbach alpha while using the svy: command.
For example, svy:alpha x1 x2, std item

But I get an error as below-
alpha is not supported by svy with vce(linearized).

First, is there a need to test the inter-reliability coefficient while using the svy command? Or is it okay if I did it without it. If I run chronbach alpha without svy, it runs fine. Thank you!

Found the Stata manual which has all the commands that go with svy. So, I believe alpha does not go with svy.

[daniel klein]

You could try and get the covariance matrix for the two variables, using correlate with aweights (as suggested here) then compute alpha from the covariance matrix. Whether this is needed in any sense, I do not know.

Best
Daniel