cronbach alpha: Scale reliability coefficient differs from test score alpha

[Steven Saxonberg]

In conducting Cronbach Alpha with STATA 14 the "scale reliability coefficient" differs from the "test score alpha" -- how is this possible? This is the result I get:

my syntax is: .alpha famsuffers preschsuf wrkwarm

my result is:
Average interitem covariance: .7442413
Number of items in the scale: 3
Scale reliability coefficient: 0.7604

Then I wrote:
alpha wanthome preschsuf wrkwarm, generate(suffering) item label

and at the bottom of the table I got
Test scale .5310093 0.6440 mean(unstandardized items)

So that means the test scale was .644 while the scale reliability coefficient was .7604. I thought they should be the same.

[William Lisowski]

I'm a bit confused, your alpha commands list different sets of variables.

The same appears to be the case in your earlier post at https://www.statalist.org/forums/for...st-score-alpha

[Steven Saxonberg]

Dear William, thank you, I see that you are right that I made a mistake in not using the same variables, Thanks for pointing that out!