Cronbachs Alpha

#1
Okay, so I'm a student and have to calculate the reliability of a test, I have the test devided into it's various measures - only one (containing 13 questions) will be run through Cronbachs as the other two are not required. I also have data from three samples - each of which represents the results from three levels of the independent variable.

So, each of the three sample results reflect a different manipulation. So do I simply put each of the three data sets through Cronbachs alpha (via SPSS), then average the three answers?

I've tried looking on the internet, but no where seems to have anything, neither do my text books.

Hair pulling stuff.

Regards

Brendan.
 
#2
Right, so I've computed Cronbach's Alpha for the three samples, S1 is.8666, S2 is .8883, and S3 is .8863. So all nice and high, can I just average these? Or should I state the Alpha for each of the three samples?

Regards

Brendan.