[R] a questtion about Cronbach's alpha more than 1

[(Ted Harding)]

On 20-Feb-08 17:34:33, Smita Pakhale wrote:
> Dear All,
> I am a novice in R and am working on my thesis for
> master's in Epidemiology & Biostatistics. While
> calculating Cronbach's alpha for the SF-36
> questionnaire, I got all 8 of them more than 1. Is it
> possible to get alpha more than 1. If yes, please help
> me understand it. I read about the negative alpha
> value on your web postings. It does not answer my
> question though.
> Thank you.
> Sincerely,
> Dr. Smita Pakhale, MD
> McGill University, Canada

You should not get alpha > 1 if you calculate it by
the standard definition. You can of course get negative
values which are arbitrarily large.

I wonder, therefore, whether
a) You have calculated it by an incorrect method or a
   non-standard definition
or

b) You have somehow been presented with negative alpha < 1
   as positive values > 1.

Can you provide more detail about how you obtained the
values you are writng about?

Best wishes,
Ted.

--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at manchester.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 20-Feb-08                                       Time: 19:55:00
------------------------------ XFMail ------------------------------