Posted by
Swank, Paul R on
Dec 04, 2006; 4:23pm
URL: http://spssx-discussion.165.s1.nabble.com/split-half-reliability-tp1072479p1072487.html
Actually, Cronbach's Alpha is equivalent to the average of all Guttman-Flanagan split halves. Spearman-Brown coefficients, since they assume parallelness, tend to over estimate the split half reliability when the halves aren't parallel. The Guttman-Flanagan split half does not assume parallelness but assumes the halves are essentially Tau equivalent, the same assumption Cronbach's alpha makes about items. In fact, the Guttman-Flanagan split half is the alpha formula applied to the halves.
Paul R. Swank, Ph.D.
Professor, Developmental Pediatrics
Director of Research,
Children's Learning Center
University of Texas Health Science Center at Houston
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of Antonio Maurandi Lopez
Sent: Monday, December 04, 2006 10:12 AM
To:
[hidden email]
Subject: Re: split-half reliability
Makes sense. If you compute one split-half reliability and then randomly divide the items into another set of split halves and recompute, and keep doing this until we have computed all possible split half estimates of reliability. Cronbach's Alpha is mathematically equivalent to the average of all possible split-half estimates.
So, if you've got a computer better use Cronbach.
Antonio Maurandi
Art Kendall escribió:
> Split half-reliability was done in the days of hand computation to
> estimate internal consistency.
>
> It is pretty much a historical artifact. It is twenty or twenty-five
> years since I have seen it actually used.
>
> Since you are posting on The SPSS list, I suggest you use Cronbach's
> alpha from the RELIABILITY procedure.
>
> Art Kendall
> Social Research Consultants
>
> AR wrote:
>
>> Dear list:
>>
>> When testing for split-half reliability is it a good number a
>> Spearman-Brown Coefficient of 0.88?
>> n=1300. and I have 16 items in my questionnaire.
>>
>> Thank you in advance.
>>
>>
>