Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: re. correct analysis of unusual outcome scale data


From   "roger webb" <roger.webb@man.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   st: re. correct analysis of unusual outcome scale data
Date   Wed, 9 Jul 2003 11:56:40 +0100

Dear Statalist

I have four outcome rating scales that measure separate domains 
of mother-infant interaction on a scale of 1 (poor) to 5 (good). (Fiori-
Cowley A, Murray L, Gunning M. Global ratings of mother-infant 
interaction at 2 and 4 months. 2nd edition. University of Reading, 
UK).

The 4 domains are:
1. non-intrusive behaviour/intrusive behaviour
2. non-intrusive speech/intrusive speech
3. non-remote/remote
4. non-silent/silent

According to the authors instructions, the four domains can be 
combined to create an 'intrusive to remote' summary score as 
follows:

((intrusive behaviour+intrusive speech)/2 - (non-remote+silent)/2)/2

This summary score runs from -2 (intrusive) to +2 (remote), with a 
score of 0 indicating that the mother is neither intrusive or remote.

Question:
I'm unsure how I should analyse these data as both extremes of 
the scale (and anything other than a zero score) denote a poor 
outcome. It seems to me that using linear regression would 
therefore yield uninterpretable coefficients. Does anyone have any 
better suggestions?

Thanks in advance
Roger Webb

School of Psychiatry & Behavioural Sciences
University of Manchester (UK)



  
      
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index