Re: st: Comparing two response variables

[Roger Newson <r.newson@imperial.ac.uk>]

It depends what you mean by "better". However, if you want to compare 
ordinal predictive power, then you can compare the Somers' D or 
Harrell's c statistics of the outcomes with respect to the predictors. 
This can be done using -lincom- after the -somersd- package, 
downloadable from SSC, which also has a set of .pdf manuals. The 
statistical issues are discussed in Newson (2002) and Newson (2010).

I hope this helps.

Best wishes

Roger


References

Newson RB. Comparing the predictive power of survival models using 
Harrell’s c or Somers’ D. The Stata Journal 2010; 10(3): 339–358. 
Purchase article for $8.75 from
http://www.stata-journal.com/article.html?article=st0198
or download pre-publication draft from
http://www.imperial.ac.uk/nhli/r.newson/papers.htm

Newson R. Parameters behind "nonparametric" statistics: Kendall's tau, 
Somers' D and median differences. The Stata Journal 2002; 2(1): 45-64. 
Download from
http://www.stata-journal.com/article.html?article=st0007


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: r.newson@imperial.ac.uk
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/

Opinions expressed are those of the author, not of the institution.

On 28/02/2011 23:20, Debs Majumdar wrote:
> Hello,
>
>     I was asked this question today "Is there any way one can say one dependent
> variable is better than the other for the following situation?"
>
>    Suppose, you have two response variables Y1 and Y2 on the same metric, one a
> composite of 8 items and the other with 10 items (same 8 items + 2 other). You
> have two predictors (X1 and X2, say) and you run the following regressions:
>
>   Y1 = a_0 + a_1*X1 + a_2*X2 + e1
>
>   Y2 = b_0 + b_1*X1 + b_2*X2 + e2
>
> Is there anyway to prove Y1 is a better measure for the trait we are measuring
> when compared to Y2?
>
> I don't have a clear cut answer for this. Is using `-sureg' appropriate for this
> case? Any help is appreciated.
>
> Thank you,
>
> Debs
>
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/