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

Re: st: comparison between a repeated ordinal measures


From   "rengaweoj" <wagnerj@edc.pitt.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: comparison between a repeated ordinal measures
Date   Thu, 31 Mar 2005 16:19:39 -0000

--- In statalist@yahoogroups.com, Joseph Coveney <jcoveney@b...> wrote:
> Joseph Wagner wrote:
> 
> Thank you for your response.  My data is in -wide- format.  In order
> to use -ologit- or -gllamm-, does my data necessarily need to be in
> -long- format?  I am guessing that either I need to or I have set up
> -ologit- and -gologit- incorrectly since the coefficient and p-value I
> get are quite different.  Looking at the table it appears there isn't
> a difference between the two time points and the results I get from
> your programming supports this.
> 
>
--------------------------------------------------------------------------------
> 
> Yes; the data need to be -reshape-d long in order to use -ologit- or
-gllamm-.  
> 
> Inspection of the -ordplot- plot leads me to the same conclusion
that you've 
> drawn:  the scores are essentially the same between the two
observation time 
> points.  (I'm sorry for my confusion over the relation between
-ordplot- and 
> -distplot-.  Thanks to Nick for straightening me out on this.)
> 
> Ralph B. D'Agostino, Sr., has written on robustness of Student's t
test for 
> ordered categorical data--did you get the chance to try -mvrepeat-
or -ttest- 
> and compare its results to the others'?
> 
> Joseph Coveney


Thanks for your help. -ttest- and -mvrepeat- gave similar results to
each other and to -ologit- in the long data.  However, -npt_s-
returned an error and -gllamm- returned "no observations"

I used -ttest- and -mvrepeat- on the wide data and these results:

                Ho: mean(hvhlths - f6hlths) = mean(diff) = 0

  Ha: mean(diff) < 0    Ha: mean(diff) != 0    Ha: mean(diff) > 0
       t =   0.6863          t =   0.6863          t =   0.6863
   P < t =   0.7535     P > |t| =   0.4929      P > t =   0.2465



         One-way MANOVA for Repeated Measures

dependent variables:   hvhlths f6hlths
                             S=1  M=-.5  N=186.5

Test            Value     F      Num DF    Den DF   Pr > F
-----------------------------------------------------------
Wilks' Lambda  0.998745  0.471     1      375.0000  0.4929


When I -reshape-d long:

somersd eval hlths, cluster(id)
------------------------------------------------------------
        |          
   eval |  Coef.        z    P>|z|     [95% Conf. Interval]
--------+----------------------------------------------------
  hlths |  -.0088652  -0.35   0.727    -.0586776    .0409471
-------------------------------------------------------------


ologit hlths eval, cluster(id)

 hlths |    Coef.       z    P>|z|     [95% Conf. Interval]
-------+-----------------------------------------------------
  eval |  -.030391    -0.35   0.727    -.2007067    .1399247
-------------------------------------------------------------


npt_s hlths, by(eval) strata(id) nodetail

gives: r(198);


gllamm hlths eval, i(id) family(binomial) link(ologit)

gives: no observations




*
*   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