Re: st: Repeated measures

[David Airey <david.airey@vanderbilt.edu>]

.

I don't see this so much as a repeated measures design as a completely  
nested design. That is, I don't think you can or want to really equate  
mean differences at trial 1 across all trial 1s. Rather you are  
probably doing 5 trials simply to gain more accuracy in the measure of  
a particular wire. Similarly, I don't think you can equate WireNo 1  
across WireTypes. If so, then Trial is nested in WireNo and WireNo is  
nested in WireType, but WireType is a fixed effect. True?

On Mar 14, 2009, at 3:16 AM, Nikolaos Pandis wrote:

> Dear Statalisters,
>
> I have the following dataset (except I have it in long format-one  
> measurement per row) :
>
> WireType	WireNo	Trial		Force
>
> A	A1	1,2,3,4,5	      5 measurements-1 per trial
> A	A2	1,2,3,4,5	      5 measurements-1 per trial
> A	A3	1,2,3,4,5	      5 measurements-1 per trial
> A 	A4	1,2,3,4,5	      5 measurements-1 per trial
> A	A5	1,2,3,4,5	      5 measurements-1 per trial
>
> B	B1	1,2,3,4,5	      5 measurements-1 per trial
> B	B2	1,2,3,4,5	      5 measurements-1 per trial
> B	B3	1,2,3,4,5	      5 measurements-1 per trial
> B	B4	1,2,3,4,5	      5 measurements-1 per trial
> B	B5	1,2,3,4,5	      5 measurements-1 per trial
>
>
> C	C1	1,2,3,4,5	      5 measurements-1 per trial
> C	C2	1,2,3,4,5	      5 measurements-1 per trial
> C	C3	1,2,3,4,5	      5 measurements-1 per trial
> C	C4	1,2,3,4,5	      5 measurements-1 per trial
> C	C5	1,2,3,4,5	      5 measurements-1 per trial
>
> D	D1	1,2,3,4,5	      5 measurements-1 per trial
> D	D2	1,2,3,4,5	      5 measurements-1 per trial
> D	D3	1,2,3,4,5	      5 measurements-1 per trial
> D	D4	1,2,3,4,5	      5 measurements-1 per trial
> D	D5	1,2,3,4,5	      5 measurements-1 per trial
>
> I am testing the forces generated by 4 different wiretypes to see if  
> there are significant differences.
> For each wiretype (categorical A B C D) I am using 5 specimens  
> (A1-5,..D1-5) and for each specimen I am repeating the measurement 5  
> times.
> The outcome is force and it is a continuous variable.
>
> This looks like a repeated measures with each trial nested in the  
> WireNo and the WireNo nested inside WireType?
>
> I was trying to use the xtgee command but I was not sure how to  
> construct it:
> xi:xtreg force i.xxxxx i.yyyyyy, i(zzzzz) ?
>
> Any help would be appreciated.
>
> Thank you very much,
>
> Nikolaos
>
>
>
>
> *
> *   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/