Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: repeated factor ANOVA


From   James prince <jsprince26@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: repeated factor ANOVA
Date   Tue, 2 Oct 2012 14:38:02 -0500

Thank you Phil. Just what I need it.
The reason for the ANOVA is to compared results to a previous analysis
of these data using SAS.
They used PROC MIXED with option REPEAT.
When I used XTMIXED, I could not match their results.

I now match their results with Stata's ANOVA

Thank you again,
Jim

On Tue, Oct 2, 2012 at 9:57 AM, Philip Ender <ender97@gmail.com> wrote:
> On 10/2/2012 James prince <jsprince26@gmail.com> wrote:
>
> Hello,
>
> I am unsure how to set-up the following problem using the ANOVA
> command. Any help would be appreciated.
>
> I have data on 600 liver patients, 300 assigned to one of two
> treatment groups (TREAT).
> Each had 4 biopsy (one in each quadrant: UR, UL, LR, LL) and the
> concentration on an enzyme (Y) measured. I also have data for sex. I
> know that I can used XTMIXED to do this, but I need to use ANOVA
>
> Thank you in advance,
> Jim
>
> --------------------------------------------------------------------------------------
>
> This is what I did:  I first tried using a repeated ANOVA but got an error:
>
> . anova y treat liver sex subj, repeat(subj)
> could not determine between-subject error term; use bse() option
>
> Then,
> If I set it up as nested, I get a result, but I am not sure if it is correct.
> anova y treat liver sex/ subj|liver/ ...
>
> You say you need to use -anova- but you don't say why you need to use
> -anova-.  For balanced designs -xtmixed- and -anova- yield equivalent
> results although somewhat differnt p-values due to chi-square vs
> F-ratio.  If the design is unbalance within subject then -xtmixed- is
> much better.  With your large sample size the p-value issue should be
> moot.
>
> All that aside, here is the -anova- code to run your analysis without
> sex.  The error term for treat is subj|treat.  There should be a treat
> by liver interaction and the repeated factor is liver.
>
> anova y treat  / subj|treat liver  treat#liver /, repeat(liver)
>
> If you want to include sex you need to treat it as a between subject
> factor.  The error term for the between subject factors is
> subj|treat#sex.  You also get more interactions
>
> anova y treat sex treat#sex / subj|treat#sex liver sex#liver
> treat#liver sex#treat#liver /, repeat(liver)
>
> -xtmixed- code is much simpler:
>
> xtmixed y treat##sex##liver || subj:, reml var
>
>
> --
> Phil Ender
> UCLA Statistical Consulting Group
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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