re: st: multiple analysis using ANOVA?

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

> Hi,
> First off excuse my ignorance as I'm an engineer attempting to do some
> statistical analysis on gene expression data I've genereated.
> Basically I have 5 genes of interest (3 candidate and 2 housekeeping) 
> and 3
> Blots which I have created myself (Triplicate). Each blot has cDNA 
> printed
> on it from samples as follows:  4 time points, two groups at each time
> point: control and experimental, with 6 animals per group. From this i
> generate ratios of each candidate gene to each housekeeping gene for 
> each
> animal over time.
> I am looking to do a multiple analysis using a two-way Anova to 
> determine if
> at any particular time point if a particular gene is elevated at a
> particular time point.
> Has anyone managed to get the Neumann Keuls test running in Stata?
> Sorry if i've confused anyone,
> Thanks in advance,
> David Hardiman
I do not think the Newman-Keuls multiple comparisons test is 
implemented in Stata. I couldn't find anything. On the other hand, most 
textbooks I've read suggest the method does not control the 
experiment-wise error to 0.05 as claimed. SAS, SPSS, and Minitab offer 
one click access to multiple comparison tests of all kinds; whether you 
should use a particular one in certain situations over others is 
another problem. There is a nice chapter on multiple comparisons in the 
new version of Maxwell and Delaney's text which speaks to this. You can 
combine anovatest with multiple test corrections in Stata 8, but that's 
about it, other than coding what you need directly from the data and 
what anova leaves behind.

-Dave

*
*   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/