Skip to content

basic anova and t-test question

2 messages · Arne.Muller@sanofi-aventis.com, Spencer Graves

Original
5 days later
Spencer Graves
# 
Does the following answer your question:

 > set.seed(1)
 > z0 <- rnorm(100)
 > p.z <- 2*pnorm(-abs(z0))
 > sum(p.z<0.05)
[1] 5
 > pchisq(sum(z0^2), 100, lower=FALSE)
[1] 0.917285

	  Some of the 100 (in this case) normal random deviates seem 
statistically significant, even though the ensemble is not.

	  spencer graves
Arne.Muller at sanofi-aventis.com wrote:

            

      
      
      
    

        
but in the t-test of the summay 2 of the 4 levels (one for
constrast) are significant. See below for some truncated output.

            

      
      
      
    

        
but I'm looking for for a "meaning". Maye you have a schenario
that explains how these differences can be created and how you'd
go ahead and analyse it further.

            

      
      
      
    

        
significant in either anova nor t-test, and a boxplot of the
factor shows that the levels are all quite similar (similar
variance and mean). Might the effect I observe be linked to
an unbalance design in the multifactorial model?