Quick Question about Null Effects in ANOVAs

#1
Hi all,

I just have one quick question:

I've been analysing a reasonably large experiment recently using a repeated-measures ANOVA (with a single between-subjects factor).

It turns out that the between-subjects factor is non-significant, so can I safely run further ANOVAs on portions of the data (to see where the differences between the other factors lie) and exclude the between-subjects factor, or would that be a bad idea? As far as I can see, it looks like the inclusion of this additional factor is making some other effects non-significant...

Any advice would be greatly appreciated!
 

Masteras

TS Contributor
#2
rm anova

you said the 1 and unique between subjects factor is non significant. I suggest you run rm anova again for the within subjects factor only and do paired t-tests (one at the time or all in one table ata once), but you will adjust the a using bonferroni correction.