one way anova overall significant, but dunnett post hoc has no single significant group?

Posted by chengfoh on Oct 18, 2012; 10:18am
URL: http://spssx-discussion.165.s1.nabble.com/one-way-anova-overall-significant-but-dunnett-post-hoc-has-no-single-significant-group-tp5715709.html

Hi hope someone can really help, this confuse me alot

I would like to compare the mean for 3 test groups to a control group fro different parameters (RBC, ALT, ALP....), so i ran a oneway ANOVA and then i conduct the post hoc using Dunnett's test..then i observed someting confusing me, whereby

There were parameter whereby when ANOVA overall P value was significant, however dunnett test gave no no single group with significant value, how possible?

Also, there were situations where overall ANOVA test was not significant, but posthoc (which i included sometime for test run) show some test groups with significant P value as compared to contorl...

Wish someone came across with these can help me, many thanks
Note: the number of subject in each group might differ, for example, each group can have 3/4 animals and is not throughouhly fixed