type II error?? - (Mar/05/2011 )

what stats should be used if difference between the means are extremely high? actually i used different post-hoc tests like LSD, Tukeys LSD, Bonferroni for multiple comparision, but all of them are showing non-significant differnce between means inspite of high difference. when i compare the means by using t-test (control with test; test with treatment1; test with treatment 2..and so on) result is highly sidnificant (p<0.0001). dont know what to do???

You could try the Mann Whitney U test, it's non-parametric and more robust to outliers.