A question for stats wizzards: ttest vs. anova

Given that a data set is normally distributed and their samples independent these two methods should give the same conclusions? Would you say this is correct? I am asking because in one particular case it does not and I am not sure what causes this discrepancy. Could it be due to power?

[Daniel Shub]

This question has nothing to do with MATLAB... An ANOVA is a generalization of the t-test to multiple (i.e., more than 2) groups. If you have two groups you should get the identical answer. If you have multiple groups then they are not the same and a single ANOVA is better.