why pooled Std err is bigger than the mean value of a group, when the sample size of three groups used in the analysis are different. Thanks.
