In this study, we focused on what the least sample size is at I and II type error level in comparisons of paired and independent group averages and used PS and PASS programmes to estimate sample size. Similar results were obtained from both programmes. In the comparison of the averages of paired and independent groups, we took 0,05, 0,025, 0,01 and 0,001 as I type error ( ), 0,10 and 0,20 as II type (ß). As error level decreases, so does the size of sample increases. In addition, the difference between population averages and the extent of standard deviation value belonging to sample values have an important effect on sample size.