I have taken two types of subjects related to two categories one has a sample size of 189 and other has 50...Can I apply independent samples T test on them ????
3
-
2$\begingroup$ Why not? Why do you think there is a problem? $\endgroup$ – kjetil b halvorsen♦ Nov 3 '14 at 10:23
-
$\begingroup$ See also stats.stackexchange.com/q/63856 $\endgroup$ – shadowtalker Nov 3 '14 at 12:19
-
1$\begingroup$ If the sample sizes are so different the test is more sensitive to failure of the equal variance assumption; if the variances are similar, it should be fine. $\endgroup$ – Glen_b Nov 3 '14 at 12:26
Add a comment
|
$\begingroup$
$\endgroup$
Yes, you can. The equality of sample numbers is not a requirement for T test at all, but both samples should be greater than 10 or 15. Considering your both sample sizes are greater than 25, you could even use Z test to compare normally distributed variables in them.
