@whuber stop closing my question, this is really inappropriate to block me from aksing my question!
I am trying again to post my question, as it is not a duplicate and not answered in any other similiar questions.
I have two random variables which are independent and normally distributed.
In this document on page 198-9 it says:
I am wondering about this formula.
In general it holds:
or on Wikipedia:
Therefore if we assume unequal variance, I would think like this:
$s_{\bar{X}_T-\bar{X}_C}=\sqrt{s^2_T/n_1+s^2_C/n_2}$
Where is my mistake here?
In the document later on for the standard error it says:
So their formula for std error corresponds to my formula for the std dev. I don't get the derivation of these $s_{\bar{X}_T-\bar{X}_C}$ vs $SE_{\bar{X}_T-\bar{X}_C}$.
