Is there a python t test for difference? [duplicate]

Question

This question already has answers here:

How to calculate the statistics "t-test" with numpy (3 answers)

Closed 8 years ago.

Is there a t test in a python package where you can test for difference? If there is how do you use it?

For example test two vectors:

a=np.random.randn(5, 7)
b=np.random.randn(5, 7)

I have found the t test : ttest_ind in statsmodels. However I would like to specify the difference to test for, this can not be passed into the ttest_ind function. Does anybody know another way to do this?

Answer 1

ttest_ind in statsmodels has a value option, to test the null hypothesis that the difference between the two means is equal to value

http://statsmodels.sourceforge.net/stable/generated/statsmodels.stats.weightstats.ttest_ind.html