How to find Student's t-test using python?

Question

Python

Sorry for the imprecisely worded question.

The teacher suggests finding the Student's t-test using this formula:



However, I noticed that I get different values ​​compared to the scipy.stats.ttest_ind method -1.6127911643120085

- my answer
is -1.6123875817242772 - the answer received through ttest_ind(ctrl, test, equal_var =False)
-1.6126205013707797 - response received via ttest_ind(ctrl, test, equal_var=True)

My function now:

ctrl = df[df['group'] == 'A']['converted'] # pandas Series, control group
test = df[df['group'] == 'B']['converted'] # pandas Series, test group

t = (np.mean(ctrl) - np.mean(test)) / np.sqrt(
    ((np.std(ctrl) ** 2) / ctrl.size) + ((np.std(test) ** 2) / test.size )
)

Please tell me if I made some mistake or the difference in values ​​is due to something else?

Answer 1

I don't know what sample you have, but just in case, I would calculate the standard deviation by setting np.std (data, ddof = 1 ).
In my opinion, the correct option is using ttest_ind(ctrl, test, equal_var=False ). I would compare with him.