How to perform a two sample t test in python using any statistics library similar to R?

Question

I can do this in R for 2 sample T-test:

t.test(x, y = NULL, alternative = c("two.sided", "less", "greater"), mu = 0, 
       paired = FALSE, var.equal = FALSE, conf.level = 0.95)

I want some function where I can pass this mu(difference in mean) parameter in Python ttest?

schrodingercat · Answer 1 · 2019-07-30T10:20:15.353

2

R has one function t.test() to perform the Student's T test, while Python employs more methods.

If you want to perform a one sample t-test with mu as the true mean μ of the population from which the data is sampled, you should use scipy.stats.ttest_1samp and pass the parameter through popmean. Docs are here.

If you want to perform a two sample t-test with mu as the difference in means, statsmodels.stats.weightstats.ttest_ind from the statsmodels module is the right function, with mu passed through value. Docs are here, as well as a link to a useful answer.

Other links you may find useful are these:

https://plot.ly/python/t-test/ - T-tests with plotly
https://towardsdatascience.com/inferential-statistics-series-t-test-using-numpy-2718f8f9bf2f T-test using Python and Numpy
https://pythonfordatascience.org/independent-t-test-python/ - T-test | Python for Data Science

edited Jul 30 '19 at 10:20

answered Aug 30 '18 at 09:05

schrodingercat

246
4
21

I have seen this function but I cannot pass the true value( or True difference) mu to this function – Ankit Agrawal Aug 31 '18 at 10:04
@AnkitAgrawal with the parameter 'popmean', which is the true mean μ of the population from which the data is sampled. – schrodingercat Aug 31 '18 at 12:10
popmean is there for 1 sampple ttest, but it's not present in two sample ttest – Ankit Agrawal Sep 03 '18 at 05:24
@AnkitAgrawal ok, so you want to test for difference. It wasn't clear from the question. The answer is here: https://stackoverflow.com/questions/20682795/is-there-a-python-t-test-for-difference – schrodingercat Sep 10 '18 at 10:48
@AnkitAgrawal the function you need is here: http://www.statsmodels.org/stable/generated/statsmodels.stats.weightstats.ttest_ind.html – schrodingercat Sep 10 '18 at 10:49
Thank you so much, I was using rpy2 till now! – Ankit Agrawal Sep 11 '18 at 12:44
@AnkitAgrawal I'll edit the answer. So it'll be more precise. – schrodingercat Sep 11 '18 at 13:00
Sure, I'll also clarify the question. – Ankit Agrawal Sep 11 '18 at 13:17
how can I do a one-sided (one tail) paired t test in python ? this https://docs.scipy.org/doc/scipy-0.19.0/reference/generated/scipy.stats.ttest_rel.html is always two sided. – seralouk Jan 15 '19 at 20:40
@serafeim here: https://stackoverflow.com/questions/15984221/how-to-perform-two-sample-one-tailed-t-test-with-numpy-scipy – schrodingercat Jul 30 '19 at 10:26

How to perform a two sample t test in python using any statistics library similar to R?

1 Answers1