Currently, student's t-test is the only option when conducting a t-test. For samples where the assumptions of equal variance and sample sizes are not met, it would be nice to be able to use Welch's t-test instead.
This can either be done by letting the user choose which test to use or just use Welch's t-test as a default (and make that transparent to the user).
Scipy's t-test function, which is currently used, has a flag to decide, which test is used.
Currently, student's t-test is the only option when conducting a t-test. For samples where the assumptions of equal variance and sample sizes are not met, it would be nice to be able to use Welch's t-test instead.
This can either be done by letting the user choose which test to use or just use Welch's t-test as a default (and make that transparent to the user).
Scipy's t-test function, which is currently used, has a flag to decide, which test is used.