What question did this study set out to answer?

The aim is to evaluate the effectiveness of t-tests and various robust methods for comparing two independent samples under non-ideal conditions.

July 6, 2026Quality & QuantityOpen Access

Neither the t-test nor Welch’s test: a case for robust alternatives in two-sample comparisons with non-ideal data

View Full Paper

Ask AI

Bookmark

Authors

Randomized simulations compare t-tests and robust alternatives for handling non-ideal data, suggesting better options.

The aim is to evaluate the effectiveness of t-tests and various robust methods for comparing two independent samples under non-ideal conditions.
Conducted simulations varying sample size, variances, and skewness to compare tests.
Performed an empirical study on gender differences using psychological scales.
Compared classical tests with robust alternatives including permutation-based Welch’s test.
T-test and Welch's test failed to maintain Type I error under unequal sample sizes and skewness.
Permutation-based Welch's test maintained nominal Type I error across conditions.
Anderson-Darling test showed high power when both mean and variability differed in skewed populations.

Joost de Winter (2026) studied this question.

View Full Paper

Ask AI

Bookmark