Why Welch's test is Type I error robustBen Derrick, P. White|The Quantitative Methods for Psychology|2016Cited by 255