Network Research Highlight: Type I Error Rates and Heteroscedastic Errors

Date: Wednesday, March 20, 2019

Work Science Center Network Member, Alice Brawley Newlin, recently published an article in the Journal of Modern Applied Statistical Methods examining some of the basic assumptions we make in the statistics we use to answer our research questions. Specifically, the research team used Monte Carlo simulations to explore the utility (or lack thereof) of preliminary tests of homoscedasticity in general linear models. The authors found that these tests may be unnecessary when testing for differences in independent slopes.

Leave a Reply

Your email address will not be published.