Abstract by Brittany Russell

Personal Infomation

Presenter's Name

Brittany Russell

Degree Level


Abstract Infomation



Faculty Advisor

Dennis Tolley


How does Heteroskedasticity Affect the Power of the F-test of a Slope Coefficient?


A common assumption of linear regression is homoskedasticity or equal variance of the errors, but this assumption is often violated in real data. In this simulation study I simulate data with one independent and one dependent variable, varying the degree of heteroskedasticity. I then create power curves of the F-test on the significance of the slope coefficient. I also simulate homoskedastic data with high and low variance and compare the power of the test on those data. The results show that heteroskedasticity does hurt the power, but the test has higher power with heteroskedastic data than with high variance homoskedastic data.