When we have very unequal sample variances, we should
A) check our work very carefully when we pool the variances.
B) not pool the variances.
C) average the variances, but not weight them with the degrees of freedom.
D) run some different analysis.

Question

Quizplus · Accepted Answer

When we have very unequal sample variances, we should not pool the variances because the assumption of essentially equal variances for each treatment group may not hold, and pooling the variances can lead to inaccurate results. Instead, we should use a method that accounts for unequal variances, such as the Welch's t-test or a non-parametric test.

When We Have Very Unequal Sample Variances, We Should