Answer:
ANOVA is considered robust to moderate departures from this assumption. But that’s not true when the sample sizes are very different.
there is no good rule of thumb for how unequal the sample sizes need to be for heterogeneity of variance to be a problem.
