Consider scenario 1 (compared with scenario 2):

Within-group variation is relatively Select an answer small large , and therefore the F statistic $\displaystyle \frac{\text{Variation among sample means}}{\text{Variation within groups}}$ will be relatively Select an answer small large , indicating that $\displaystyle {H}_{{0}}$ will probably Select an answer be rejected not rejected , and we'll conclude that the population means Select an answer may be equal are not all equal