Question

3. Imagine three treatment groups are being compared using Analysis of Variance. What effect would each of the following have on an F score (make it smaller, make it larger, or cause no change)? Explain why.

a. A large within-group error.

b. One group with a mean that is two standard deviations away from the other groups.

c. A great deal of overlap between two groups.

d. Adding a bunch of outliers (extreme scores) to all three groups being compared.

Answer 1

F = MS between / MS error

consider the above formula for the answers below

a. A large within-group error.

it will make smaller the vale of F because it comes denominator

b. One group with a mean that is two standard deviations away from the other groups

it will increase the variance BETWEEN groups, so it will make larger the F score

c. A great deal of overlap between two groups

it will decrease the variance BETWEEN groups, so it will make the F score smaller

d. Adding a bunch of outliers (extreme scores) to all three groups being compared

it will increase the variance of with-in group , so F score will get smaller