An insurance company has offices in all 50 states and the District of Columbia. It plans to use an ANOVA to compare the average sales (in dollars) generated per agent among the states and DC. To simplify the analy- sis, assume that it has an equal number of agents in each state (20 per state).
(a) If se = +3,500, then how different must the average sales per agent in one state be from the average sales per agent in another state in order to be statistically significant? Use the Bonferroni approach to adjust for the effect of multiplicity.
(b) If managers of the insurance company believe that average sales per agent in most states are within about +2,000 of each other, is there much point in doing this study?