A retrospective study examined the link between baldness and the incidence of heart disease. In the study, 1435 middle-aged men were selected at random and examined to see whether they showed signs of Heart Disease (or not) and what amount of Baldness they exhibited (none, little, some, or much).

A student runs a two factor ANOVA on these data and finds the following ANOVA table;

a) Comment on her analysis. What problems, if any, do you find with the analysis?
b) What sort of analysis might you do instead?

Number of Men Baldness None None Little Little Heart Disease Yes No Yes No Yes No Yes No 251 331 165 221 195 185 52 Some Much Much 35 Mean DF Squares Square F-Ratio P-value Sum of Source Baldness 3 62441.375 20813.792 17.956 0.020 Heart disease 1485.125 1485.125 .281 0.340 Error 3 3477.375 1159.125