Consider a study that seeks to compare four diets (low-fat, low-carb, Zone, and DASH), plus a control condition (five groups of people). Two hundred participants are randomly assigned and trained to follow one of the four diets or are assigned to the control condition in which they receive training on general healthy eating. Approximately 40 participants are assigned to each of the five groups. Each participant's weight is recorded before and after the study period. Suppose this study also seeks to compare the effects of exercise in conjunction with each of these diets, so the researchers assign half of the participants in each diet group to also participate in an exercise program. Why wouldn't a single-factor ANOVA be appropriate for comparing the diet and exercise conditions?