Refer to the energy expenditure data of Exercise 12.2.9. Each subject's expenditure value (V) is the average of two measurements made on different occasions. It might be proposed that it would be better to use the two measurements as separate data points, thus yielding 14 observations rather than 7. If this proposed approach were used, one of the conditions for inference would be highly doubtful. Which one, and why?