We conducted this study to determine whether 3 days of dietary records from randomly selected days or from consecutive days more closely estimated the energy and nutrient intake values obtained from 16 days over a 1-year period. Participants consisted of 228 men and women, black and white. No one of three random-day samples or one of five consecutive-day samples better estimated energy and nutrient intake. The group means for intake of energy and seven nutrients were similar for random and consecutive days. However, for energy, protein, fat, carbohydrate, and calcium, the means of the absolute differences between the values for 16 days and those for the random days were significantly smaller than the corresponding means between the values for 16 days and those for consecutive days. Random-day samples may be preferable for describing dietary intake of individuals and small groups when the gain in reliability is worth the cost.