Figure 3.

Relationship between dietary energy density (expressed in kcal g⁻¹) (predictor variable) and energy intake (kcal d⁻¹) and total food intake (g d⁻¹). Analysis was derived from 6155 subjects self-recording their food intake by 24 h recall over 4 days [77,78]. No corrections for plausibility of dietary energy intake were made. Sloping lines were a random sample of 40 individual regression lines, based on 4 days, to give an indication of inter-individual variation. All association were significant at p < 0.001. (Online version in colour.)