Dietary restraint and self-reported meal sizes: diary studies with differentially informed consent.

Psychometric methods were used to explore the reliability and criterion validity of self-reported food intake in studies of dietary restraint. In Study 1 the reliabilities over days of daily aggregate intakes and of intakes at meals at particular times of day were assessed in 7 day food diaries by 27 low-BMI females. The sizes of particular meals correlated poorly with each other and with the total of all other meals; daily aggregate intakes also had poor reliability (Cronbach's alpha). Individuals meal sizes were consistent from day to day, with high inter-correlations between meal sizes, high correlations between meals at particular times and the sum of the remainder and high reliabilities. Aggregate intake had moderate criterion validity. Of individual meals, only breakfast achieved criterion validity, but there was a significant cubic component in its relationship with restraint. In Study 2, young male and female participants with various BMIs, completed a food diary on a single day. Again, aggregate daily intake had low reliability. Total intake and breakfast both had criterion validity, dietary restraint correlating negatively with total intake and breakfast size in the whole sample and in females, but there were significant quadratic components in the relationships. In contrast, restraint correlated positively with lunch size in the whole sample and in males. The combination of low reliability of individual meals as estimates of total intake, and the low criterion validity of all meals except breakfast, suggests that it may be inappropriate to study dietary restraint using aggregate self-reported intake measures.