Objective: Validation of a self-monitoring 'portions' measurement of fruit and vegetable (F&V) consumption against a standard of weighed intakes. Design: Component of a randomized controlled trial. Setting: Subjects attended research centres in Reading and Glasgow for instruction and monitoring but undertook free-living dietary changes at home. Subjects: A study sample of 42 adult men and women fulfilling the main recruitment criterion of eating less than five F&V portions/day but contemplating increasing intakes and providing weighted baseline reported energy intakes exceeding (estimated basal metabolic rate´1.1). Interventions: Subjects attended an intensive group advice session which included the specific relationship of high F&V intake with reduced risk of disease; practicalities; portion definition and measurement recording. The target was to exceed five F&V portions/day for 8 weeks. Main outcome measures: Self-recorded simultaneous weighed inventories and F&V portion measures. Results: Data from subjects who were not evident under-recorders showed correlations between portion and weighed intakes of r=0.73, (P<€0.000), although the portions measure tended to under-estimate intakes. Using 80 g/portion the '5-a-day' concept tends to create false negatives (namely consumption could be greater than 400 g whilst recording fewer than five discrete portions) but rarely false positives (namely recorded consumption of less than 400 g did not give measures of more than five discrete portions). Conclusions: The data suggest that the five portions F&V/day health message, if used in conjunction with defined discrete portions, would encourage desirable consumption exceeding 400 g. Sponsorship: Ministry of Agriculture, Fisheries and Food. | 
