'How many calories did I just eat?' An experimental study examining the effect of changes to serving size information on nutrition labels.

<bold>Objective: </bold>To test modifications to nutrition label serving size information on understanding of energy (calorie) content among youth and young adults.<bold>Design: </bold>Participants completed two online experiments. First, participants were randomly assigned to view a beverage nutrition label with a reference amount of per serving (250 ml), per container (473 ml) or a dual-column format with both reference amounts. Participants were then randomized to view a cracker nutrition label which specified a single serving in small font, a single serving in large font, or the number of servings per bag with single serving information below. In both experiments, participants estimated energy content. Logistic regression analysis modelled correct energy estimation. Finally, participants reported their preference for serving size display format.<bold>Setting: </bold>Canada.<bold>Subjects: </bold>Canadian youth and young adults (n 2008; aged 16-24 years).<bold>Results: </bold>In experiment 1, participants randomized to view the nutrition label with per container or dual column were more likely to correctly identify energy content than those using per serving information (P<0·01). For experiment 2, the serving size display format had no association with correct energy estimation. The majority of participants (61·9 %) preferred the serving size format that included servings per package.<bold>Conclusions: </bold>Labelling foods with nutrition information using a serving size reference amount for the entire container increased understanding of energy content. Consumers prefer nutrition labels that include more prominently featured serving size information. Additional modifications that further improve consumers' accuracy should be examined. These results have direct implications for nutrition labelling policy. [ABSTRACT FROM AUTHOR]