The Effect of a Variety of Nutrition Label Format Alignments on Recognition, Application and Total Comprehension Scores of Consumers

Abstract

In order for a nutrition label to be effective, consumers must comprehend the format and find it visually appealing. FDA was frustrated in its attempt to find a format that consumers preferred and could comprehend. Our study evaluated five food label formats [tabular control, bar graph (Geiger, et al., 1991). highlighted bar graph, adjectival, and pie chart] to determine if there were any significant correlations between consumer preference and recognition, application and total comprehension scores. We hypothesized, by changing only the format and holding all other information and information load constant, that there should be significant correlations between preference scores (utilities) and comprehension scores, or at the very least, no significant differences (negative correlations). A shopping mall intercept study using a computer interactive interview was completed on 121 subjects in a major mall in Salt Lake City, Utah, as a component of a larger study. Formats were printed on Swanson's TV dinners. Each respondent answered a block of four comprehension questions for each format which assessed recognition, application and total comprehension. Order of labels and questions responses were randomized. Differences among comprehension of formats were determined by Friedman's nonparametric two-way ANOVA with an associated multiple comparison test. Preference scores were determined using adaptive conjoint analysis for each format after the comprehension questions were completed. Correlations were performed between utility scores for each format and the number of correct questions for recognition, application and total comprehension questions. Consumers significantly preferred (p<.001) the bar graph (Geiger, et al., 1991), tabular control and highlighted bar graph formats. On the total comprehension, tabular control, bar graph (Geiger, et al., 1991) and highlighted bar graph were not significantly different and scored highest (p<.05). The only significant correlation was found for the adjectival format and its recognition score (p<.05). A Kruskal-Wallis one-way non-parametric ANOVA was performed for each format by category (for 0 through 4 correct scores). Those consumers scoring 1 (50 %) correct for the recognition questions for the pie chart format had a higher utility (p<.05) for the pie chart format than consumers scoring 0 or 2 correct for the recognition questions. In conclusion, there were only two significant positive correlations between comprehension scores and consumer preference for food label format. However, preference and total comprehension score were highest for tabular control, bar graph (Geiger, et al., 1991) and highlighted bar graph formats, which suggests that comprehension is not significantly inversely related to preference.