Proposition for a shortened version of the Leeds Food Preference Questionnaire (LFPQ)

The Leeds Food Preference Questionnaire (LFPQ) assesses, among other components, food preference (FP) using a procedure that requires both time and concentration. Its use may therefore be difficult in a complex protocol. In this article, we assessed the suitability of two shorter versions: 1) the LFPQ-S₂, using two instead of four foods per group, and 2) the food preference questionnaire (FPQ-S₁₆), using the 16 same foods as the LFPQ, displayed simultaneously, that have to be ranked in order of preference. We assessed the reliability of these short versions using correlations and interchangeability (Bland and Altman) analyses. Sixty-nine healthy French adults completed the original LFPQ using foods adapted to the French context, and then the FPQ-S₁₆ procedure before (hungry state) and after (satiated state) a classic lunch. Finally, they had to select the two foods that best depict the characteristics of each group. The two foods receiving the most votes were used to calculate FP scores with the LFPQ-S₂ procedure. The FP scores were no different between the three questionnaires and were modified similarly by lunch intake. The FP scores also highly correlated between the two short versions and the LFPQ (mean r = 0.83 and 0.88 in the hungry state and r = 0.86 and 0.87 in the satiated state for LFPQ-S₂ and FPQ-S₁₆, respectively). Finally, the FPQ-S₁₆ was considered to be a better surrogate of the LFPQ than the LFPQ-S₂. Scores of specific categories (groups of foods categorized by one property; e.g. high-fat foods) showed modest-to-good agreement, whereas scores of combined categories (groups of foods categorized by combined properties, e.g. high-fat and sweet foods) showed poor-to-modest agreement. Overall, the two short versions of the LFPQ could replace it to assess FP in a protocol in which time is limited. The FPQ-S₁₆ appears to be a better solution than the LFPQ-S₂, based on interchangeability analyses.