Effects of social approval bias on self-reported fruit and vegetable consumption: a randomized controlled trial

BACKGROUND: Self-reports of dietary intake in the context of nutrition intervention research can be biased by the tendency of respondents to answer consistent with expected norms (social approval bias). The objective of this study was to assess the potential influence of social approval bias on self-reports of fruit and vegetable intake obtained using both food frequency questionnaire (FFQ) and 24-hour recall methods. METHODS: A randomized blinded trial compared reported fruit and vegetable intake among subjects exposed to a potentially biasing prompt to that from control subjects. Subjects included 163 women residing in Colorado between 35 and 65 years of age who were randomly selected and recruited by telephone to complete what they were told would be a future telephone survey about health. Randomly half of the subjects then received a letter prior to the interview describing this as a study of fruit and vegetable intake. The letter included a brief statement of the benefits of fruits and vegetables, a 5-A-Day sticker, and a 5-a-Day refrigerator magnet. The remainder received the same letter, but describing the study purpose only as a more general nutrition survey, with neither the fruit and vegetable message nor the 5-A-Day materials. Subjects were then interviewed on the telephone within 10 days following the letters using an eight-item FFQ and a limited 24-hour recall to estimate fruit and vegetable intake. All interviewers were blinded to the treatment condition. RESULTS: By the FFQ method, subjects who viewed the potentially biasing prompts reported consuming more fruits and vegetables than did control subjects (5.2 vs. 3.7 servings per day, p < 0.001). By the 24-hour recall method, 61% of the intervention group but only 32% of the control reported eating fruits and vegetables on 3 or more occasions the prior day (p = 0.002). These associations were independent of age, race/ethnicity, education level, self-perceived health status, and time since last medical check-up. CONCLUSION: Self-reports of fruit and vegetable intake using either a food frequency questionnaire or a limited 24-hour recall are both susceptible to substantial social approval bias. Valid assessments of intervention effects in nutritional intervention trials may require objective measures of dietary change.