OBJECTIVE: To examine differences in low energy intake reporting between intervention and control groups during a dietary intervention trial. DESIGN: Retrospective data analysis from a subcohort of participants in the Polyp Prevention Trial (PPT), a 4-year, multisite, randomized, controlled dietary intervention trial. Intervention consisted of educational material and counseling sessions supporting a low-fat, high-fiber diet. Baseline and annual demographics, behavioral characteristics, energy intake (EI) based on self-reported 4-day food records, and height and weight of participants were collected at baseline and annually. Basal metabolic rate (BMR) was estimated (using the Schofield equation) to calculate EI/BMR. SUBJECTS: Of the 443 participants (302 male, 141 female) at baseline, 195 (43.3%) were younger than 60 years, and 394 (91%) were white. At Year 4, 383 participants remained: 186 (122 men, 64 women) in the intervention group, and 197 (133 men, 64 women) in the control group. STATISTICAL ANALYSES: Using either paired t tests or analysis of variance, the differences between the means for EI, weight, and EI/BMR were compared at baseline, Year 1, and Year 4 for the participants who remained at Year 4. The Goldberg EI/BMR cutoff value of 1.06 (for plausible EI) identified participants who reported low EI. Linear regression was used to quantify the association of various risk factors to EI/BMR and for multivariate analyses within groups. chi(2) contingency table analysis quantified differences of low energy reporting within groups. RESULTS: At baseline, 46.8% of women and 11.6% of men reported lower than plausible EI. Only men had a significant increase in low energy reporting after randomization. At Year 1, 18.9% of intervention group men reported low EI compared with 9.8% of control group men (P<.05). At Year 4, 23.0% of intervention group men reported low EI compared with 12.8% of control group men (P<.05). CONCLUSIONS/APPLICATIONS: Difference in low EI reporting between intervention and control groups could distort results from dietary intervention trials; interpretation of findings from dietary trials must include this potential bias. Intervention study design should include dietary intake data collection methods that are not subject to such bias (ie, biomarkers and performance criteria) to measure intervention compliance.