Psychometric characteristics of the Three-Factor Eating Questionnaire-18 and eating behavior in undergraduate students

Purpose: To assess the psychometric characteristics of the Three-Factor Eating Questionnaire-18 (TFEQ-18) and to estimate the prevalence of cognitive restraint (CRes), uncontrolled eating (UE), and emotional eating (EE) among Brazilian undergraduate students. Methods: A total of 775 students completed TFEQ-18 (62.1% females; total mean age = 20.9 years [SD = 2.8]; females mean age: 20.8 years [SD = 2.8]; males mean age: 21.2 years [SD = 2.7]). Total sample was randomly separated in test/validation samples. An assessment of psychometric characteristics was conducted to each subsample with confirmatory factor analysis considering the indices: chi-square per degree of freedom ratio (χ2/df), Comparative Fit Index (CFI), Tucker–Lewis Index (TLI), and Root Mean Square Error of Approximation (RMSEA). The mean score of each TFEQ-18 factor was estimated. The prevalence of CRes, UE, and EE was calculated with 95% confidence intervals (95%CI) and compared by sex and weight status. Results: The psychometric characteristics of the TFEQ-18 were adequate in both samples (test/validation: χ2/df = 3.58/2.86; CFI = 0.938/0.958; TLI = 0.928/0.951; RMSEA = 0.081/0.069). The prevalence of students with moderate to exacerbated scores of CRes, UE, and EE was 34.4% (95%CI 31.1–37.7%), 35.6% (95%CI 32.2–39.0%), and 43.1% (95%CI 39.6–46.6%), respectively. There were no statistically significant differences in prevalence according to sex or weight status. Conclusions: The TFEQ-18 presented adequate psychometric characteristics and indicated a high prevalence of CRes, UE, and EE among the students. This research emphasizes the importance of investigating aspects of eating behavior to best direct clinical and educational strategies for reducing the risk of adopting unhealthy eating behaviors among undergraduate students. Level of evidence: Level V, cross-sectional descriptive study.