Accuracy of self-reported heights and weights in a predominately low-income, diverse population living in the USA

Objectives: This study explored the accuracy of self-reported heights and weights and factors associated with self-reported bias in a diverse American sample. Methods: Demographic, self-reported, and measured height and weight data from different studies with the same PI were compiled into one SPSS file and analyzed with paired t-tests to detect differences between self-reported and actual values. Kruskal-Wallis tests followed by pairwise t-tests detected differences among age, ethnicity, sex, income, and education. Stepwise regression analyses were done using anthropometric differences as the dependent variable and age category, sex, and ethnicity as independent variables to explore which variable was most predictive of anthropometric differences. Results: Individuals over-reported height and under-reported weight leading to an under-calculated BMI from self-reported height and weight by 0.6-1 kg/m². These under-calculations of BMI led to misclassifications of obesity by 3, 6, 8, and 4% for African American, Euro-American, Native American women, and total women, and by 5, 6, 8, and 8% by African American, Euro-American, Native American men, and total men. Older individuals and males over-reported height more than younger individuals and females. African American females over-reported height to a lesser extent than other ethnicities. Asian males over-reported height to a lesser extent than other ethnicities. Conclusions: Self-reported heights and weights lead to invalid results. Most individuals over-report height and under-report weight, resulting in an inaccurate underweight and obesity prevalence. Being misclassified into the incorrect BMI category could result in inappropriate healthcare treatment. Age, ethnicity, and sex appear to influence the misreporting of height and weight.