Can self-reported height and weight be relied upon?

Methods Volunteers were instructed how to measure their weight, height and waist circumference (WC). Self-reported values were compared with direct measurements. Accuracy was assessed using simple (self-report - actual) and percentage difference [(self-report - actual)/(actual measurement)]. The distribution of differences (in weight, height and BMI) across age and BMI classes was calculated plus Pearson (parametric) and Spearman (non-parametric) coefficients of correlation, to assess relation of differences (simple and percentage) with actual values. For percentage differences, classes were created to explore differences in mean values of actual measurements across various difference classes, using analysis of variance. Results Eight hundred and fifty-seven workers took part; 585 (68%) provided all requested data. 'Statistical analysis showed that men and the whole group underestimated their BMI due to overestimating their height and underestimating body weight'. Similar trends were seen in females, especially the centrally obese ones (WC > 80 cm), but women as a group were more accurate than men in anthropometric self-reports. Males > 40 years of age underestimated their weight. Conclusions This study showed that the differences between actual and self-reported values depend on the actual values and self-reported anthropometric measurements cannot be relied upon, at least in males. Females seem to provide more accurate reports than men and we could consider their measurements reliable, although a further study with a larger number of female participants would be needed.