This Is AuburnElectronic Theses and Dissertations

Testing Accuracy of Body Size Estimation Among Boys




Dawkins, Rachel

Type of Degree



Human Development and Family Studies


The current study utilized this sample to explore a very different question: the associations between boys’ estimates of their body size/shape and potential biasing factors such as the boys’ preferred size, parents’ perceptions and boys’ physical dimensions derived from 3D body scans. The sample used for this study consisted of 119 mother-son pairs, who participated in a study of attitudes and practices related to the purchase of boys clothing. The boys were between the ages of 8 and 14. During data collection, the boys were body scanned using a [TC]² NX 12 Body Scanner and both boys and their mothers answered questions pertaining to body image including the Stunkard silhouettes (Stunkard, et al., 1983), which were used to measure estimated body size. We treated BMI as the operational definition of accuracy and found that BMI explained 41% of the variance in boys’ size estimates. The inclusion of five potential biasing factors and three control variables added an additional 27% to the variance accounted for. Since nearly two- thirds of the total variance explained was linked to our accuracy indicator, it appears that boys in the current study were relatively accurate. However, significant biases were also noted. Boys’ estimates of their own size, reflected in their selection of Stunkard silhouette was biased toward their preferred size, with boys wishing they were bigger selecting larger silhouettes than would be expected from their BMI and with boys wishing they were smaller selecting smaller ones. Boys’ size estimates were also biased in the direction of their mothers’ perceptions of their size. When their mothers viewed them as bigger (smaller), the boys selected a larger (smaller) silhouette than would be expected based only on BMI. We also looked at biasing effects for ii mothers’ size from the perspective of self-reports and sons’ reports. When sons viewed mothers as larger (smaller), they tended to select a larger (smaller) silhouette for themselves, as if identifying with mothers’ size. However, when mothers described themselves as larger, the biasing effect on the boys’ size estimates was in the opposite direction. Controlling for boys’ perceptions of mothers’ size, boys whose mothers said they were larger (smaller) estimated their own size to be smaller (larger) than would be expected on the basis of BMI. The final potential biasing factor examined was a factor score created from actual the body scans which provided objective measures of various parts of the body (chest, hips, waist, bicep, thigh, and distance from floor to waist on the back side). Although the factor score did not uniquely contribute to boys’ size estimates, when we included the three control variables, income, ethnicity (white versus other), and age, in the model the association between the factor score and the boys’ size estimates neared significance even controlling for BMI, suggesting that the a combined measure of objectively assessed body parts could be a better measure of accuracy than BMI. Additional analysis revealed that two body parts (waist and hips) seem to be more meaningful to boys as they estimate their body sizes. Boys for whom these parts were larger (smaller) described themselves as larger (smaller) than would be expected from their BMI alone. Given the mothers’ role in biasing their sons’ size estimates, we looked to see if sons’ discrepancies from the BMI norm in BMI affected mothers’ size estimates for sons. We expected that larger discrepancies in the overweight direction would lead mothers to more seriously underestimate their son’s size, but actually found the opposite. Mothers revealed a significant bias in their estimates only when sons were below the BMI norm and under those circumstances they selected a smaller size than would be expected.