Document Type

Thesis - Open Access

Award Date


Degree Name

Master of Science (MS)

Department / School

Health and Nutritional Sciences


The purpose of this investigation was to determine an effective and simple method of estimating body density of college-age women (n=51), ages 18 to 26 years, using age and selected circumference measurements (thorax, waist, abdomen, hip, gluteal fold and thigh). Two subproblems were also investigated. The first subproblem was to determine if the sum of the selected circumferences was a better predictor of body density than when each circumference was used individually as a predictor. The second subproblem was to determine if body density could best be predicted by a linear or curvilinear relationship. Statistical analysis of the data included descriptive statistics for the subjects and multiple regression analyses to determine the best predictors of body density from circumferences. lntraclass correlations were computed to determine test-retest reliability of the data collected. A regression equation using four circumferences (thorax, abdomen, gluteal fold and thigh) was derived to predict body density (R 2=.65). Cross-validation was employed to determine the validity of the regression equation. The four independent circumferences (thorax, abdomen, gluteal fold and thigh) were significant (p<.05) in their prediction of body density and the inclusion of five or all six did not add significantly to the prediction of body density (p>.05). The sums of the circumferences, the square of these sums, the square of the independent circumferences and age were not significant predictors of body density (p>.05). The data of the validation and cross-validation samples were highly correlated with the actual body density values for each respective sample (r=.70 and r=. 73, respectively).

Library of Congress Subject Headings

Body composition

Women college students

South Dakota State University Theses



Number of Pages



South Dakota State University