Date of Award

January 2018

Degree Type

Open Access Thesis

Degree Name

Master of Science (MS)


Exercise and Sport Science

First Advisor

Michael T. Lane

Second Advisor

Aaron D. Sciascia

Third Advisor

James M. Larkin


Intro: Body composition is an important metric to evaluate overall health. Having reliable body composition testing methods are critical to ensure that an individual is receiving correct data in which to base health, nutrition, and lifestyle decisions. With technological improvements, there are more reliable, high-quality testing devices for body composition than in the past. However, there remains questions on the validity of the devices to properly measure body composition. There is a need for a more accurate, simplistic testing methodology.

Purpose: The purpose of this paper was to examine the use of dual energy x-ray absorptiometry (DEXA) based formulas to evaluate body volume (BV). The second purpose was to examine the validity of using predicted body volume measurements in four-compartment body composition models.

Methods: Subjects were tested on three devices designed for body composition metrics; DEXA for lean body mass, bone mineral content, and fat mass, BodPod for body volume, and a bioimpedance spectroscopy (BIS) device using dual electrode tabs (SFB7) for total body water. The measured metrics were used to compute two DEXA based predicted BV equations, Wilson, et al. (2012) and Smith-Ryan, et al. (2017). The results were then compared to measured BodPod BV. Second, this study calculated two different four-compartment model formulas, Withers, et al. (1998) and Wang, et al. (2005), using a DEXA calculated body volumes. These models were then compared to a four-compartment model using BodPod measured body volume.

Results: 90 healthy adults (50 females and 40 males), aged 18 to 66 years (median 23 years), BMI 18 to 34 (median 25), weight 45 to 115 kg (mean 73.64 ± SD 14.35 kg), height 150 to 191 cm (171.07 ± 9.98 cm), BodPod data was collected and used as a standard for comparison to the DEXA based body volume formulas; (Wilson et al. 2012) and (Smith-Ryan, et al., 2017). BodPod measured BV mean of 70.36 ± SD 13.85 L, Wilson, et al. (2012) 70.88 ± 13.54 L, and Smith-Ryan, et al. (2017) 70.02 ± 14.23 L. ANOVA yielded no statistical difference between the three groups (p=.915). Among the Withers, et al. (1998) formula, paired t-test of BodPod/Withers yielded a significant difference lower than Wilson/Withers (z = -6.178, p ≤ .001) and Smith-Ryan/Withers yielded a significant difference lower than BodPod/Withers (z = -5.052, p ≤ .001). Among Wang, et al. (2005) formula, Wilcoxon signed-rank test of BodPod/Wang yielded a significant difference lower than Wilson/Wang (z = -5.816, p ≤ .001) and Smith-Ryan/Wang a significant difference lower than BodPod/Wang (z = -4.690, p ≤ .001). These significant differences indicate the predicted BV equations are not equivalent to using measured BV and yielded inaccurate results

Conclusion: The use of DEXA based BV formulas is a viable replacement for other BV testing methodologies for use in four-compartment testing models. However, both the Wilson, et al. (2012) and Smith-Ryan, et al. (2017) formulas failed to be a viable replacement for measured BodPod values. A new formula, A new formula resulted from this study. Further studies may be needed before a formula can be utilized in four-compartment models.