Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School

Physical Education

First Advisor

Matthew Vukovich


With the use of the meta-analytical approach, the purpose of this study was summarize the current state of knowledge examining the validity of five body composition methods; skinfolds (SKF), dual energy x-ray absorptiometry (DXA), air displacement plethysmography (ADP), bioelectrical impedance analysis (BIA) and the 3 compartment method (3-C) in relation to hydrostatic weighing (HW) across age and gender differences. For the interpretation of the meta-analysis, a small effect size (ES) was :'.S0.20, a moderate ES 􀃆0.50, and large ES was 􀃆0.80. For the purpose of the current study, the desired ES was zero indicating an accurate estimation of percent body fat (%BF) to that from HW. Mean ES were calculated for the total group (males+ females), males and females for each of the age groups(< 40 years old, 40-59 years old,􀃆 60 years old), and mixed-ages with each of the body composition methods (SKF, DXA, ADP, BIA and 3-C). Thirty six studies met the inclusion criteria for the analysis and were used in the current study. From these studies, 203 ES (64 group, 76 female, 73 male) were calculated representing 2,373 subjects (1449 women, 924 men). The average sample size was 51 ± 44 (range: 7 to 220) subjects per study. All methods underestimated %BF in relation to HW. Overall, the 3-C method yielded the smallest ES of (-0.05 ± 0.83 ). ADP provided the second smallest mean ES of (-0.06 ± 0.41). The methods DXA and BIA followed with mean ES of (-0.07 ± 0.68), (-0.18 ± 0.56), respectively. SKF had the largest ES of (- 1.11 ± l .9 4) indicating a severe underestimation of %BF when compared to % BF from HW. The accuracy of the experimental methods (DXA, ADP, 3 -C, SKF, and BIA) in comparison to HW was better for females. When the ES was based solely on gender, females generated a small effect with a mean ES (-0.24 ± 1.33 ) and that of the males (- 0.39 ± 1.09) yielded a moderate effect. The largest ES values were found in the 60 year old and over age group. The results indicated that aging does have an effect on body composition, which is evident by the large mean ES in the 􀃤 60 year old populations. %BF was underestimated within all age groups except for mixed age group (0.14 ± 0.44; range: -0.42 to 1.19) and mixed age females (0.15 ± 0.67; range: -1.68 to 2.02). In conclusion, the meta-analysis review of the included studies suggested that aging has a large effect on the accuracy of body composition measurements. The mean ES for females (-0.24 ± 1.3 3: -6.66 to 2.02) generated a smaller effect than that derived from males (-0.3 9 ± 1.09; -6.27 to 2.03). The accuracy of the mean %BF estimation was very closely related between the 3 -C method (-0.05 ± 0.83 ), ADP (-0.06 ± 0.41), and DXA (- 0.07 ± 0.68) based on the mean ES calculation in relation to the mean %BF from HW.

Library of Congress Subject Headings

Body composition -- Measurement.
Aging -- Physiological aspects.
Body composition -- Sex differences.
Anthropometry -- Methodology.


South Dakota State University



Rights Statement

In Copyright