Document Type

Book Chapter

Publication Date



International Business and Management



Publication Title

Handbook of Physical Education Research: Role of School Programs, Children's Attitudes and Health Implications


Objectives: To determine whether an association exists between performance on various physical fitness activities (PFs) and body mass index (BMI) in a sample of U. S. middle school students. Are there sex based differences in this association and does an athletic bias exist regarding BMI as a measure of obesity?
Study Design: The Pennsylvania Department of Health instituted the Active Schools Program to encourage daily physical education (PE) in middle schools. This analysis uses the pre-assessment of 9,123 students on four PFs together with information that allowed calculation of BMI and BMI for age- and sex-percentiles (B%). Students were placed in a 16 cell partition based on their sex- and age-adjusted performance on four PFs. Various definitions of athletic and non-athletic are examined based on this partition .
Regressions on the logistic transform of B%, L = Ln(B%/[100-B %]), and linear regressions on BMI were performed using the four PFs together with athletic and non­-athletic dummy variables.
Results: All models place the rank ordering of the effect of increased PF on body mass as mile run then push-ups then curl-ups. Increasing push-ups and curl-ups decreases L and BMI at a decreasing rate but increased mile performance (faster mile times) decreases L and BMI at an increasing rate. Increased back-saver sit and stretch increases L and BMI. Females see greater effect from increased push-ups and curl-ups and males see greater effect from decreased mile run. An asymmetry exists between those defined as athletes and non-athletes. Athletic females have a smaller athletic bias than males . For example, if both have a B% of 85, the best guess is ΔB%Athletic Female = 2.5, 95% CI [0.3, 4.7] and ΔB%Athletic Male = 5.1, [2.8, 7.5]. BMI regressions allow estimating Δweight = ΔW and estimated percentage Δweight = ΔW% associated with being athletic: ΔWAthletic Female = 3.4, [0.9, 5.9] pounds and ΔW% Athletic Female = 3.0%, 95% CI [0.8%, 5.2%], and ΔWAthletic Male = 6.2, [3.8, 8.5] pounds and ΔW% Athletic Male = 5.6%, [3.4%, 7.7%] for students of median height and BMI.
Conclusion: Strong performance on individual PFs does decrease B% and BMI, but doing well on multiple PFs has the reverse effect as long as one of the PFs is the mile run. This provides evidence of an athletic bias in middle school aged students.


Published as:
Erfle, Stephen. "Performance on Physical Activities and Athletic Bias in Body Mass Index in Middle School Students." In Handbook of Physical Education Research: Role of School Programs, Children's Attitudes and Health Implications, edited by Ricky Todaro, 25-49. New York: Nova Science Publishers, Inc., 2014.

This author post-print is made available on Dickinson Scholar with the permission of the publisher. For more information on the published version, visit Nova Science Publisher's Website.

Included in

Education Commons