A study to examine the validity and reliability of a selected group of physical fitness tests within a primary school setting

Buck, Robin

A study to examine the validity and reliability of a selected group of physical fitness tests within a primary school setting

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Human Biology

Faculty

Faculty of Health Sciences

Abstract

Purpose: This study represents an effort to provide an evidence-based approach to selecting field-based fitness measures for children and youth for inclusion in a test battery to be used within the school setting. The approach used by the author considers the following criteria to determine which youth fitness test items would be most appropriate for inclusion: (1) the strength of the fitness tests' association with health markers in youth, (2) the fitness tests integrity (validity and reliability), (3) the relative feasibility of the test administration in the field and (4) the measurement error of the fitness test. This study aims to identify valid and reliable physical fitness tests with low measurement errors for use in South African schools Method: A literature review was performed to determine the predictive and criterion validity of physical fitness tests in children. Tests with strong predictive evidence and good criterion validity were selected for further reliability testing. A group of 290 male and female children aged 6 to 13 years performed a physical fitness testing protocol three times within a 7-day period. The typical error of measurement (TE), coefficient of variation (CVTE) and smallest worthwhile change (SWC) were calculated for each sex and age group (grade). Changes in measurement were categorised as small, medium or large based on effect sizes. Results: The 20m shuttle run (SRT), hand grip strength (HGS), standing long jump (SLJ), body mass index (BMI), and waist circumference (WC) tests demonstrate strong predictive evidence and good criterion validity. Typical error for the 20mSRT was 0.9 stages. Small changes in 20mSRT (0.4 stages) are challenging for the test to detect. Medium (1.0 stages) and large (1.9 stages) changes in the 20mSRT are more interpretable as they exceed the “noise” (TE) of the measurement. Small (0.6 kg/m2), medium (1.6 kg/m2) and large (2.5 kg/m2) changes in BMI are all larger than the TE (0.2 kg/m2). The BMI test has a ‘good' ability to detect small, medium and large changes. TE for HGS was 1.3 kg. Small changes in HGS (1.3kg) are difficult for the HGS test to detect. Medium (3.3kg) and large (5.3kg) changes in HGS are more interpretable as they exceed the “noise” (TE) of the measurement. The WC test has a ‘good' ability to detect small (1.5 cm), medium (3.8 cm) and large (6.1 cm) changes as they are larger than the TE (1.5 cm). TE for SLJ was 8.8 cm. Small changes in SLJ (5.4 cm) are difficult for the SLJ test to detect. Medium (13.5 cm) and large (21.5 cm) changes in SLJ are more interpretable as they exceed the “noise” (TE) of the measurement. Conclusion: The field-based physical fitness tests that demonstrated strong predictive validity and good criterion validity from the literature and low measurement error from the reliability data are HGS, SLJ, BMI, WC and 20mSRT. These fitness tests could be included in a battery of tests for implementation within the school setting. Researchers and practitioners can use the SWC and TE as guidelines to set targets when determining the extent to which performance changes in these fitness tests are practically significant.