Reliability and validity of sports accelerometers during static and dynamic testing
- Publication Type:
- Journal Article
- International Journal of Sports Physiology and Performance, 2015, 10 (1), pp. 106 - 111
- Issue Date:
© 2015 Human Kinetics, Inc. Purpose: To investigate the validity and reliability of accelerometry of the SPI-ProX II dual data logger (GPSports, Canberra, Australia). Methods: Controlled laboratory assessments determined the accuracy and reproducibility of raw accelerometer data. Intra- and interdevice reliability assessed the ability of the SPI-ProX II accelerometers to repeatedly measure peak gravitational accelerations (g) during impact-based testing. Static and dynamic validity testing assessed the accuracy of SPI-ProX II accelerometers against a criterion-referenced accelerometer. Dynamic validity was assessed over a range of frequencies from 5 to 15 Hz. Results: Intradevice reliability found no differences (P < .05) between 4 SPI-ProX II accelerometers, with a low coefficient of variation (1.87-2.21%). SPI-ProX II accelerometers demonstrated small to medium effect-size (ES) differences (0.10-0.44) between groups and excellent interdevice reliability, with no difference found between units (F = 0.826, P = .484). Validity testing revealed significant differences between devices (P = .001), with high percentage differences (27.5-30.5%) and a large ES (>3.44). Conclusions: SPI-ProX II accelerometers demonstrated excellent intra- and interaccelerometer reliability. However, static and dynamic validity were poor, and caution is recommended when measuring the absolute magnitude of acceleration, particularly for high-frequency movements. Regular assessment of individual devices is advised, particularly for mechanical damage and signal-drift errors. It is recommended that guidelines be provided by the manufacturer on measuring shifts in the base accelerometer signal, including time frames for assessing accelerometer axis, magnitude of errors, and calibration of accelerometers from a stable reference point.
Please use this identifier to cite or link to this item: