Secondary Logo

Journal Logo

Short Research Report

Reliability and Validity of a Medicine Ball–Contained Accelerometer for Measuring Upper-Body Neuromuscular Performance

Roe, Gregory1,2; Shaw, William1; Darrall-Jones, Joshua1,2; Phibbs, Padraic J.1,2; Read, Dale1,2; Weakley, Jonathon J.1,2; Till, Kevin1,2; Jones, Ben1,2,3

Author Information
Journal of Strength and Conditioning Research: July 2018 - Volume 32 - Issue 7 - p 1915-1918
doi: 10.1519/JSC.0000000000002470
  • Free



Professional sport relies on many different technologies to quantify and monitor training outputs and provide feedback to coaches and athletes (6). Examples include the use of global positioning systems for monitoring distances, velocities and accelerations (2), timing gates to monitor linear speed (7), and linear position transducers to provide force, velocity, and power of barbell lifts (9). However, practitioners regularly use new technology without understanding the validity and reliability of the devices in use (6).

Recently a new technology allowing for measurement of velocity from medicine ball throws has been made available to practitioners. This may offer a diagnostic tool for the assessment of upper-body neuromuscular performance, which is important for monitoring fatigue (8), measuring and tracking performance (10), and use in upper-limb rehabilitation programs (1). However, this technology is yet to be validated, or assessed for reliability. Therefore, the aim of this study was to examine the between-day reliability and validity of a medicine ball–contained accelerometer (MBA) for assessing throwing velocity during an upper-body power exercise.


Experimental Approach to the Problem

Testing was undertaken during the preseason period, throughout which subjects typically engaged in 8 individual training sessions across 5 days per week including 2 upper-body and 2 lower-body resistance training sessions, along with rugby skill training and conditioning. Subjects were excluded if they had a current upper-limb injury or history of upper-limb injury that prevented them from performing the testing procedure.


Ten male professional rugby union players (± SD age 19.7 ± 1.1 years, body mass 98.3 ± 13.2 kg, height 186.2 ± 7.6 cm) were recruited from a professional rugby union club. All players were in their second year of their professional contracts. Ethics approval was granted by Leeds Beckett University’s Ethics Board (Leeds, West Yorkshire, UK), and informed written consent was acquired from all participants.


Reliability data of the 8-kg MBA (Ballistic Ball; Assess2Perform, Colorado, USA) were collected on 2 separate days (7 days apart) during a week at the end of the playing season. Testing was undertaken at the same time of day to ensure that diurnal variation did not affect performance. All participants were provided with regular dietary advice from the club's nutritionist, but dietary intake was not controlled for in this study. Subjects had approximately 48-hour rest before each testing session. Subjects were in a supine position on the floor, on their back with hips and knees bent to approximately 45° and feet flat on the floor. The MBA was held with elbows extended and hands supporting the medicine ball from underneath. Elbow width was at the discretion of the participant. Subjects were instructed to throw the MBA into the air as hard as possible without their feet, buttocks, back, or shoulders leaving the floor. Subjects performed 2 sets of 3 throws. Intraset rest was approximately 20 seconds while interset rest was 2 minutes. The highest score achieved was used in the final analysis.

The validity of the MBA was assessed against a criterion measure (Optioelectronic system; Qualisys—Qqus system, software version 2.14) in a University biomechanics laboratory, where 3 subjects performed 25 throws each. Five (18 mm) reflective markers were attached to the MBA (superior, inferior poles, and a 3-marker cluster on the anterior face) and were tracked during each throw using 8 Qualisys cameras sampling at 200 Hz. The kinematic data were processed and filtered in Qualisys, before data were transferred to Visual 3D (version 6). In Visual 3D, the 5 tracking markers were used to create 3 virtual markers within the MBA center (offset 50% between superior and inferior markers). Anterior split (offset 50% between 2 cluster tracking markers, on a plane with the med ball center) and anterior y (offset 50% between the top cluster marker and the anterior split marker). A virtual segment was then created for the MBA using these virtual markers. The advantage of creating virtual markers to track the object kinematics is to minimize any error that may come from one marker moving or shifting on the surface of the MBA. A pipeline (segment velocity) was used to extract the velocity signal for segment “med ball” against the “laboratory” as a reference. The peak velocity achieved during the upward phase of the throw was used for analysis against the MBA's output.

Statistical Analyses

The between-day reliability statistic of typical error (TE) was calculated as;where Sdiff is the SD of the difference score and converted to a coefficient of variation (CV; TE expressed as a percentage) for all tests using a Microsoft Excel spreadsheet (4). The standardized TE was rated as trivial (<0.19), small (0.2–0.59), medium (0.6–1.19), or large (1.2–1.99) (4). The smallest worthwhile change was calculated as 0.2 multiplied by the between-subject SD and calculated as a percentage of the mean (5).

The agreement between the criterion measure (Qualisys) and the practical measure (MBA) was assessed using an excel spreadsheet (4) designed to calculate the mean bias () standard error of the estimate (STEXY function) and Pearson correlation coefficient, all with 90% confidence limits (4). The mean bias and standard error were standardized using the SD of the criterion to allow for qualitative rating. The standardized mean bias was rated as trivial (<0.19), small (0.2–0.59), medium (0.6–1.19), or large (1.2–1.99) (4). The standardized standard error was rated as trivial (<0.1), small (0.1–0.29), moderate (0.3–0.59), or large (>0.59) (4). The magnitude of correlation was rated as trivial (<0.1), small (0.1–0.29), moderate (0.3–0.49), large (0.5–0.69), very large (0.7–0.89), or nearly perfect (0.9–0.99) (4).


The between-day CV of the test was 2.2% (2.0–4.6%) (raw; 0.11 m·s−1 [0.10–0.23 m·s−1]) and was rated as small. The smallest worthwhile change was 1.5% (raw; 0.07 m·s−1). When the medicine ball was compared with the criterion measure, mean bias was moderate (7.9% [6.6–9.2%], raw; 0.39 m·s−1 [0.33–0.45 m·s−1]), TE of the estimate was moderate (4.9% [4.2–5.7%], raw; 0.24 m·s−1 [0.21–0.28 m·s−1]), whereas the correlation was almost perfect (r = 0.91 [0.87–0.94]). The regression plot is presented in Figure 1.

Figure 1.:
Regression plot for agreement between the criterion measure (Qualisys) and the practical measure (medicine ball).

The regression equation to estimate the criterion measure (Y) from the practical measure (X) is


This study examined the reliability and validity of an MBA for assessing throwing velocity during an upper-body power exercise. Between-day reliability analysis revealed a small TE of 2.8% (2.0–4.6). Practitioners wishing to track upper-body neuromuscular performance over time using an MBA can use this statistic to assess whether a meaningful change has occurred. Hopkins (3) proposed a method whereby the change score of an individual (±error bars representing the CV) is graphed with an important threshold (e.g., the smallest worthwhile change). A change is “clear” when the error bars lie outside of the important threshold and “unclear” when the error bars cross the important threshold (3). It must be pointed out that the subjects in this study were trained athletes, and as such, the between-day error may be different for other populations of different training backgrounds and athletic abilities (Figure 2).

Figure 2.:
An example of change in the performance of 2 athletes. Data are percentage change in an individual's performance (±CV error bars) with gray area representing the smallest worthwhile change. Adapted from Hopkins (3).

In addition, correlation analysis demonstrated an almost perfect relationship between the MBA and criterion measure, indicating excellent validity. However, the findings show that the MBA overestimated throwing velocity by 7.9% and demonstrated a moderate standard error (4.9%) when compared with the criterion measure. Therefore, practitioners using an MBA to measure upper-body neuromuscular performance must take into account the error and bias when making inferences about such performance. Practitioners wishing to estimate the criterion from the MBA may do so using the equation provided while appreciating the associated error with this.

Practical Applications

Practitioners using an MBA to assess neuromuscular performance of the upper body must take into account the overestimation and error associated with such assessment with respect to a criterion measure. However, as the error associated with between-day testing was small and testing is easy to implement in applied practice, an MBA may provide a useful tool for monitoring upper-body neuromuscular performance over time in trained athletes.

This study investigated the reliability and validity of an MBA for measuring throwing velocity during a throwing task. Although it exhibited an almost perfect relationship with a criterion measure, it moderately overestimated throwing velocity and had a moderate standard error. Despite this, the between-day assessment error was only small, making it a potentially useful test to monitor changes in upper-body neuromuscular performance over time in trained athletes.


1. Borms D, Maenhout A, Cools AM. Upper quadrant field tests and isokinetic upper limb strength in overhead athletes. J Athl Train 51: 789–796, 2016.
2. Cummins C, Orr R, O'Connor H, West C. Global positioning systems (GPS) and microtechnology sensors in team sports: A systematic review. Sports Med 43: 1025–1042, 2013.
3. Hopkins WG. How to interpret changes in an athletic performance test. Sportscience 8: 1–7, 2004.
4. Hopkins WG. Spreadsheets for analysis of validity and reliability. Sportscience 19: 36–42, 2015.
5. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Med Sci Sports Exerc 41: 3–13, 2009.
6. Malone JJ, Lovell R, Varley MC, Coutts AJ. Unpacking the black box: applications and considerations for using GPS devices in sport. Int J Sports Physiol Perform 12: S218–S226, 2017.
7. Roe G, Darrall-Jones J, Black C, Shaw W, Till K, Jones B. Validity of 10 HZ GPS and timing gates for assessing maximum velocity in professional rugby union players. Int J Sports Physiol Perform12: 836–839, 2017.
8. Roe G, Darrall-Jones J, Till K, Phibbs P, Read D, Weakley J, Jones B. Between-days reliability and sensitivity of common fatigue measures in rugby players. Int J Sports Physiol Perform 11: 581–586, 2016.
9. Scott BR, Duthie GM, Thornton HR, Dascombe BJ. Training monitoring for resistance exercise: Theory and applications. Sports Med 46: 687–698, 2016.
10. Wang R, Hoffman JR, Sadres E, Bartolomei S, Muddle TWD, Fukuda DH, Stout JR. Evaluating upper-body strength and power from a single test: The ballistic push-up. J Strength Cond Res 31: 1338–1345, 2017.

testing; monitoring; fatigue

© 2018 National Strength and Conditioning Association