Pro-agility unpacked: Variability,comparability and diagnostic value

Abstract

Change of direction (COD) is an important component of athlete performance and measuring and comparing athletes is an integral aspect of strength and conditioning practice. This article aimed to determine pro-agility shuttle utility, by quantifying variability and normative values for different sports, skill-levels and positions. Limitations of the pro-agility shuttle are identified, as are future research directions. A total of 67 studies were included for review. Pro-agility shuttle reliability was reported in 10 studies across 6 sports; however, comprehensive reliability statistics were absent in most papers. Additionally, only reliability of total-time from stopwatch and timing lights were reported. Data of 32,891 subjects in 12 sports (American football, basketball, cricket, general athletes, hockey, lacrosse, recreational athletes, resistance-trained athletes, rugby, soccer, swimming, and tennis) were extracted and aggregated, establishing sport, skill-level (elite, sub-elite, and novice) and positional normative values, where practical. Elite athletes showed the fastest performance times, whereas sub-elite and novice athletes showed similar spreads in performance, suggesting similar athletic capabilities. In conclusion, the pro-agility shuttle currently has limited diagnostic value and the variability of smaller performance sub-components within pro-agility shuttle should be examined. Furthermore, the value of other technologies such as smart phone, inertial sensor or radar should be investigated.

Keywords

Athletic performance change of direction shuttle test

Introduction

Change of direction (COD) ability is one of the key determinants of successful athletic performance in both field and court-sport athletes.^1,2 The ability to change direction efficiently provides an indication of an athlete's underlying physiological capabilities, such as multi-directional reactive strength and anaerobic power^3,4 and can be vital during key instances within a game which influence winning or losing (i.e. line breaks in rugby.⁵⁾ Therefore, the assessment of COD ability has become important in many sporting codes.^6–9 One of the most common tests to measure COD ability is the pro-agility shuttle test.¹⁰ Providing insight into acceleration, deceleration and COD, the pro-agility test is comprised of two 180° changes of direction over a total of 18.28 m (20 yards).^11–14

The pro-agility has been used to determine team sport athletes’ COD performance in sports such as basketball,⁶ cricket,¹⁵ ice hockey,¹⁶ lacrosse,^17,18 and rugby.¹⁹ In some studies, the pro-agility test has also been used to distinguish positional differences in athletic qualities between sporting codes, for example, soccer and lacrosse.^20,21 The pro-agility has also been used to benchmark individual athlete ability against other athletes for talent identification and recruitment within the American football combine.^6,15,22,23 This is due to the sharing of similar game-related movements, such as performing short accelerated sprints with rapid deceleration and high degrees of COD when transitioning from attacking to defending.^18,24,25 In order for practitioners to benefit from performance assessment, they need to be knowledgeable as to whether the tests they use are reliable and accurately assess athlete performance. To be assured in the tests used reliably reflect athlete performance, tests must be deemed consistent across multiple testing occasions.²⁶ However, despite its widespread use there is little understanding as to the validity and reliability of the pro-agility test and therefore is a focus of this review.

One of the goals of collecting sport performance data is to gather normative information that provides insight into the representation and spread of typical performance, relative to the sport, performance level or individual.²⁷ By providing normative data, it enables practitioners to make appropriate comparisons between player performance to those of other groups (i.e. player positions and skill level). Additionally, practitioners can use the established values to advise practical and feasible goal setting.¹⁰ Therefore, establishment of pro-agility normative values is critical for the identification, monitoring and development of athlete performance and provides a focus of this review.

Existing reviews in the area of COD have been conducted, providing summaries of the different testing and training qualities of COD.^12,24,28 Unfortunately, they fail to discuss the utility of specific COD tests, such as the pro-agility, with previous work only focusing on the broader literature. Despite the use of pro-agility as a performance assessment, little thought or critique has been given to the utility of the test Given the preceding information, the aims of this review were:1) to quantify the variability associated with the pro-agility test; 2) establish normative data; and, 3) identify current limitations and future research directions. By taking such an approach, users of the pro-agility will have a better appreciation of the utility and limitations of this test

Methods

Study design

A systematic review was conducted in accordance with the Preferred Reporting Item for Systematic Reviews and Meta-analysis (PRISMA) guidelines²⁹ and was defined by the Population, Intervention, Comparison, and Study Design (PICOS) model. This review aimed to determine the utility of the pro-agility, through quantifying variability and reliability, establish normative performance values of different sports, skill levels and player positions, and identify limitations and areas of future research for the pro-agility shuttle.

Search strategies

A systematic search of four electronic data bases (SPORTDiscus, PubMed, ScienceDirect, and OVID journals) was undertaken between January and May 2020 to identify original research articles published from the earliest available records up to and including May 2020. The keywords ‘pro-agility’, OR ‘20 yard shuttle’, OR ‘5-10-5 shuttle’ were used in Boolean logic for query. The reference sections of the selected studies were also examined for identification of other applicable studies.

Study and screening selection

To be included studies needed to meet the following criteria; 1) measurement of the pro-agility; 2) a detailed description of the methods and technology used; 3) state the number of subjects and descriptive statistics of characteristics (age, height, and body mass); and 4) state the sport and skill level of the subjects. Studies that included the pro-agility were initially included in the first screening phase (n = 282). Additionally, studies must have been written in English. The initial search of the literature was conducted by the author. A second search of the literature, using the same parameters, was conducted by a second author, resulting in 4.47% fewer articles than the initial search. The discrepancy in number of articles found was solved via consensus to use the results of the initial search. All potentially applicable articles were included in a database, where afterwards they were read in full by one author. Specific aspects of the articles were then discussed to establish and ensure a consensus was reached with the other authors, regarding the article(s) inclusion, where relevant. To determine the number of eligible studies a three-stage screening process was implemented; 1) Removal of duplicate studies (n = 127); 2) Screening of article title and abstract. Studies that were deemed to be ‘out of scope’ (did not contain pro-agility time data) were excluded (n = 65); and, 3) Exclusion of studies that did not meet the inclusion criteria after screening the full text (n = 31). An additional eight eligible articles were included after reference checks, resulting in 67 studies included for analysis in this review. The selected studies comprised of 38 acute studies, 16 intervention studies, and 13 archival data (stored data) studies (Figure 1).

Figure 1.

Flow diagram of study selection process.

Data extraction

One author extracted the data using a custom designed standardised excel database (version 16.0, Microsoft, Redmond, WA, USA). A second author validated a cross-section of these ratings for quality assessment control. Consensus was attained for the quality score of each article. In this case, an additional reviewer was not required for resolution of scoring issues. General study information (i.e. author, year), subject characteristics (i.e. sample size, gender, age, body mass, height, sport, performance level), type of study (i.e. acute, training, archival), methods of assessment (i.e. testing equipment, surface), and primary outcome measures (i.e. means and standard deviations of velocity or time) were extracted. Descriptive information relating to the sport and performance level were used to categorise each of the participants. Subjects who were not identified with a sport were grouped as “general athletes”. In the case that data for different sports was reported in an article, performance results were categorised in the appropriate sport for analysis. There is a wide array of definitions for elite, sub-elite, and novice athletes.³⁰ Therefore, in order to clearly differentiate between groups, skill level was defined by the competitive level of athletes in their respective sport in this review. Elite athletes were classified to be competing at National Collegiate Athletic Association (NCAA) divisions 1 or 2, national, international, or professional competitive levels. Sub-elite athletes being those of NCAA division 3 athletes,^17,30 regional level club athletes, or undrafted athletes. With novice athletes being classed high school or recreational athletes.

Study quality assessment

To assess the standard of the included studies, a quality scale designed to evaluate research conducted in athletic-based environments, was implemented.³¹ This scale was modified and utilised based upon the scale created by Brughelli, Cronin,³² using a combination of items from the Cochrane, Delphi, and PEDRO, as shown in Table 1. The quality of each study was evaluated on 10 items (2 points per item): inclusion criteria stated, subject assignment, intervention description, control groups, dependent variables definition, assessment methods, study duration, statistics, results section, and conclusions. The total quality score for each study ranged between 0 to 20, where a score of 0 = clearly no; 1 = yes, not detailed; and 2 = yes, clearly detailed.

Table 1.

Study quality score.³²

Number	Item	Score
1	Inclusion criteria stated	0-2
2	Subjects assigned appropriately	0-2
3	Intervention described	0-2
4	Control group	0-2
5	Dependant variable defined	0-2
6	Assessments practical	0-2
7	Duration	0-2
8	Statistics appropriate	0-2
9	Results detailed	0-2
10	Conclusions insightful	0-2

Data analysis

IBM SPSS statistical software package (version 25.0; IBM Corporation, New York, USA) was used to analyse the data from the studies collected. Data was reported using mean and standard deviation (±SD). Subjects were grouped by sport, skill level (novice, sub-elite, and elite), and position where appropriate. The level of relative reliability intraclass correlation coefficient (ICC) was described as: ≤0.50 poor, 0.50–0.75 moderate, 0.75–0.90 good, and >0.90 excellent reliability.³³ Absolute reliability coefficient of variation (CV%) was deemed acceptable at a level ≤10%.^34,35

Results

Overview of studies

A summary of the included studies for this review can be found in Table 2. Data was gathered from 11 different sports in 68 population samples over 67 studies as summarised in Table 3 and herewith; most sports were found in the sub-elite and novice skill level, however, there appears to be an ample spread of data through the three skill levels. The 67 studies comprised a sample size of 32,891 subjects of different skill levels (elite: 6,631, sub-elite: 382, novice: 25,878), with an average sample size of 472.79 ± 1193 (range 12–4603) for elite, 27.3 ± 19.7 (range 12–84) for sub-elite, and 924.21 ± 2505 (range 11–9203) for novice subjects. The most common sport was American football (Table 3), with the least common sports being basketball, cricket, ice hockey, resistance-trained athletes, and tennis.

Table 2.

Description of the studies in the review.

Author	Sport	Age	Skill level	Number of Subjects	Mean ± SD	Quality Score
Corey et al. 2020³⁶	American football	20 + years	Elite	1263	4.39 ± 0.18	14
Daichi et al. 2017³⁷	American football	26.2 ± 3.9	Elite	245	4.51 ± 0.19	18
Davis et al. 2004³⁸	American football	Not reported	Elite	46	4.17 ± 0.27	12
Gains et al. 2010³⁹	American football	18.8 ± 0.4	Elite	24	4.63 ± 0.06	17
Mann et al. 2016⁴⁰	American football	20.5 ± 1.2	Elite	64	4.47 ± 0.29	17
Nesser et al. 2008⁴¹	American football	Not reported	Elite	29	4.50 ± 0.30	15
Nimphius et al. 2013⁴²	American football	20 ± 2.0	Elite	66	4.53 ± 0.33	17
Nuzzo, 2015²³	American football	Not reported	Elite	4603	4.39 ± 0.19	13
Robbins, 2011⁴³	American football	Not reported	Elite	1136	4.43 ± 0.30	14
Robbins, et al. 2013⁴⁴	American football	Not reported	Elite	1712	4.66 ± 0.02	14
Stodden and Galitski, 2010⁴⁵	American football	Not reported	Elite	84	4.17 ± 0.09	17
Hoffman et al. 2011⁴⁶	American football	Not reported	Sub-elite	289	4.33 ± 0.19	18
Leutzinger et al. 2018⁴⁷	American football	Not reported	Sub-elite	7160	4.67 ± 0.18	13
Lockie et al. 2012⁴⁸	American football	16.53 ± 0.77	Sub-elite	36	4.85 ± 0.53	15
Newsom and Probst, 2016⁴⁹	American football	Not reported	Sub-elite	25	4.58 ± 0.16	12
Yuasa et al. 2018⁵⁰	American football	19.9 ± 0.9	Sub-elite	17	4.42 ± 0.13	18
Caswell et al. 2016⁵¹	American football	12.4 ± 1.32	Novice	819	5.60 ± 0.23	17
Collins et al. 2018⁵²	American football	15.93 ± 0.96	Novice	15	5.03 ± 0.21	18
Dupler et al. 2010⁵³	American football	9-12 grade	Novice	2327	4.63 ± 0.06	14
Ghigiarelli, 2011⁵⁴	American football	Not reported	Novice	161	4.45 ± 0.19	14
Gillen et al. 2019⁵⁵	American football	Not reported	Novice	9203	4.49 ± 0.18	14
Gillen et al. 2019⁵⁶	American football	Not reported	Novice	7214	4.60 ± 0.30	13
Lockie et al. 2016⁵⁷	American football	20.11 ± 1.60	Novice	62	4.59 ± 0.20	17
McKay et al. 2018⁵⁸	American football	Not reported	Novice	7478	4.92 ± 0.48	14
McKay et al. 2020¹⁰	American football	Not reported	Novice	7478	4.92 ± 0.48	13
Banda et al. 2019⁶	Basketball	Not reported	Elite	12	4.72 ± 0.29	16
Bishop et al. 2017¹⁵	Cricket	26.2 ± 5.3	Sub-elite	14	4.75 ± 0.18	18
Cesar et al. 2017⁵⁹	General Athletes	20.0 ± 0.42	Elite	160	4.64 ± 0.33	16
Jones et al. 2010⁶⁰	General Athletes	Not reported	Sub-elite	46	5.39 ± 0.24	17
Wagner et al. 2014⁶¹	General Athletes	Not reported	Sub-elite	15	5.47 ± 0.13	16
Faigenbaum et al. 2006⁶²	General Athletes	15.5 ± 0.9	Novice	30	5.14 ± 0.03	17
Faigenbaum et al. 2007⁶³	General Athletes	13.5 ± 0.14	Novice	17	5.60 ± 0.00	18
Gillen et al. 2018⁶⁴	General Athletes	10.9 ± 2.1	Novice	69	6.05 ± 0.35	17
Jones and Lorenzo, 2013⁶⁵	General Athletes	11.8 ± 0.07	Novice	157	5.92 ± 0.11	18
Markovic et al. 2007⁶⁶	General Athletes	20.1 ± 1.10	Novice	93	4.97 ± 0.08	19
Sekulic et al. 2013⁶⁷	General Athletes	21.1 ± 0.71	Novice	63	5.52 ± 0.45	17
Stewart et al. 2014⁶⁸	General Athletes	16.7 ± 0.6	Novice	44	4.93 ± 0.44	15
Toyomura et al. 2018⁶⁹	General Athletes	22.8 ± 2.20	Novice	18	10.97 ± 0.01	17
Cordingley et al. 2019¹⁶	Ice Hockey	14 ± 1	Novice	92	4.86 ± 0.18	14
Vescovi and McGuigan, 2008 ^20‡	Lacrosse	19.7 ± 1.1	Elite	79	4.99 ± 0.24	16
Greene et al. 2019⁷⁰	Lacrosse	20.25	Sub-elite	20	5.20 ± 0.78	16
Hoffman et al. 2009)¹⁷	Lacrosse	19.2 ± 1	Sub-elite	22	4.93 ± 0.05	16
Sell et al. 2018⁷¹	Lacrosse	19.6 ± 1.6	Sub-elite	41	4.39 ± 0.03	18
Vescovi et al. 2007¹⁸	Lacrosse	19.8 ± 1.1	Sub-elite	84	4.99 ± 0.23	14
Carlson et al. 2019⁷²	Recreational Athletes	20.5 ± 0.5	Novice	11	5.53 ± 0.14	17
Chan et al. 2018⁷³	Recreational Athletes	29.8 ± 9.9	Novice	16	5.25 ± 0.04	19
Nikolenko et al. 2011⁷⁴	Recreational Athletes	23.40 ± 1.88	Novice	20	5.04 ± 0.30	15
Zweifel et al. 2017⁷⁵	Resistance-trained athletes	22.15 ± 2.2	Novice	26	4.92 ± 0.07	14
Freitas et al. 2018⁹	Rugby	Backs: 22.0 ± 2.3, Forwards: 23.1 ± 3.2	Elite	24	Not reported	16
Freitas et al. 2019⁵	Rugby	23.6 ± 1.41	Elite	36	Not reported	17
Beyer et al. 2016⁷⁶	Rugby	21.2 ± 1.4	Sub-elite	12	5.06 ± 0.28	13
Speirs et al. 2016⁷⁷	Rugby	18.1 ± 0.5	Sub-elite	18	4.66 ± 0.13	20
La Monica et al. 2016¹⁹	Rugby	20.2 ± 1.6	Novice	25	5.15 ± 0.01	18
Lockie and Jalilvand, 2017⁷⁸	Soccer	20.10 ± 1.20	Elite	20	5.07 ± 0.17	18
Lockie et al. 2017⁷⁹	Soccer	20.2 ± 1.2	Elite	21	5.08 ± 0.13	18
Lockie et al. 2018⁸	Soccer	20.19 ± 1.20	Elite	26	5.09 ± 1.60	16
McFarland et al. 2016⁸⁰	Soccer	18-23	Elite	36	5.00 ± 0.51	15
Risso et al. 2017⁸¹	Soccer	Starters: 20.4 ± 1.3, Non-starters: 20.1 ± 1.2	Elite	22	5.07 ± 0.04	17
Vescovi and McGuigan, 2008 ^20‡	Soccer	19.9 ± 0.9,	Elite	51	4.88 ± 0.20	16
Kavaliauskas et al. 2017⁸²	Soccer	22.00 ± 8.00	Sub-elite	14	5.99 ± 0.04	19
Magal et al. 2009⁸³	Soccer	Not reported	Sub-elite	12	4.88 ± 0.11	15
Derek et al. 2020⁸⁴	Soccer	INC: 15.75 ± 0.92; LEV: 15.1 ± 0.42; CG: 16.0 ± 0.57	Novice	46	5.20 ± 0.09	18
Millar et al. 2020⁸⁵	Soccer	Hip Thrust Group: 15.7 ± 0.8, Squat Group: 15.3 ± 0.7	Novice	15	5.26 ± 0.01	16
Moran et al. 2018⁸⁶	Soccer	pPHV – EG: 10.4 ± 0.8, CG: 10.0 ± 1.0; PPHV – EG: 13.6 ± 0.7, CG: 14.5 ± 1.0	Novice	42	5.44 ± 0.29	17
Vescovi and McGuigan, 2008 ^20‡	Soccer	15.1 ± 1.6	Novice	83	4.91 ± 0.22	16
Halil et al. 2013⁸⁷	Swimming	11.5 ± 0.07	Novice	20	6.08 ± 0.29	12
Kemal et al. 2013⁸⁸	Swimming	11.62 ± 0.07	Novice	21	6.06 ± 0.04	10
Eriksson et al. 2015¹	Tennis	14 ± 1.6	Novice	34	5.10 ± 0.04	18

‡ Denotes use of an article in more than one sport.

LEV = level elevation group, INC = incline elevation group, CG = control group, EG = experimental group, pPHV-EG = pre-peak height velocity experimental group, PPHV = post-peak height velocity experimental group.

Table 3.

Number of population samples by sport and skill level.

	Basketball	Cricket	General Athletes	Ice Hockey	Lacrosse	NFL	Recreational Athletes	Resistance-trained Athletes	Rugby	Soccer	Swimming	Tennis	Total
Elite	1		1		1	11			2	6			22
Sub-elite		1	2		4	5			2	2			16
Novice			8	1		9	3	1	1	4	2	1	30
Total	1	1	11	1	5	25	3	1	5	12	2	1	68

Quality score

For acute studies items 3 and 4 on the modified scale could not be assessed, additionally archival studies, where data collection came from historical archives, items 3, 4, and 7 on the modified scale could not be assessed and was excluded for these studies, therefore, these studies are assessed on a scale of 0 to 16 and 0 to 14, respectively. Quality score for all studies averaged 15.9 (± 2.09). Acute studies averaged 14.2 (± 1.72) out of 16, for intervention studies 17.5 (± 1.51) out of 20, and for archival data studies a 13.4 (± 0.77) out of 14 (Table 1). Acute study quality was affected by the inclusion or exclusion of item 1 and 9 in Table 1 (i.e. inclusion criteria stated, and use of results detailed). Of the 38 acute studies four did not state inclusion criteria, and 14 of the 38 studies ‘maybe’ stated inclusion criteria. Seventeen of the 38 acute studies ‘maybe’ stated detailed results (i.e. results lacked detail or were not presented clearly). Intervention study quality was affected by the inclusion or exclusion of item 4 and 8 (Table 1) (i.e. use of a control group (CG), and appropriate use of statistics). Nine of the 16 intervention studies included the use of a CG, as it was practical to use this study design in the tested population (e.g. randomised-control trial). Five of the 16 intervention studies did not use appropriate statistical analysis, where inferential statistics, within-group or between-group reliability, or effect size (ES) analysis may not have been conducted. Quality score for archival studies averaged 13.4 ± 0.77 out of 14 (Table 1). Archival study quality was affected by inclusion or exclusion of item 1, 2, and 6 in Table 1 (i.e. inclusion criteria clearly stated, the assignment of subjects, and results detailed). Ten of the 13 archival studies clearly described inclusion criteria for the subjects. Eleven of the 13 archival studies assigned subjects appropriately, based on playing position and skill level. Three of the 13 archival studies reported results in insufficient detail (i.e. reporting between group differences).

Reliability and technology

Reliability of the pro-agility was found to be reported in five different sports and in general athletic populations (Table 4). The change in mean denotes the systematic bias or random error of measurement,³⁵ and only three studies have reported the change in mean over multiple testing occasions.^15,40,64 The change in mean varied by 0.60% to 1.71%. The greatest change in the mean was found in general athletes,⁶⁴ the smallest change was noted in elite American football athletes.⁴⁰

Table 4.

Reliability and technology.

Author	N =	Sport	Skill level	Timing Technology	Number of trials x sessions (between trial length)	Reliability type	Change in mean (%)	ICC	CV (%)
Gains et al. 2010³⁹	24	American football	Elite	Stopwatch	2 × 2 (1 week)	Between-session		Turf: 0.92, grass: 0.98
Mann et al. 2016⁴⁰	64	American football	Elite	Stopwatch	2 × 2 (1 week)	Between-session (ICC), within-session (CV)	0.60%	0.914	1.9
Lockie et al. 2012⁴⁸	36	American football	Sub-elite	Timing light (dual beam)	2 × 1	Within-session (ICC)		0.93
Bishop et al. 2017¹⁵	14	Cricket	Sub-elite	Timing light (single beam)	3 × 2 (18 weeks)	Between-session (ICC), within-session (CV)	1.05%	0.852	1.35
Gillen et al. 2018⁶⁴	69	General athletes	Novice	Timing light (single beam)	2 × 2 (5 days)	Between-session (ICC), within-session (CV)	1.71%	6-9y: 0.86, 10-11y: 0.87, 12-15y: 0.80	6-9y: 4.24, 10-11y: 3.65, 12-15y: 4.95
Sekulic et al. 2013⁶⁷	63	General athletes	Novice	Timing light (single beam)	3 × 1	within-session (CV)			0.06
Stewart et al. 2014⁶⁸	44	General athletes	Novice	Timing light (single beam)	2 × 1	Relative within-day, reliability (ICC), between test (CV)		0.90	2.19
Carlson et al. 2019⁷²	11	Recreational athletes	Novice	Timing light (single beam)	2 × 2 (1 week)	Between session (ICC)		0.802
Speirs et al. 2016⁷⁷	18	Rugby	Sub-elite	Timing light (single beam)	3 × 2 (5 weeks)	cohens d		0.89
Eriksson et al. 2015¹	34	Tennis	Novice	Timing light (single beam) and Stopwatch	3 × 2 (3 days)	between session and within session reliability (ICC)		Between-session: Timing light: 0.91 (0.83–0.96), Stopwatch: 0.95 (0.90–0.97) Within-session: Timing light - S1: 0.95, S2: 0.96; Stopwatch - S1: 0.95, S2: 0.96

Key: ICC = intraclass-corelation coefficient, CV = coefficient of variation, 6-9y = 6 to 9 year old, 10-11y = 10 to 11 years old, 12-15y = 12 to 15 years old, S1 = session 1, S2 = session 2.

The ICC gives insight into rank order consistency between repeated testing occasions.³³ Nine studies reported ICCs,^{1,15,39,40,48,64,68,72,77} with seven studies reporting between-session reliability.^{1,15,39,40,64,68,72} The ICC ranged from 0.80 to 0.98, the lowest reliability was found in a novice general athlete population,⁶⁴ the highest ICC reported in an elite American football populations.³⁹

The CV provides insight into the absolute consistency of measures and is calculated as a ratio of the SD to the mean³⁵ and presented as a percentage. Only five studies reported the CV for within-day reliability.^{15,40,64,67,68} The CV ranged from 0.06% to 4.95%. The lowest and highest CVs were found in novice general athletes,^64,67 indicating low individual consistency and high variability of within-session measurement within the novice population.

In terms of the reliability associated with the various technologies, it seems that infrared timing light technology (ICC = 0.80–0.96; CV = 0.06% - 4.95%) is less stable than stopwatch measures (ICC = 0.91–0.98; CV = 1.9%). In saying this, if variability in performance is picked up by a more accurate technology (i.e. infrared timing lights) it may be resultant of biological error (i.e. fluctuation in performance) which stopwatch is unable to detect. It should be noted that most studies used single beam timing light technology. Seven of these studies^{1,15,64,67,68,72,77} reported the reliability of single beam timing lights, only one study⁴⁸ reported the reliability of dual beam timing lights, and three studies^1,39,40 reported reliability of stopwatches.

Overview of skill level

Mean pro-agility time across all studies was 5.00 ± 0.86 s, ranging from 4.17 to 6.06 s (Figure 2A). Averaged times for skill levels were 4.61 ± 0.29 s in elite, 4.91 ± 0.44 s in sub-elite, and 5.32 ± 1.14 s in novice athletes. When examining the performance times of each skill level, an overlap in pro-agility performance range was observed between the sub-elite and novice skill levels (Figure 2B). However, when using an averaged trendline, it was observed that there was a 0.30 s difference in mean pro-agility time between elite and sub-elite (4.61 ± 0.29 and 4.91 ± 0.44 s, respectively) and a 0.41 s difference between sub-elite and novices (4.91 ± 0.44 and 5.32 ± 1.14 s, respectively).

Figure 2.

Averaged and skill level performance quartiles.

In Figure 2A quartiles for averaged data across all studies are presented. The lower quartile can be observed between 4.17–4.58 s and represents the fastest 25% of pro-agility times, with the middle 50% of times between 4.58–5.19 s, and the slowest 25% between 5.19–6.06 s presented in the upper quartile. The quartile rankings for each skill level (elite, sub-elite, and novice) are presented in Figure 2B. The fastest 25% of performance times for each skill level are between 4.17–4.39 s in elite athletes, 4.33–4.60 s in sub-elite athletes, and 4.37–4.86 s in novice athletes.

Overview of sports

An overview of the performance times per sport is shown in Figure 3. Based on percentile ranks of all included literature, the fastest 25% (lower quartile) of pro-agility times include elite and sub-elite American football athletes. The middle 50% of all times was represented by elite general athletes (4.64 ± 0.33 s), basketball (4.72 ± 0.29 s), cricket (4.75 ± 0.18 s), novice American football (4.80 ± 0.42 s), ice hockey (4.86 ± 0.18 s), sub-elite and novice rugby athletes (4.86 ± 0.28 and 5.15 ± 0.01 s, respectively), elite and sub-elite lacrosse athletes (4.99 ± 0.24 and 4.88 ± 0.34 s, respectively), resistance-trained athletes (4.92 ± 0.07 s), elite soccer (5.06 ± 0.04 s), and tennis athletes (5.10 ± 0.04 s). With the slower 25% (upper quartile) of athletes being recreational athletes, (5.27 ± 0.25 s), novice soccer (5.30 ± 0.12 s), novice general athletes (5.94 ± 1.81 s), and swimming athletes (6.05 ± 0.12 s).

Figure 3.

Normative performance timeline.

As for positional performance findings (Table 5), American football was the only sport to provide pro-agility times in novice, sub-elite, and elite skill groups. Interestingly, at both novice and sub-elite levels pro-agility time was not reported in the fullback position. In lacrosse, only sub-elite positional performance was reported. Similarly, the same was observed in novice rugby and elite soccer players.

Table 5.

Pro-agility positional differences.

Sport	Position	Skill level
Sport	Position	Novice	Sub-elite	Elite
American football	Defensive back	4.52 ± 0.31	4.60 ± 0.11	4.25 ± 0.11
	Quarter back	4.62 ± 0.34	4.65 ± 0.10	4.34 ± 0.04
	Running back	4.59 ± 0.30	4.50 ± 0.10	4.30 ± 0.10
	Fullback			4.45 ± 0.14
	Wide receiver	4.53 ± 0.30	4.58 ± 0.08	4.27 ± 0.12
	Tight end	4.64 ± 0.35	4.68 ± 0.06	4.42 ± 0.12
	Offensive linemen	5.11 ± 0.40	5.40 ± 0.49	4.84 ± 0.32
	Defensive linemen	4.79 ± 0.49	5.06 ± 0.60	4.71 ± 0.35
	Linemen (offensive and defensive)	4.97 ± 0.50	5.35 ± 0.46	4.84 ± 0.38
	Line-backer	4.64 ± 0.34	4.62 ± 0.04	4.36 ± 0.13
Lacrosse	Starter		4.92 ± 0.22
	non-starters		4.94 ± 0.13
	attackers		4.81 ± 0.34
	defenders		4.77 ± 0.33
	midfielders		4.74 ± 0.32
	goal keepers		4.91 ± 0.14
Rugby	Forwards	5.14 ± 0.30
	Backs	5.16 ± 0.30
Soccer	Starter			5.04 ± 0.17
	non-starters			5.09 ± 0.17
	attackers			4.88 ± 0.00
	defenders			5.14 ± 0.29
	midfielders			5.11 ± 0.08
	goal keepers			4.94 ± 0.00

Discussion

This systematic review aimed to identify the reliability and established normative values for performance time in the pro-agility shuttle from the available literature. To the authors knowledge, this was the first study to establish variability and normative data of pro-agility performance across different sports, skill levels, and positions. Pro-agility times from relevant literature have been collated and synthesised to provide an overview of test variability, categorical performance values, and guide identification of ability relative to sport, skill level, or player position in the pro-agility shuttle (where applicable).²⁷ However, there was limited data pertaining the use of the pro-agility shuttle test across the different skill levels in each sport.

Hopkins,³⁵ believes measures of change in mean, relative (ICC), and absolute (CV) consistency need to be reported to fully understand the reliability of measures.^35,89 Only three studies^15,40,64 reported all three measures of reliability, only one study⁶⁸ reported the CV and ICC, and six studies^{1,39,48,67,72,77} reported a single reliability measure.

To quantify the repeatability of measures for a given performance task, reliability of measures should be determined over multiple testing occasions i.e. test-retest reliability. Additionally, to understand the effectiveness of a training programme, coaches must be confident that the tests they use are consistent across multiple testing occasions. Generally, three testing occasions are needed to observe whether changes in measures are becoming more consistent. Only seven studies^{1,15,39,40,64,72,77} reported test-retest reliability and these were only conducted over two testing occasions. All other studies calculated within session reliability. Furthermore, the pro-agility as it stands provides limited diagnostic value due to total-time being a single measure of performance.¹² Whereas advancing diagnostic capabilities of the pro-agility to differentiate between acceleration, deceleration, and COD measures would provide further insight into the capabilities of underlying physiological components is of value to applied practitioners, and may be used to guide COD programming.

From this review, it appears that stopwatches and timing lights are the most commonly used technologies to measure performance time of the pro-agility shuttle (Table 4). While these technologies are portable and easily available resources to applied practitioners, the technologies reported in these articles are not without limitations. It would be logical to assume that measures from infrared timing light technology would be more stable than stopwatch measures. This assumption is not supported with the presented data, as there is greater variability reported for infrared timing light technology, compared to stopwatch measures. However, data pertaining to infrared timing light technologies was represented more in novice and sub-elite athletes. This may increase variability due to less developed neuromuscular capacity to perform more complex movement patterns resulting in more variation in performance.^90,91 Additionally, stopwatch measures reliability was limited to two sporting populations (American football and tennis). Therefore, while present data does not support the above assumption, we suggest additional research be conducted to establish a meaningful conclusion.

Interestingly, a single study quantified performance using dual beam timing light technology.⁴⁸ While timing lights provide practitioners with more precise measurement of time, Cuthbert, Dos’ Santos⁹² reported moderate to relative reliability in single-beam timing lights (ICC = 0.63–0.86), acknowledging this could be improved with the use of dual beam timing lights.⁹³ Previous sprinting studies with dual beam timing lights have been shown to elicit greater accuracy of measurement compared to single beam lights^94,95 as dual beam systems are thought more accurate and reliable due to both lights having to be broken to record a time i.e. mitigating false triggers such as a hand breaking the infrared beams. However, this was counter to findings in this review, as dual beam was observed to be less reliable (ICC = 0.76) in sub-elite American football athletes,⁴⁸ compared to studies using single beam lights (ICC = 0.80–0.90) in sub-elite cricket and rugby, and novice general athletes.^{15,16,64,67,68,72,77}

While some single beam timing lights have built-in software to prevent false triggers from occurring, such as the SPARQ XLR8 timing system (SPARQ Products, Oconomowoc, WI, USA) used by Carlson, Fowler,⁷² other studies using single beam timing lights did not utilise this technology.^{1,15,64,67,68,77} Therefore, while both single and dual beam timing lights can be used reliably, it is clear that dual beam timing lights provide more accurate measures. However, single beam and stopwatches can be used, through practitioners should take caution when interpreting results with regards to consistency of the technology used to measure pro-agility performance.

Some scientists have arbitrarily chosen an analytical goal of the CV being 10% or below and ICC greater than 0.70 for measures to have acceptable reliability.²⁶ In terms of this review all studies measuring reliability reported acceptable levels of reliability for the pro-agility shuttle., However, more variability was observed in novice athletes and general athlete populations, compared to elite and sub-elite American football (Table 4). Therefore, it is recommended that practitioners be aware of the performance variability associated with the pro-agility in novice and general athlete population.

The results show the pro-agility shuttle is used in practice and research across a number of sports.^{1,10,48,50,67,68,87,88} American football comprised nearly 40% of all reviewed studies and 95% of all subjects. This could be due to the fact that most American football testing batteries include the pro-agility shuttle, and test performance has implications for NFL draft status.^{10,22,23,36,47} However, since other sports, such as soccer, cricket, and tennis include many changes of direction,^7,15,96,97 the utility of the pro-agility shuttle could benefit sports which require multiple high degree directional changes, as a measure of repeated 180° COD ability.

From the review, it is clear that skill level plays an important factor in pro-agility performance (see Figure 2B). As would be expected, elite athletes have been found to complete the pro-agility faster than sub-elite and novice athletes, across all sports, excluding lacrosse (Table 2). Elite athletes are likely faster as a result of more developed neuromuscular systems, enabling greater force capacity to perform more complex movement skills with high synergistic muscle activation and high rates of force development.^90,91 Interestingly, findings from the similar spread in performance, indicate that sub-elite and novice athletes are characterised by similar athletic capabilities when it comes to 180° COD performance.⁸² It is important to note however, the variation in number of studies found for the data presented in each sport, in terms of skill level and player positions. For example, lacrosse performance data gathered was mostly in sub-elite athletes,^17,18,70,71 with elite athlete performance being reported in a single study.²⁰ Therefore, we recommend caution to be taken when interpreting values representative of single studies or skill levels.

The type of sport athletes participate in also appears to influence pro-agility performance. This is because each sport has specific requirements for performance. American football athletes had the fastest pro-agility time in comparison to any other sports reported in this review. These faster performance times may be due to the historic use of the pro-agility in the NFL combine to identify specific components of athletic fitness.^22,36,98 American football athletes perform better than other sports because it either might represent sport specific movement patters for this sport, or because these athletes spend more time practicing this movement. For example, because match performance may be more reliant on short accelerative and decelerative ability in American football athletes, sprints are shorter than that performed by rugby athletes.^13,99,100

Additionally, applied practitioners need to select and administer performance assessments appropriately based on the context of the sport and physiological capabilities to be measured.¹⁰¹ It would be thought that the pro-agility shuttle's current diagnostic capabilities are inefficient and limit the applicability of this test across a variety of sports, due to requiring large horizontal forces for multiple fast lateral movement over a small distance over ground.^11,12,98 However, research has reported use of the pro-agility in novice youth swimmers to assess athlete ability to efficiently apply vertical and horizontal forces to initiate movement, and change direction relating to start and quick turn performance.^87,88,102 Therefore, maximising ground reaction force and minimising ground reaction time should be a focus of programming for strength and conditioning coaches who are looking to develop athlete COD capabilities.¹⁰³

It is also of great importance for applied practitioners and coaches to understand the differences in performance values between the playing position of their athletes, and why these differences exist Robbins⁴³ substantiated this posit by concluding that athletic capabilities are partially moderated by playing position. The normative results of this review also indicate that mean performance times for each player position differ for American football, lacrosse, rugby, and soccer (Table 5). Therefore, positional normative values for the pro-agility should be used by practitioners who wish to be able to discern between positional performance differences.

A recent study by LaPlaca and McCullick,⁹⁸ found the pro-agility shuttle to be correlated to player performance. They concluded certain performance test results are significant for different player positions in American football. The study reported the pro-agility shuttle was correlated with better grades for offense. In offensive positions, pro-agility performance for players in the fullback position showed a correlation to the ability to avoid more tackles (r = −0.753). However, pro-agility performance weakly correlated with number of games played in centre and quarterback positions (r = −0.29, −0.34, respectively). This weaker correlation may be explained, for example, by being in the middle of the offensive line creating the nonessential need for centres to perform 180° degree change of directions, due to the amount of support which can be offered by nearby teammates.⁹⁸ In defensive positions, they reported weak correlations to pro-agility shuttle performance in defensive tackle position with more pressure, hits, and sacks per pass rush snap count (r = −0.22, −0.23, −0.25, respectively). Faster pro-agility shuttle performance was correlated, although weakly, with more interceptions per pass coverage snap count (r = −0.29). Additionally, in the strong safety position slower pro-agility shuttle performance was correlated with tackling performance (r = 0.37). However, these findings are unsurprising, due to the performance demands of playing in defensive positions does not inherently require high 180° COD ability (i.e. ability to decelerate and change direction suddenly).^23,43,98

Pro-agility times for rugby players (novice backs and forwards) has been reported in one study only.¹⁹ There was no observable difference in pro-agility times between the two positional groups (5.14 ± 0.3 and 5.16 ± 0.3 (respectively)). In the case for lacrosse athletes, where minimal differences were observed between attack, midfield, and defender positions, with goal keepers reported to be the slowest of the sub-elite athletes. While attackers, midfielders, and defence players cover the field, the goal keepers cover the goal, thereby moving the least of all players. Goals in lacrosse are smaller than those used in hockey. Therefore, first step quickness may be more influential than 180° COD (i.e. they can stay in the frontal plane more instead of transitioning between sprinting, decelerating, and turning). This would be the opposite for soccer athletes; individual position times for goal keep (4.94 ± 0.00), defenders (5.14 ± 0.29), midfielders (5.1 ± 0.08), and attackers (4.88 ± 0.00) were reported in elite athletes.⁸ Interestingly, faster pro-agility times were reported in goal keep and attacker positions. The physical and technical constraints imposed upon goal keepers may explain this, as they are required to perform high-intensity lateral movements and sprints over 5 m with a high rate of force development and application, while being co-ordinated.^104,105 As for swimming athletes, they presented the slowest times (6.05 ± 0.12). It would be expected for swimmers to present the slowest times over all sports, as assessments should reflect the demands of the sport, and swimmers may have the slowest times because of the limited applicability of over ground assessment to water sports. Nonetheless, further research needs to be conducted into positional differences in other sporting codes at the different skill levels before a comprehensive analysis of positional differences can be determined.

This review was the first to establish normative performance values for the pro-agility shuttle, across different sporting codes. While being the catalyst, this also presented itself as a limitation for this review. Given the limited data presented in literature, comprehensive identification of normative pro-agility times could not be established, except for American football, throughout the range of sports and their respective skill levels. Therefore, an overview of pro-agility performance was identified, with normative values for each skill level and player position being reported where appropriate. Additionally, values represented by few or single studies should be interpreted with caution. Given the paucity of research in basketball, cricket, ice hockey, resistance trained athletes, and tennis athletes future research into pro-agility performance in these sports is recommended. Additional evidence in the aforementioned sports would be valuable to further understand the representation of pro-agility performance and better establish pro-agility normative values. It should also be noted that athletes from different genders, specific training methods, and training ages were not characterised in this review. Future research should explore the inclusion of other technologies, such as additional timing lights, smartphone camera, inertial sensor or radar technology to provide advancement to the diagnostic protocol of the pro-agility test Nonetheless, future analysis of pro-agility performance and diagnostic protocol advancement presents itself to be necessary in order to establish comprehensive normative data within a wide range of sports and provide insightful value to applied practitioners.

The results of this review demonstrate the utility of the pro-agility shuttle, concluding the test to be reliable when using stopwatch and timing light measurement technology. It would seem from the data reviewed that a comprehensive understanding of the reliability of the pro-agility is yet to be established. Test-retest methodologies that use a comprehensive suite of reliability statistics are essential to fully understand the reliability of the pro-agility test Furthermore, establishing variability between single and dual beam timing lights, and the reliability associated with other technologies such as smartphone camera, radar, contact mat, etc. is needed. This review provided normative values that practitioners can use to understand performance in sport, skill level, and player positions. It was also the first to collectively provide normative pro-agility performance data across a range of sports and present areas requiring investigation. It was apparent the sporting context influenced pro-agility performance. However, whether the movement patterns associated with these sports is assessed specifically within the pro-agility is unclear i.e. the ecological validity of the pro-agility across these sports for quantifying COD capability. Additionally, the current diagnostic value of the pro-agility is limited as it provides practitioners with a total-time. Higher level diagnostics could be achieved by breaking the test into sub-components; however, the reliability, validity, and utility of these sub-component performance measures would need to be established. Additionally, whether other technologies such as inertial sensors, smartphone videography, and radar can add value to the diagnostics needs to be explored.

Practical applications

Practitioners wishing to understand their athletes’ 180° COD ability should look to use the pro-agility shuttle to assess repetitive 180° COD ability. The pro-agility shuttle can help provide insight into acceleration, deceleration, and 180° COD capabilities: critical movements in most field and court sports. As such, practitioners should look to use normative values as guidelines to compare relative performance to gauge and monitor athlete performance. Furthermore, practitioners should think about maximising ground reaction forces while minimising ground reaction time in order to develop pro-agility performance. In addition, the normative values established, across the various sports examined, may be used to set attainable goals appropriate to the sport, position, and skill level in context. While practitioners should utilise the technology available to them for pro-agility shuttle assessment. We recommend practitioners use dual beam timing lights as a first choice for talent identification and athletic development, as they have the greatest accuracy of measurement, however, for daily monitoring stopwatches are sufficient, and best practice guidelines should be followed. For example, some single beam timing lights may have different software to prevent false triggers from occurring. Therefore, practitioners should utilise the same timing technology to allow for consistency of measurement.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iDs

James W. D. Forster

Aaron M. Uthoff

References

Eriksson

Johansson

Back

. Reliability and criterion-related validity of the 20-yard shuttle test in competitive junior tennis players. Open Access J Sports Med 2015; 6: 269–276. 2015/09/01.

Sheppard

Young

. Agility literature review: classifications, training and testing. J Sports Sci 2006; 24: 919–932.

Young

James

Montgomery

. Is muscle power related to running speed with changes of direction? J Sports Med Phys Fitness 2002; 42: 282–288.

Reilly

Williams

Nevill

, et al. A multidisciplinary approach to talent identification in soccer. J Sports Sci 2000; 18: 695–702.

Freitas

Alcaraz

Calleja-González

, et al. Differences in change of direction speed and deficit between male and female national rugby sevens players. J Strength Conditioning Res 2019: 1–7. DOI: 10.1519/JSC.0000000000003195

Banda

Beitzel

Kammerer

, et al. Lower-body power relationships to linear speed, change-of-direction speed, and high-intensity running performance in DI collegiate women's Basketball players. J Hum Kinet 2019; 68: 223–232.

Lockie

Callaghan

Jeffriess

. Analysis of specific speed testing for cricketers. J Strength Conditioning Res 2013; 27: 2981–2988. 2013/02/27.

Lockie

Moreno

Lazar

, et al. The physical and athletic performance characteristics of division I collegiate female soccer players by position. J Strength Conditioning Res 2018; 32: 334–343.

Freitas

Alcaraz

Bishop

, et al. Change of direction deficit in national team rugby union players: is there an influence of playing position? Sports 2018; 7. 2018/12/24. DOI: 10.3390/sports7010002.

10.

McKay

Miramonti

Gillen

, et al. Normative reference values for high school–aged American football players: proagility drill and 40-yard dash split times. J Strength Conditioning Res 2020; 34: 1184. DOI: 10.1519/JSC.0000000000002930

11.

Nimphius

. High-performance training for sports. In: Joyce

Lewindon

(eds) High-Performance training for sports. Champaign, IL: Human Kinetics, 2014, pp.185–198.

12.

Nimphius

Callaghan

Bezodis

, et al. Change of direction and agility tests: challenging our current measures of performance. Strength Conditioning J 2018; 40: 26–38.

13.

Brechue

Mayhew

Piper

. Characteristics of sprint performance in college football players. J Strength Conditioning Res 2010; 24: 1169–1178.

14.

Jones

Nimphius

. Change of direction and agility. Performance assessment in strength & conditioning. London: Routledge, 2018, pp.140–165.

15.

Bishop

Herridge

Turner

. Monitoring changes In power, speed, agility and endurance in elite cricketers during the off-season. J Strength Conditioning Res 2017; 34: 2285–2293. 2017/06/29. DOI: 10.1519/jsc.0000000000002077.

16.

Cordingley

Sirant

MacDonald

, et al. Three-year longitudinal fitness tracking in top-level competitive youth ice hockey players. J Strength Conditioning Res 2019; 33: 2909–2912.

17.

Hoffman

Ratamess

Neese

, et al. Physical performance characteristics in national collegiate athletic association division III champion female lacrosse athletes. J Strength Conditioning Res 2009; 23: 1524–1529. 2009/07/22.

18.

Vescovi

Brown

Murray

. Descriptive characteristics of NCAA division I women lacrosse players. J Sci Med Sport 2007; 10: 334–340.

19.

La Monica

Fukuda

Miramonti

, et al. Physical differences between forwards and backs in American collegiate rugby players. J Strength Conditioning Res 2016; 30: 2382–2391.

20.

Vescovi

McGuigan

Relationships between sprinting, agility, and jump ability in female athletes. J Sports Sci 2008; 26: 97–107.

21.

Vescovi

Brown

Murray

. Positional characteristics of physical performance in division I college female soccer players. J Sports Med Phys Fitness 2006; 46: 221.

22.

McGee

Burkett

. The national football league combine: a reliable predictor of draft status? J Strength Conditioning Res 2003; 17: 6–11. 2003/02/13.

23.

Nuzzo

. The national football league scouting combine From 1999 to 2014: normative reference values and an examination of body mass normalization techniques. J Strength Conditioning Res 2015; 29: 279–289.

24.

Bourgeois

McGuigan

Gill

, et al. Physical characteristics and performance in change of direction tasks: a breif review and training considerations. J Aust Strength Conditioning 2017; 25: 104–117.

25.

Jones

Bampouras

Marrin

. An investigation into the physical determinants of change of direction speed. J Sports Med Phys Fitness 2009; 49: 97–104. 2009/02/04.

26.

Atkinson

Nevill

. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 1998; 26: 217–238.

27.

O’Donoghue

. Normative profiles of sports performance. Int JPerform Anal Sport 2005; 5: 104–119.

28.

Falch

Rædergård

Tillaar

. Effect of different physical training forms on change of direction ability: a systematic review and meta-analysis. Sports Med Open 2019: 5. DOI: 10.1186/s40798-019-0223-y

29.

Moher

Shamseer

Clarke

, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015; 4: –1.

30.

Lorenz

Reiman

Lehecka

, et al.

What performance characteristics determine elite versus nonelite athletes in the same sport?

Sports Health 2013; 5: 542–547.

31.

McMaster

Gill

Cronin

, et al. The development, retention and decay rates of strength and power in elite rugby union, rugby league and American football. Sports Med 2013; 43: 367–384.

32.

Brughelli

Cronin

Levin

, et al. Understanding change of direction abilility in sport: a review of resistance training studies. Sports Med 2008; 38: 1045–1063.

33.

Koo

. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 2016; 15: 155–163.

34.

Uthoff

Oliver

Cronin

, et al.

Prescribing target running intensities for high-school athletes: can forward and backward running performance be autoregulated?

Sports 2018; 6: 1–10.

35.

Hopkins

. Measures of reliability in sports medicine and science. Sports Med 2000; 30: 1–15.

36.

Corey

Randall

. A comparison of the national football league's Annual national football league combine 1999–2000 to 2015–2016. J Strength Cond Res 2020; 34: 1–11. DOI: 10.1519/JSC.0000000000002682

37.

Yamashita

Asakura

Ito

, et al. Physical characteristics and performance of Japanese top-level American football players. J Strength Conditioning Res 2017; 31: 2455–2461.

38.

Davis

Barnette

Kiger

, et al. Physical characteristics that predict functional performance in division I college football players. J Strength Conditioning Res 2004; 18: 115–120.

39.

Gains

Swedenhjelm

Mayhew

, et al. Comparison of speed and agility performance of college football players on field turf and natural grass. J Strength Conditioning Res 2010; 24: 2613–2617.

40.

Mann

Ivey

Mayhew

, et al. Relationship between agility tests and short sprints: reliability and smallest worthwhile difference in national collegiate athletic association division I football players. J Strength Conditioning Res 2016; 30: 893–900.

41.

Nesser

Huxel

Tincher

, et al. The relationship between core stability and performance in division I football players. J Strength Conditioning Res 2008; 22: 1750–1754.

42.

Nimphius

Geib

Spiteri

, et al. “Change of direction deficit” measurement in division I American football players. J Aust Strength Conditioning 2013; 21: 115–117.

43.

Robbins

. Positional physical characteristics of players drafted into the national football league. J Strength Conditioning Res 2011; 25: 2661–2667.

44.

Robbins

Goodale

Kuzmits

, et al. Changes in the athletic profile of elite college American football players. J Strength Conditioning Res 2013; 27: 861–874.

45.

Stodden

Galitski

. Longitudinal effects of a collegiate strength and conditioning program in American football. J Strength Conditioning Res 2010; 24: 2300–2308.

46.

Hoffman

Ratamess

Kang

. Performance changes during a college playing career in NCAA division III football athletes. J Strength Conditioning Res 2011; 25: 2351–2357.

47.

Leutzinger

Gillen

Miramonti

, et al. Anthropometric and athletic performance combine test results among positions within grade levels of high school–aged American football players. J Strength Conditioning Res 2018; 32: 1288–1296.

48.

Lockie

Schultz

Callaghan

, et al. Physiological profile of national-level junior American football players in Australia. Serb J Sports Sci 2012; 6: 127–136.

49.

Newsom

Probst

. Relationship between performance tests and yards per carry average in collegiate running backs. Morehead State University, 2016.

50.

Yuasa

Kurihara

Isaka

. Relationship between toe muscular strength and the ability to change direction in athletes. J Hum Kinet 2018; 64: 47–55.

51.

Caswell

Ausborn

Diao

, et al. Anthropometrics, physical performance, and injury characteristics of youth American football. Orthop J Sports Med 2016; 4: 2325967116662251. 2016/09/02.

52.

Collins

Coburn

Galpin

, et al. Relationships and reliability between a drive block test and traditional football performance tests in high school offensive line players. J Strength Conditioning Res 2018; 32: 3423–3432.

53.

Dupler

Amonette

Coleman

, et al. Anthropometric and performance differences among high-school football players. J Strength Conditioning Res 2010; 24: 1975–1982.

54.

Ghigiarelli

. Combine performance descriptors and predictors of recruit ranking for the top high school football recruits from 2001 to 2009: differences between position groups. J Strength Conditioning Res 2011; 25: 1193–1203.

55.

Gillen

Shoemaker

McKay

, et al. Performance differences between national football league and high school American football combine participants. Res Q Exercise Sport 2019; 90: 227–233.

56.

Gillen

Shoemaker

McKay

, et al. State population influences athletic performance combine test scores in high school-aged American football players. Int J Exerc Sci 2019; 12: 256–262. https://digitalcommons.wku.edu/ijes/vol12/iss6/2/ (accessed 2 April 2019).

57.

Lockie

Lazar

Orjalo

, et al. Profiling of junior college football players and differences between position groups. Sports 2016; 4. 2016/08/05. DOI: 10.3390/sports4030041.

58.

McKay

Miramonti

Gillen

, et al. Normative reference values for high school-aged American football players. J Strength Conditioning Res 2018; 43: 2849–2856. 2018/03/01. DOI: 10.1519/jsc.0000000000002532.

59.

Cesar

Edwards

Hasenkamp

, et al. Prediction of athletic performance of male and female athletes measured by triple hop for distance. Trends Sport Sci 2017; 24: 19–25.

60.

Jones

Matthews

Murray

, et al. Psychological correlates of performance in female athletes during a 12-week off-season strength and conditioning program. J Strength Conditioning Res 2010; 24: 619–628. 2010/01/23.

61.

Wagner

Oden

Glave

, et al. Development of agility utilising a multidimentional modality of plyometrics. J Fitness Res 2014; 3: 49–59.

62.

Faigenbaum

Kang

McFarland

, et al. Acute effects of different warm-up protocols on anaerobic performance in teenage athletes. Pediatr Exerc Sci 2006; 18: 64.

63.

Faigenbaum

McFarland

Keiper

, et al. Effects of a short-term plyometric and resistance training program on fitness performance in boys age 12 to 15 years. J Sports Sci Med 2007; 6: 519–525. 2007/01/01.

64.

Gillen

Miramonti

McKay

, et al. Test-retest reliability and concurrent validity of athletic performance combine tests in 6–15 year-old male athletes. J Strength Conditioning Res 2018; 32: 2783–2794.

65.

Jones

Lorenzo

. Assessment of power, speed, and agility in athletic, preadolescent youth. J Sports Med Phys Fitness 2013; 53: 693–700. 2013/11/20.

66.

Markovic

Jukic

Milanovic

, et al. Effects of sprint and plyometric training on muscle function and athletic performance. J Strength Conditioning Res 2007; 21: 543–549. 2007/05/29.

67.

Sekulic

Spasic

Mirkov

, et al. Gender-specific influences of balance, speed, and power on agility performance. J Strength Conditioning Res 2013; 27: 802–811. 2012/05/15.

68.

Stewart

Turner

Miller

. Reliability, factorial validity, and interrelationships of five commonly used change of direction speed tests. Scand J Med Sci Sports 2014; 24: 500–506.

69.

Toyomura

Mori

Tayashiki

, et al. Efficacy of downhill running training for improving muscular and aerobic performances. Applied Physiology. Nutr Metab (Lond) 2018; 43: 403–410.

70.

Greene

Perryman

Cleary

, et al. Core stability and athletic performance in male and female lacrosse players. Int J Exerc Sci 2019; 12: 1138–1148. 2019/12/17.

71.

Sell

Prendergast

Ghigiarelli

, et al. Comparison of pysical fitness parameters for starters vs. Nonstarters in an NCAA division I men's Lacrosse team. J Strength Conditioning Res 2018; 32: 3160–3168. 2018/09/15.

72.

Carlson

Fowler

Lawrence

. Agility and vertical jump performances are impacted by acute cool exposure. J Strength Conditioning Res 2019; 33: 1648–1652.

73.

Chan

Yung

. Sprint cycling training improves intermittent run performance. Asia Pac J Sports Med, Arthroscopy, Rehabil Technol 2018; 11: 6–11. 2018/03/20.

74.

Nikolenko

Brown

Coburn

, et al. Relationship between core power and measures of sport performance. Kinesiology 2011; 43: 163–168.

75.

Zweifel

Vigotsky

Contreras

, et al. Effects of 6-week squat, deadlift, or hip thrust training program on speed, power, agility, and strength in experienced lifters: a pilot study. Journal of Trainol 2017; 6: 13–17.

76.

Beyer

Fukuda

Miramonti

, et al. Spatial awareness is related to moderate intensity running during a collegiate rugby match. Int J Exerc Sci 2016; 9: 599–606.

77.

Speirs

Bennett

Finn

, et al. Unilateral vs. Bilateral squat training for strength, sprints, and agility in academy ruygby players. J Strength Conditioning Res 2016; 30: 386–392.

78.

Lockie

Jalilvand

. Reliability and critereon validity of the arrowhead change-of-direction speed test for soccer. Facta Univ: Ser Phys Educ Sport 2017; 15: 139–151.

79.

Lockie

Jalilvand

Moreno

, et al. Yo-yo intermittent recovery test level 2 and its relationship with other typical soccer field tests in female collegiate soccer players. J Strength Conditioning Res 2017; 31: 2667–2677.

80.

McFarland

Dawes

Elder

, et al. Relationship of two vertical jumping tests to sprint and change of direction speed among male and female collegiate soccer players. Sports 2016; 4 2016/02/16. DOI: 10.3390/sports4010011.

81.

Risso

Jalilvand

Orjalo

, et al. Physiological characteristics of projected starters and non-starters in the field positions from a division I women's Soccer team. Int J Exerc Sci 2017; 10: 567–579.

82.

Kavaliauskas

Kilvington

Babraj

. Effects of in-season uphill sprinting on physical characteristics in semi-professional soccer players. J Sports Med Phys Fitness 2017; 57: 165–170.

83.

Magal

Smith

Dyer

, et al. Seasonal variation in physical performance-related variables in male NCAA division III soccer players. J Strength Conditioning Res 2009; 23: 2555–2559. 2009/11/17.

84.

Derek

Shane

Matthew

. Combined sprint interval, plyometric, and strength training in adolescent soccer players: effects on measures of speed, strength, power, change of direction, and anaerobic capacity. J Strength Conditioning Res 2020; 34: 1–12. DOI: 10.1519/JSC.0000000000003476

85.

Millar

Colenso-Semple

Lockie

, et al. In-season hip thrust vs. Back squat training in female high school soccer players. Int J Exerc Sci 2020; 13: 49–61.

86.

Moran

Parry

Lewis

, et al. Maturation-related adaptations in running speed in response to sprint training in youth soccer players. J Sci Med Sport 2018; 21: 538–542.

87.

Halil

Nurtekin

Dede

, et al. Effect of vertical jump on quickness, agility, acceleration and speed performance in children swimmer. Ovidius Univ Ann Ser Phys Educ Sport/Sci, Mov Health 2013; 13: 347–351.

88.

Kemal

Cecilia

Metin

, et al. Evaluation of agility with regard to different start positions. Ovidius Univ Ann Ser Phys Educ & Sport/Sci, Mov & Health 2013; 13: 343–346.

89.

Hopkins

. A new view of statistics. Sport Sci 2015; 19: 36–42.

90.

Cormie

McGuigan

Newton

. Developing maximal neuromuscular power. Sports Med 2011; 41: 17–38.

91.

Duchateau

Baudry

. Neuromuscular aspects of sport performance, Volume XVII. 2010, pp.216–253.

92.

Cuthbert

Dos’ Santos

Thomas

, et al. The application of change of direction deficit to evaluate cutting ability. J Strength Cond Res 2017; 33: 2138–2144. DOI: 10.1519/JSC.0000000000002346

93.

Haugen

Buchheit

. Sprint running performance monitoring: methodological and practical considerations. Sports Med 2016; 46: 641–656. 2015/12/15.

94.

Yeadon

Kato

Kerwin

. Measuring running speed using photocells. J Sports Sci 1999; 17: 249–257.

95.

Earp

Newton

. Advances in electronic timing systems: considerations for selecting an appropriate timing system. J Strength Conditioning Res 2012; 26: 1245–1248.

96.

Bloomfield

Polman

O’Donoghue

. Physical demands of different positions in FA premier league soccer. J Sports Sci Med 2007; 6: 63–70.

97.

Leone

Comtois

Tremblay

, et al. Specificity of running speed and agility in competitive junior tennis players. J Med Sci Tennis 01/01 2006; 11: 10–11.

98.

LaPlaca

McCullick

. National football league scouting combine tests correlated to national football league player performance. The J Strength Conditioning Res 2020; 34: 1317–1329.

99.

Docherty

Wenger

Neary

. Time-motion analysis related to the physiological demands of rugby. J Human Mov Stud 1988; 14: 269–277.

100.

Deutsch

Maw

Jenkins

, et al. Heart rate, blood lactate and kinematic data of elite colts (under-19) rugby union players during competition. J Sports Sci 1998; 16: 561–570.

101.

McGuigan

Cormack

Gill

. Strength and power profiling of athletes: selecting tests and how to use the information for program design. Strength Conditioning J 2013; 35: 7–14.

102.

Kilani

Al-Tuieb

Kilani

. Analysis of the sprint start, swimming start and volleyball push-off in generating impulse momentum. Int Conf Biomech Sports. 2013.

103.

Bishop

Read

Stewart

, et al. Strength and conditioning for sprint swimming. Strength Cond J 2013; 35: 1–6.

104.

Spratford

Mellifont

Burkett

. The influence of dive direction on the movement characteristics for elite football goalkeepers. Sports Biomech 2009; 8: 235–244.

105.

Di Salvo

Benito

Calderon

, et al. Activity profile of elite goalkeepers during football match-play. J Sports Med Phys Fitness 2008; 48: 443.