Secondary Logo

Journal Logo

Original Research

Relationship Between Agility Tests and Short Sprints

Reliability and Smallest Worthwhile Difference in National Collegiate Athletic Association Division-I Football Players

Mann, J. Bryan1,2; Ivey, Pat A.1; Mayhew, Jerry L.3,4; Schumacher, Richard M.3; Brechue, William F.4

Author Information
The Journal of Strength & Conditioning Research: April 2016 - Volume 30 - Issue 4 - p 893-900
doi: 10.1519/JSC.0000000000001329
  • Free

Abstract

Introduction

Although horizontal running speed is a key component in athletics, sport performance is dominated by change of direction (COD) movements. This component of physical competency has been termed agility and relates the quickness by which the performer can change direction. Change of direction requires acceleration interspersed with complete deceleration to change direction and initiate a sports skill. There are many tests that purport to assess agility and perhaps one factor confounding the issue with COD maneuvers is the variety of tests used to measure this component of sport. Some COD tests combine sprints with a weaving movement through obstacles (i.e., Illinois agility run (35); Australian Football League [AFL] test (9)), whereas others include a combination of forward and backward sprinting and sliding movements (i.e., SEMO test (17), T-test (35)). Still other tests rely on sprinting in different directions with frequent COD (i.e., pro-agility or I-Test (36) and 3-cone drill (35)).

The most common tests to evaluate agility currently used in Division-I football are the I-Test and 3-cone drill (3-CD). Both these tests are essential parts of the National Football League (NFL) Combine profile (24,25,30) and are often accepted on face validity as being adequate measures of agility, although the singular concept of agility remains questionable (38). In addition, there is growing attention being paid to short sprints such as the 10-yd dash to evaluate a player's acceleration ability and is another test increasingly used to assess player potential in the NFL Combine and for college and high school prospects (38). Although tests of agility have been widely studied (6,15,19,21,23,27–29,31–33), there remain 2 key issues that have not been adequately addressed: (a) the relationship between performance in agility and the contribution of straight-ahead acceleration to specific COD movements and (b) an assessment of the magnitude of change in performance that constitutes a true change in agility and 10-yd sprint performance.

Because the objective of most football strength and conditioning programs is the continual improvement of performance, determining the amount of improvement in COD and 10-yd sprint tests that represents more than random fluctuation evident across repeated trials would provide strength and conditioning professionals with a sound basis for judging the impact of their training programs. Although reliability documentation is available for the I-Test (35) and 3-CD (35), no such data exist for the 10-yd sprint. Furthermore, there have been no assessments of the smallest worthwhile difference (SWD) for any of these tests. The SWD is now widely accepted as a useful criterion for judging the amount of change that reflects a meaningful improvement in performance as opposed to random trial-to-trial variation (13). Such information is lacking for many tests used in football for assessing player performance level or ability. Thus, it would be beneficial to strength and conditioning professionals to evaluate the absolute and relative reliabilities of these tests to aid in greater scientific assessment of player potential and performance. Absolute reliability analysis provides an indication of the consistency with which players are able to repeat a performance on a given test over time (2,12). Relative reliability provides an assessment of the consistency with which players maintain their ranking within a group when performing a specific test (2,12). In addition, estimation of the SWD provides a minimum value, which allows the evaluation of the meaningfulness of changes in performance that reflect real improvement beyond the random variation noted during repeated trials (13). Therefore, the purpose of this study was to determine the absolute and relative reliabilities and SWD of the I-Test, 3-CD, and 10-yd sprint in major college football players.

Methods

Experimental Approach to the Problem

Most Division-I college football programs now widely use agility tests to evaluate players' COD ability. However, there seems to be a lack of estimates of the reliability of these tests in major college football players. This study was designed to assess the absolute and relative reliabilities of the I-Test, 3-CD, and 10-yd sprint to determine the smallest SWD for each to allow an assessment of the effectiveness of various training programs to promote improvements in performance and assess the relationship between tests of agility and the 10-yd sprint. Players were tested during 2 consecutive weeks at the same time of day (1200–1530 hours) during an off-season winter conditioning period (second week of March). The training load the day before testing was light to insure maximum rest during subsequent performance. Strong verbal encouragement was given by coaches and players during all performances to maximize their effort.

Subjects

Sixty-four players (age = 20.5 ± 1.2 years, height = 185.2 ± 6.1 cm, body mass = 107.8 ± 20.7 kg) from a successful Division-I program that was consistently ranked in the top 25 in the country were measured at the conclusion of a 6-week off-season conditioning program (Table 1). All players had a minimum of 5 years' experience performing sprints and agility drills. Only players who were free of any lower extremity injuries within the previous year were evaluated. The testing program was part of the regular training procedures for the team, approved by the Institutional Review Board, and all players signed a consent document to participate before testing. No players below the age of 18 years were included in the study. The study conforms to the Code of Ethics of the World Medical Association (approved by the ethics advisory board of Swansea University) and required players to provide a signed informed consent document before participation.

Table 1
Table 1:
Mean, SD, and range of agility and sprint tests in college football players (n = 64).

Procedures

All testing was performed on an indoor artificial turf with the player wearing multicleat turf shoes, gym shorts, and tee shirt. Each player performed 2 trials of each test during 2 consecutive weeks, with the better time used to represent each week. Rest interval between trials was a minimum of 3 minutes. Test order was randomized by position groups to minimize the effect of fatigue on test performance. Agility tests were hand timed by experienced timers, whereas the 10-yd sprint was timed electronically (Speed Trap, model II; Brower Timing Systems, Draper, UT, USA). All 3 tests began on the player's volition. Hand timing in the agility tests commenced with the first indication of movement by the player (starting position is described below). For the 10-yd sprint, all players started from a 3-point stance, and the start of each trial using a touch pad start (ground hand) and an infrared beam stopped timing; the stop beam was placed 75 cm above the ground.

3-Cone drill

The player assumed a 3-point stance with the down-hand on an imaginary line next to the first cone. Initially the player sprinted forward 5 yds, touched the second cone, performed a 180° turn, and returned to touch the first cone. After another 180° turn, the player sprinted back to the second cone, executed a 90° turn to the right, and sprinted toward the third cone. The player circled the third cone on his left, sprinted back to the second cone, executed a 90° turn to the left around the cone, and sprinted back past the starting line to finish the test (35).

I-Test

The I-Test, also known as the pro-agility or the 5-10-5 test, measures the ability to change direction laterally to the right and left. The player assumed an initial position in a 3-point stance facing forward while straddling the midline. Each player had the choice of initiating the sprint to the right or left on the first trial; the second trial was initiated in the opposite direction. The test commenced with the player sprinting 5 yd and touching a line with his foot, then turning 180° and sprinting 10 yds to the other outside line and touching it with his foot. The test was completed by performing another 180° turn and sprinting back across the midline (35).

10-yd Sprint

The 10-yd sprint was timed as part of a 40-yd sprint. The 10-yd distance was set with a steel tape, and an infrared timing box that marked the end of distance was placed there.

Statistical Analyses

Separate paired t-tests were performed between each pair of trials, with Bonferroni correction for multiple t-tests, to determine systematic bias and establish the 90% confidence intervals for the difference between successive trials. Typical error of measurement was computed using the formula:

, where MSE represented the mean square error from the repeated-measures analysis of variance for the 2 trials of each performance (12). Coefficient of variation (CV%) for each test was calculated from the ratio of the average of individual SDdiff for successive trials divided by the mean of successive trials (13). The 90% limits of agreement (LoA) was estimated by the method suggested by Bland and Altman (3). The SWD was calculated using a 90% CI with the formula:

(13). Intraclass correlation coefficient (ICC) was calculated according to the method detailed by Weir (37). Pearson correlations were used to evaluate the relationships among all variables. Partial correlations and multiple regression analysis were used to hold the impact of some variables constant or determine the relative contribution of several variables to the prediction of speed or agility. Significance level was set at p ≤ 0.05 for all analyses, and power exceeded 0.80 for all analyses.

Results

Performance data for each test are given in Table 1. There was no significant difference between trials for any performance test, although there was a trend toward a slower second trial for the 3-CD (p = 0.06). Players (n = 41) with I-Test times less than 4.60 seconds on the first trial produced significantly slower (p < 0.01) second trials than players (n = 23) with initial times greater than 4.60 seconds (p = 0.52). Likewise, players (n = 39) with times less than 7.50 seconds in the first 3-CD trial had significantly slower second trial times (p < 0.03) than players (n = 25) with initial times greater than 7.50 seconds (p = 0.68). However, in the 10-yd sprint, players (n = 21) who were slower than 1.85 seconds initially produced significantly faster times (p = 0.02) on the second trial, whereas players (n = 43) who ran faster than 1.85 on the first trial did not produce a significantly different second trial (p = 0.32). The percent difference between duplicate trials averaged less than 1% for all tests.

Bland-Altman plots were constructed and 90% CIs between trials for each test were determined (Figure 1). The LoA established indicated high agreement between trials. Absolute and relative reliabilities are shown in Table 2. The high ICC values between trials indicated excellent levels of consistency (absolute reliability) for players to approximate their same performance during repeated trials for all 3 tests. Players should be able to repeat multiple trials of all 3 tests with less than 2% variation. Relative reliability was calculated for each test using rank-order correlations as suggested by Atkinson and Nevill (2). The high rho correlations indicated that players were “good” for the 3-CD and 10-yd sprint and “acceptable” for the I-Test at maintaining their relative ranking in a test-retest situation (Table 3).

Figure 1
Figure 1:
Bland-Altman plots (average of trials vs. difference between trials) illustrating level of agreement between trial 1 and trial 2 for I-Test, 3-CD, and 10-yd sprint.
Table 2
Table 2:
Absolute and relative reliabilities for 3-cone test, I-Test, and 10-yd sprint in Division-I college football players (n = 64).
Table 3
Table 3:
Correlation among demographic and performance variables in football players (n = 64).

Regression of one trial on another was used to illustrate the existence of any significant degree of bias (14). As shown in Figure 2, the regression line was near the line of identity across the full range of data for each test illustrating the lack of bias in any of the current tests in the overall sample. Despite the tendency for some players to have slower times in the second trial, there was no systematic bias as determined by paired t-tests performed between each pair of trials, with Bonferroni correction for multiple t-tests. The lack of systematic bias was also confirmed by Bland-Altman plots for each test allowing establishment of 90% CIs for the difference between successive trials.

Figure 2
Figure 2:
Regression of trial 3 on trial 2 for 3-CD, I-Test, and 10-yd sprint illustrating the lack of bias between trials.

The I-Test and 3-CD were highly correlated (r = 0.94; Table 2), which supports a strong commonality between them (r2 × 100 = 88%). The correlations between COD tests and 10-yd sprint were high, being higher with the I-Test (r = 0.85) than with the 3-CD (r = 0.69; Table 2), also supporting a strong relationship between tests. Removing the effect of 10-yd sprint performance significantly reduced the association between I-Test and 3-CD (r = 0.79), although a moderate of commonality still remained (r2 × 100 = 62%).

There were also a high positive correlation between COD tests and body mass and between sprint and body mass (Table 3). Removing the influence of body mass by partial correlation decreased the relationship between COD tests only slightly (r = 0.85), still suggesting a substantial degree of commonality between the COD tests (r2 × 100 = 72%). Furthermore, when the influence of body mass was removed by partial correlation, the relationship between 10-yd performance and COD decreased significantly (p < 0.01) for both I-Test (r = 0.63) and 3-CD (r = 0.68), reducing the commonality between them to less than 46%. The interaction of 10-yd performance and body mass was further illustrated when stepwise multiple regression analysis was used to predict 3-CD using these 2 variables. The 10-yd sprint was the first item selected followed by body mass to estimate both COD tests. The multiple Rs for estimating I-Test (R = 0.87) and 3-CD (R = 0.84) from 10-yd sprint and body mass were similar, as were the coefficients of variation (SEE/mean × 100 = 3.4% and 2.8%, respectively). The percent contribution of 10-yd sprint to the known variance of each test (64% and 74%, respectively) was substantially greater than that for body mass (36% and 26%, respectively).

The SWD provides a statistically derived value for determining whether a player has made a “real,” physiologically significant improvement in performance beyond his random variations in time that will be produced with multiple trials. As shown in Table 2, the SWD to indicate physiologically significant change in performance were 0.28 seconds and 0.30 seconds, respectively, for the I-Test and 3-CD. When SWD was expressed as a percent (6.6 and 3.7%, respectively; Table 3) to account for individual differences among players, the correlations between body mass and improvement in performance time were significant and negative for I-Test (r = −0.76), 3-CD (r = −0.75), and 10-yd sprint (r = −0.65), indicating that heavier players would need to make greater decreases in time to be considered meaningful enhancements.

Discussion

This is the first study to document the test-retest reliability and SWD for the I-Test and 3-CD in major college football players. The high between-trial correlations and low CV% provide support for the consistency with which Division-I football players are able to perform these tests. The SWDs noted in this study are the first documentation of the relative change necessary in major college players to show a meaningful (i.e., significant) improvement in COD resulting from training programs. The typical error was less than the SWD for all 3 tests, indicating that each was “good” for evaluating the amount of change required to signify real improvement (13). The slightly greater trial-to-trial variability noted for the I-Test was reflected in the greater SWD% required to signify real change in performance level beyond random variation.

The consistency with which players are able to perform these COD tests was slightly better for the 3-CD than for the I-Test, although their typical errors were similar (Table 2). Hopkins (13) noted that the typical error usually contains elements of both biological variation and technical error. In the case of both the I-Test and 3-CD, the technical error could be the variation contributed by hand timing. One approach to removing the technical error would be to use an electronic timing system for these tests as has been done in other studies (1,22,23). However, comparable values for the CV% between COD tests in the current study with those noted when using electronic timing (22) would suggest that timing is most likely a small part of the trial-to-trial variability.

Limb asymmetry could be a factor contributing to the variability noted in both COD tests. However, in a study on American football players, Hoffman et al. (11) observed no significant effect of leg power asymmetry measured from single-leg vertical jumps on 3-CD performance. The authors suggested that perhaps the nondominant leg was a major contributor to deceleration in preparation for the turn in the 3-CD. Lockie et al. (20) noted that the highest correlation between the T-test and isokinetic torque was registered at the slow speed (30° s−1) eccentric motion during leg flexion, which may lead to the suggestion that greater ability to slow running speed (decelerate) to make the COD in an agility test is a major contributor to performance. Interestingly, they found that faster COD times were associated with greater isokinetic torque asymmetries in team sport participants (rugby, basketball, soccer). The authors speculated that players may have used the stronger leg to provide compensation to produce faster acceleration out of a turn. Indeed, the high correlations between 10-yd sprint and COD in the current study would partially support that concept, although leg strength per se was not measured in the present study.

Spiteri et al. (31) noted that stronger subjects had less bilateral deficit (∼6%) in isometric squat strength compared with the weaker subjects (∼9%), but no comparison of COD times with bilateral strength difference was conducted. The authors found that team sport athletes with greater isometric squat strength produced significantly faster velocities on the initial step after a COD. More recent studies by Spiteri et al. (32,33) have emphasized the contribution of relative strength to COD performance. Meylan et al. (22) concluded that COD ability during a simple 10-m, 2-turn test was not significantly different between dominant and nondominant legs determined by maximal jumping tests. To the contrary, Hart and colleagues (8,9) found significantly faster times in a modified version of the Australian Football League's (AFL) agility test for both right-dominant (∼9%) and left-dominant (∼9%) individuals that manifested a 5–10% performance deficit between the 2 directions. The authors felt that the current procedure in AFL combine testing using only 1 direction of the test placed those who were left-leg dominant at a disadvantage. Both these factors could apply to the current study because our players ran only to the right in the 3-CD. Because most players tend to be right-leg dominant (9) and owing to 2 right-leg turns vs. 1 left-leg turn in our 3-CD, it is possible that some players in the current study might have produced times that were as much as 10% slower than their potential ability because the test was run only to the right.

The high correlations of COD tests with the 10-yd sprint in the current study would indicate a strong reliance on acceleration for success in both the I-Test and 3-CD in college football players (Table 2). This finding agrees with some studies (6,29,36) but disagrees with others (18,28,39). Several factors might contribute to this controversy. One potential difference would be the need to differentiate between acceleration (as indicated by a 10-yd sprint) and speed (as measured by a 40-yd sprint). In college football players, acceleration peaks at 10 yds and by 30 yds, it is basically zero, whereas running speed is not changing after 20-yds (5). A second factor would be to emphasize the difference in size (i.e., body mass) between football players and other athletes. Previous work has shown that body mass is negatively correlated with 9.1-m acceleration in football players (4). Muscle strength and power as measured by 1 repetition maximum squat and power clean were correlated with acceleration at 10 yds only when considered relative to body mass (4). Thus, it follows that players with greater body mass have more inertia to overcome when attempting to decelerate, change direction, and accelerate (10), making a greater strength-to-mass ratio desirable.

The relationship between COD and 10-yd sprint time agrees with some studies that suggests that speed and agility are related (7,15,19,21,29) but disagrees with others that concluded that agility is a separate component from speed (18,27,28). This controversy might arise from the various COD tests used by different investigations and the failure to differentiate between acceleration (short sprints) and running speed (longer sprints). The higher association between sprint speed and COD seems to occur when both tests are longer (23), or in COD tasks that have fewer alterations in course (35). An additional problem arises with failure to account for body mass, especially in heavier athletes such as football players. When body mass was removed by partial correlation in the current study, the relationship between straight-ahead acceleration and COD decreased significantly (p < 0.01) for both 3-CD (r = 0.68) and I-Test (r = 0.63), reducing the commonality between speed and COD in football players to less than 46%. Although the I-Test and 3-CD are routinely used in American football, they may not be the most appropriate selections to assess COD in this population. Different positions requiring different movements may require different forms of agility testing. This viewpoint could be controversial, challenging the one-test-fits-all philosophy. Players who are required to change direction primarily while moving forward would probably perform differently than those who initiate movement backward before executing a COD. Thus, receivers and running backs might be more appropriately evaluated by tests such as the AFL test (8,9) or COD with acceleration (19), whereas linebackers and defensive backs could be better evaluated by tests such as the SEMO test (17), a backward 3-CD test, or backward T-test. Offensive and defensive lineman might require yet a different COD test, which has not yet been determined. This might be a controversial issue because some studies have shown a large degree of commonality among various COD tests (26,35), whereas others have not (34).

Practical Applications

Change of direction is a fundamental maneuver in most field sports and thus deserves close attention in the evaluation of players. The I-Test and 3-CD have been used traditionally to assess agility in American football players at all levels, although no sound rationale has been identified supporting their general use. Both tests seem to be measuring a fundamental component of COD, which could suggest a redundancy in their use. However, the generality of each test might not allow specific identification of the particular movements common to different positions in football. Future work might focus on development of position-specific COD tests that better identify player potential.

Acceleration as measured by 10-yd time seems to be an integral part of these COD tests in football players. Previous work has shown that body mass is a factor in COD tests in as much as it impacts the strength-to-body mass ratio (4) and the ability to control deceleration and re-acceleration. Thus, players with a greater strength-to-body mass ratio may produce better COD times because of better control of the deceleration preceding the turn and the ability to reach a higher rate of acceleration out of the turn. This would suggest the need for a comprehensive resistance training program to aid improvement in COD tasks (16).

References

1. Alemdaroglu U. The relationship between muscle strength, anaerobic performance, agility, sprint ability and vertical jump performance in professional basketball players. J Hum Kinet 31: 99–106, 2012.
2. Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 26: 217–238, 1998.
3. Bland JM, Altman DG. Statistical method of assessing agreement between two methods of clinical measurement. Lancet 327: 307–310, 1986.
4. Brechue WF, Mayhew JL, Piper FC. Characteristics of sprint performance in college football players. J Strength Cond Res 24: 1169–1178, 2010.
5. Brechue WF, Mayhew JL, Piper FC, Houser JJ. Comparison between hand- and electronic-timing of sprint performance in college football players. Mo J Health Phys Educ Recreation Dance 18: 50–58, 2008.
6. Chaouachi A, Manzi V, Chaalali A, Wong DP, Chamari K, Castagna C. Determinants analysis of change-of-direction ability in elite soccer players. J Strength Cond Res 26: 2667–2676, 2012.
7. Gains GL, Swedenhjem AN, Mayhew JL, Bird HM, Houser JJ. Comparison of speed and agility performance on field turf and natural grass. J Strength Cond Res 24: 2613–2617, 2010.
8. Hart NH, Spiteri T. Quantifying performance deficits between limbs using a modified version of the AFL agility test. J Aust Strength Cond 21: 105–107, 2013.
9. Hart NH, Spiteri T, Lockie RG, Nimphius S, Newton RU. Detecting deficits in change of direction performance using the preplanned multidirectional Australian Football League agility test. J Strength Cond Res 28: 3552–3556, 2014.
10. Hewitt JK, Cornin JB, Hume PA. Kinetic factors affecting fast and slow straight and change-of-direction acceleration times. J Strength Cond Res 27: 69–75, 2013.
11. Hoffman JR, Ratamess NA, Klatt M, Faigenbaum AD, Kang J. Do bilateral power deficits influence direction-specific movement patterns? Res Sports Med 15: 125–132, 2007.
12. Hopkins WG. Measures of reliability in sports medicine and science. Sports Med 30: 1–15, 2000.
13. Hopkins WG. How to interpret changes in an athletic performance test. Sportscience 8: 1–7, 2004.
14. Hopkins WG. Bias in Bland–Altman but not regression validity analyses. Sportscience 8: 42–46, 2004.
15. Jones P, Bampouras TM, Marrin K. An investigation into the physical determinants of change of direction speed. J Sports Med Phys Fitness 49: 97–104, 2009.
16. Keiner M, Sander A, Wirth K, Schmidtbleicher D. Long-term strength training effects on change-of-direction sprint performance. J Strength Cond Res 28: 223–231, 2014.
17. Kirby RF. A simple test of agility. Coach and Athlete 25: 30–31, 1971.
18. Little T, Williams AG. Specificity of acceleration, maximum speed, and agility in professional soccer players. J Strength Cond Res 19: 76–78, 2005.
19. Lockie RG, Schultz AB, Callaghan SJ, Jeffriess MD, Berry SP. Reliability and validity of a new test of change-of-direction speed for field-based sports: The change-of-direction and acceleration test (CODAT). J Sports Sci Med 12: 88–96, 2013.
20. Lockie RG, Schultz AB, Jeffriess MD, Callaghan SJ. The relationship between bilateral differences of knee flexor and extensor isokinetic strength and multi-directional speed. Isokinet Exerc Sci 20: 211–219, 2012.
21. Mayhew JL, Piper FC, Schwegler TM, Ball TE. Contribution of speed, agility, and body composition to anaerobic power measurement in college football players. J Appl Sports Sci Res 3: 101–106, 1989.
22. Meylan C, McMaster T, Cronin J, Mohammad NI, Rogers C, Deklerk M. Single-leg lateral, horizontal, and vertical jump assessment: Reliability, interrelationships, and ability to predict sprint and change-of-direction performance. J Strength Cond Res 23: 1140–1147, 2009.
23. Munro AG, Herrington LC. Between-session reliability of four hop tests and the agility T-test. J Strength Cond Res 25: 1470–1477, 2011.
24. Robbins DW. Positional characteristics of players drafted into the National Football League. J Strength Cond Res 25: 2661–2667, 2011.
25. Robbins DW. Relationships between National Football League combine performance measures. J Strength Cond Res 26: 226–231, 2012.
26. Robins DW, Goodale T. Evaluation of the physical test battery implemented at the National Football League combine. Strength Cond J 24: 1–10, 2012.
27. Salaj S, Markovic G. Specificity of jumping, sprinting, and quick change-of-direction motor abilities. J Strength Cond Res 25: 1249–1255, 2011.
28. Sassi RH, Dardouri W, Yahmed MH, Gmada N, Mahfoudhi ME, Gharbi Z. Relative and absolute reliability of a modified agility T-test and its relationship with vertical jump and straight sprint. J Strength Cond Res 23: 1644–1651, 2009.
29. Sheppard JM, Young WB, Doyle TLA, Sheppard TA, Newton RU. An evaluation of a new reactive agility and its relationship to sprint speed and change of direction speed. J Sci Med Sport 9: 342–349, 2006.
30. Sierer SP, Battaglini CL, Mihalik JP, Shields EW, Tomasini NT. The National Football combine: Performance differences between drafted and nondrafted players entering the 2004 and 2005 drafts. J Strength Cond Res 22: 6–12, 2008.
31. Spiteri T, Cochrane JL, Hart NH, Haff GG, Nimphius S. Effect of strength on plant foot kinetics and kinematics during a change of direction task. Eur J Sport Sci 13: 646–652, 2013.
32. Spiteri T, Newton RU, Binetti M, Hart NH, Sheppard JM, Nimphius S. Mechanical determinants of faster change of direction and agility performance in female basketball athletes. J Strength Cond Res 29: 2205–2214, 2015.
33. Spirteri T, Nimphius S, Hart NH, Specos C, Sheppard JM, Newton RU. Contributions of strength characteristics to change of direction and agility performance in female basketball athletes. J Strength Cond Res 28: 2415–2423, 2014.
34. Sporis G, Jukic I, Milanovic L, Vucetic V. Reliability and factorial validity of agility tests for soccer players. J Strength Cond Res 24: 679–686, 2010.
35. Stewart PF, Turner AN, Miller SC. Reliability, factorial validity, and interrelationships of five commonly used change of direction speed tests. Scand J Med Sci Sports 24: 500–506, 2014.
36. Vescovi JD, McGuigan MR. Relationship between sprinting, agility, and jump ability in female athletes. J Sports Sci 26: 97–107, 2008.
37. Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res 19: 231–240, 2005.
38. Young W, Farrow D. A review of agility: Practical applications for strength and conditioning. Strength Cond J 28: 24–29, 2006.
39. Young WB, James R, Montgomery I. Is muscle power related to running speed with change of direction? J Sports Med Phys Fitness 42: 282–288, 2002.
Keywords:

I-test; 3-cone drill; sprint acceleration; performance evaluation

Copyright © 2016 by the National Strength & Conditioning Association.