The 40-yd sprint has become the most widely used method for evaluating sprint speed among athletes in various sports. It is a key component of any test battery in American football (2,5,11,13,20,21) and a significant determinant of player potential in the National Football League (NFL) (18,19). Players may be timed by electronic systems or by individuals with differing levels of expertise using hand-held stopwatches. The difference between hand-held stopwatches and electronic systems seems to vary widely, with some studies showing as little as 0.08-second to as much as 0.31-second difference between the 2 methods (3,10,15,16). The study that noted only a 0.08-second average difference between hand timing and electronic timing (16) used a standing start and initiated the electronic timing system when the front thigh passed through a photocell beam. Therefore, while the hand timers started their watches on the first movement of the subject, the potential delay in activating the electronic system may have accounted for the smaller difference between timing methods (7,16). Most of the timers in previous studies have been experienced in the use of stopwatches. A factor, which has not been explored fully, is the effect of the level of experience of the timers on the accuracy of measurement of short sprints.
Given the importance of the 40-yd sprint in evaluating a player's ability and potential in football at the major college level, it would be important to assess the difference to be expected between hand timing and electronic timing, as well as the difference between experienced timers and novice timers. In addition, establishing the reliability and smallest worthwhile difference (SWD) of each method would provide the strength and conditioning specialist with a standard by which to gauge significant changes in sprint performance. Therefore, the purpose of this study was to evaluate the validity of hand timing by experienced and novice timers compared with electronic timing and to establish the reliability and SWD of each method for the 40-yd sprint in college football players.
Experimental Approach to the Problem
The 40-yd sprint is an essential element in the evaluation of American football players and a fundamental component of the NFL combined for determining professional potential (22). Given the importance placed on 40-yd sprint time for assessing player ability at the college level and determining future professional status, slight variations in timing could mean the difference in selection by specific college teams or might involve a significant amount of money in professional salary (14). This study sought to evaluate the variability and degree of difference between hand timing and electronic timing, with the additional observation on the effect of experience in the use of a hand-held stopwatch on the accuracy of timing. Major college players were assessed for 40-yd sprint speed at the conclusion of their winter conditioning period using both electronic timing and hand-held stopwatch timing by experienced and inexperienced timers.
The 40-yd sprints were timed by experienced and novice timers. The experienced timers consisted of 2 strength and conditioning coaches with over 20 years of experience, timing the sprint and agility performance of college football players and other college athletes. The 4 novice timers were prephysical therapy students who have never timed running speed performance and had only a few minutes to practice using a stopwatch before actual timing.
Eighty-one players (age = 20.5 ± 1.2 years; height = 184.9 ± 6.6 cm; weight = 102.5 ± 18.1 kg) from an elite NCAA Division I team volunteered to participate. Measurements were conducted at the conclusion of an off-season conditioning program immediately before spring practice. Each player had undergone extensive resistance training, as well as speed and agility drills. The testing program was part of the regular training procedures for the team, and all players signed a consent document to participate. No players under the age of 18 years were included in the study.
Players performed designated group warm-ups that included light jogging, moderate-paced high-knee drills, and various stretching exercises. After the group warm-ups, players were free to perform additional individual stretching and movement activities. Players wore a multicleat turf shoe, gym shorts, and tee shirt during testing.
All sprint trials were performed on an indoor artificial turf with the distance measured with a steel tape. Each player ran 2 trials separated by a minimum of 5-minute rest. Starting position was a 3-point stance (1 hand on the ground). Electronic timing used a touch pad start (ground hand), and an infrared beam stop placed 75 cm above the ground (SpeedTrap, Model II; Brower Timing Systems, Draper, UT, USA). Each trial was also timed by 2 experienced and 4 novice timers using hand-held stopwatches (Model SC-5-5; Robic Inc., Orange, CA, USA). Timers were stationed perpendicular to and 6 m from the finish line. Timers were instructed to start their watch on first movement of the player and stop their watch coincident with the upper-body crossing the finish line. There was no communication among timers during the testing.
Repeated-measures analysis of variance was used to evaluate the differences among hand timers and for comparing hand timing with electronic timing. Post hoc testing was performed using the Bonferroni method. Trial-to-trial reliabilities for hand and electronic timing were assessed using intraclass correlation coefficients (ICCs) as suggested by Weir (23). Coefficient of variation (CV%) was determined by dividing the SDbetween trials by the mean of the 2 trials. The SWD was determined as 1.65 × typical error × √2 (12). Smallest worthwhile difference percentage was calculated by dividing the SWD by the mean of successive trials. Effect size (ES) was determined using Cohen's d statistic (4).
There were no significant differences between trials 1 and 2 whether timed electronically (5.16 ± 0.36 and 5.14 ± 0.35, respectively; p = 0.93) or by hand (4.92 ± 0.33 seconds and 4.91 ± 0.34, respectively; p = 0.84). One novice timer had significantly faster times than each of the other novice timers (Table 1). Removing the faster times of this timer and averaging the remaining 3 novice timers' scores did not alter observations as there was still no significant difference between trial 1 (4.95 ± 0.35 seconds) and trial 2 (4.94 ± 0.34 seconds). Likewise, there was no significant difference between trials 1 and 2 for either experienced timers (Table 1). The ESs between trials 1 and 2 were considered trivial for novice timers (ES = 0.06), experienced timers (ES = 0.08), and electronic timing (ES = 0.06). Reliability for experienced timers (ICC = 0.972) was slightly higher than for novice timers (ICC = 0.934).
Given that there was no difference between trials and owing to the small ES, the fastest time from the 2 trials was used for further analysis. The average of the best hand times for the 2 experienced timers (4.90 ± 0.34 seconds) was not significantly different from the average of the 4 novice timers (4.86 ± 0.33 seconds), with a very high correlation between them (r = 0.98; Figure 1).
The average of the best times for all hand timers (4.87 ± 0.33 seconds) was significantly faster than that observed with electronic timing (5.12 ± 0.35 seconds) by 0.25 ± 0.08 seconds (ES = 0.71), but there was a consistent relationship between the 2 times (ICC = 0.985; 95% CI = 0.985 to 0.994; Figure 2). The average of best hand times for the 2 experienced timers (4.90 ± 0.34 seconds) was significantly faster than the fastest electronic time (5.12 ± 0.35 seconds) by 0.22 ± 0.07 seconds (ES = 0.63), with a consistent relationship between the 2 (ICC = 0.988). The average of best hand times for the novice timers was significantly faster than the fastest electronic time by 0.26 ± 0.08 seconds (ES = 0.74), with a consistent relationship between the 2 (ICC = 0.987).
Reliability between the 2 electronic trials was comparable with the reliability between the hand-timed trials for all timers with similar total error, CV%, and 90% limits of agreement (Table 2). The SWD to indicate a meaningful change in 40-yd sprint time was 0.12 seconds for electronic timing and 0.14 seconds for hand timing (Table 2).
Fifty-one percent of the players produced the fastest electronic time during the second trial, whereas 56% produced the fastest time with hand timing. There was a significant difference (p = 0.008) between hand timing and electronic timing when comparing the slowest 30% of players (−0.28 ± 0.08 seconds) and the fastest 30% of players (−0.22 ± 0.05 seconds) (Figure 3). Thus, the slower the player, the greater the tendency for hand time to underestimate electronic time.
The major finding in this study was that there is not likely to be a significant difference between experienced and novice timers when timing 40-yd sprint in college football players. Hand timing in the 40-yd sprint is likely to produce significantly faster times than with electronic timing regardless of the experience of the timers. Electronic timing of the 40-yd sprint is considered the most accurate method of evaluating sprint speed in college football players, and therefore, based on the present observations, hand timing would be considered invalid as it produced consistently faster than actual times. However, the high ICCs indicated that hand timing produces highly reliable data between trials by both experienced and inexperienced timers and with electronic timing. Furthermore, the high rank-order correlation between hand times and electronic times (rho = 0.98) indicated that players would have been categorized correctly regardless of timing method. Therefore, the ability to accurately categorize players by speed (fastest to slowest) seems intact with hand timing.
Interestingly, the magnitude of the difference between electronic and hand timing in this study was similar to the average individual's hand reaction time 0.22 seconds reported by Haywood and Teeple (9). The difference between hand and electronic timing observed here was slightly less than that noted previously between 2 experienced timers and electronic time (−0.27 ± 0.06 seconds) when sprinting on an outdoor turf field (6) but significantly greater than the difference (−0.16 ± 0.12 seconds) noted when players ran indoors on a rubberized floor (1,3). In another study where players ran on an indoor synthetic running track (15), 7 experienced timers had a greater difference (−0.31 ± 0.07 seconds) when compared with electronic timing. If the values from all these studies are pooled, the average difference to be expected between hand timing and electronic timing would be −0.25 ± 0.09 seconds, which approaches the widely accepted value (−0.24 seconds) for differences between hand and electronic timing (17).
Haugen et al. (7) have indicated that the major factor contributing to the difference between hand and electronic timing is most likely involved with the start. They noted that electronic timing may actually start 0.04 seconds before visual hand movement from a starting pad can be detected by video analysis. In this study, the difference between electronic and hand time was greater in the slower runners (Figure 2). The slower players tend to be the bigger such that body movements may occur while the hand remains on the starting pad, possibly resulting in hand timing starting sooner than electronic timing (7). Body movement in the smaller players might tend to be more coincidental with lifting the hand off the starting pad as their initial movement, which is likely to be more noticeable to the timers. In a separate study, Haugen et al. (8) noted that different infrared beam height configurations at the finish could affect electronic clock stopping time. Although timers in this study were instructed to stop their watches when the chest passed the finish line, there could be variation in when they perceived that to occur. Thus, further study in which the differences between hand and electronic timing at both the start and finish are evaluated might provide more definitive information on where the major discrepancy lies between the 2 techniques.
The important aspects of evaluating speed are differentiating performances among individuals, tracking changes after training or injury, and judging readiness for competition. Despite the universal use of the 40-yd sprint as an indicator of speed in football, little is known about the test-retest reliability and consequently the level of change, which might be classified as physiologically meaningful after training. The SWD has become the method for discriminating physiological difference based on trial-to-trail variation and learning effects (12). Although absolute reliability was high (trial-to-trial ICCs >0.98) and ES was small (Cohen's d value <0.11), calculation of the SWD provides a number by which strength and conditioning specialists can judge physiologically significant changes resulting from training. The SWD for both electronic and hand timing in this study (0.12 electronic time and 0.14 hand time) is comparable with the 0.14 seconds observed in Division II players (6). Hopkins (12) has suggested that %SWD is better to estimate the desired level of improvement necessary from training because of individual differences in the level of performance. This would suggest that the faster players in this study would need to improve their 40-yd time by 0.11 seconds, whereas the slower players would need to improve by 0.13 to indicate a meaningful change. Thus, interpretation of physiologically meaningful change might differ according to the level of ability of the player but would be independent on timing method.
To our knowledge, this study is the first to compare the abilities of experienced and novice timers for recording sprint times. The lack of significant difference between the 2 levels of experience and the comparable relationship of each with electronic timing suggest that newcomers can learn the technique quickly. Further study might be interesting to isolate the components of the difference between hand and electronic timing (i.e., start vs. finish) to determine whether further refinements can be made. In addition, an evaluation of the NFL combine technique would be of interest. This technique uses a manual start initiated by an individual positioned approximately 10 yd in front of the player and an infrared stop. Although this might remove any anticipation of timers at the finish, the reaction time of the starter might still be a factor in establishing a correct time. Owing to the weight placed on player speed at all positions in the NFL, slight discrepancies in 40-yd sprint time could amount to big changes in employment status and salary in the professional ranks.
Hand timing, regardless of the experience of timers, allows for an accurate classification (ranking) of sprint performance among college football players, although the times may be significantly faster than electronic timing. Furthermore, statistical assessment of test-retest data should be performed on other teams to evaluate the consistence of the SWD observed in this study. Smallest worthwhile difference allows for an understanding of real changes (as opposed to random variations) and an assessment of the effect of various training procedures for improving speed. Hand timing, while less sophisticated than electronic timing, does not seem to significantly impact the SWD or the inferences of player ranking from timed performance.
1. Brechue WF, Mayhew JL, Piper FC. Equipment and running surface alter sprint performance of college football players. J Strength Cond Res 19: 821–825, 2005.
2. Brechue WF, Mayhew JL, Piper FC. Characteristics of sprint performance in college football players. J Strength Cond Res 24: 1169–1178, 2010.
3. Brechue WF, Mayhew JL, Piper FC, Houser JJ. Comparisons between hand- and electronic-timing of sprint performance in college football players. Mo J Hlth Phys Educ Rec Dance 18: 50–58, 2008.
4. Cohen J. Statistical Power Analysis for Behavioral Sciences (2nd ed.). Mehwah, New Jersey: Lawrence Erlbaum Associates, 1988.
5. Dupler TL, Amonette WE, Coleman AE, Hoffman JR. Anthropometric and performance differences among high school football players. J Strength Cond Res 24: 1975–1982, 2010.
6. Gains GL, Swedenhjelm AN, Mayhew JL, Bird HM, Houser JJ. Comparison of speed and agility performance of college football players on field turf and natural grass. J Strength Cond Res 24: 2613–2617, 2010.
7. Haugen TA, Tonnessen E, Seiler SK. The difference is in the start: Impact of timing and start procedure on sprint running performance. J Strength Cond Res 26: 473–479, 2012.
8. Haugen TA, Tonnessen E, Svendsen IS, Seiler SK. Sprint time differences between single- and dual beam timing systems. J Strength Cond Res 28: 2376–2379, 2014.
9. Haywood KM, Teeple JB. Representative simple reaction and movement time series. Res Quart 47: 855–856, 1976.
10. Hetzler RK, Stickley CD, Lundqist KM, Kimura IF. Reliability and accuracy of handheld stopwatches compared with electronic timing in measuring sprint performance. J Strength Cond Res 22: 1969–1976, 2008.
11. Hoffman JR, Ratamess NA, Kang J. Performance changes during a college playing career in NCAA Division III football athletes. J Strength Cond Res 25: 2351–2357, 2011.
12. Hopkins WG. Measures of reliability in sports medicine and science. Sports Med 30: 1–15, 2000.
13. Jacobson BH, Conchola EG, Glass RG, Thompson BJ. Longitudinal morphological and performance profiles for American, NCAA Division I football players. J Strength Cond Res 27: 2347–2354, 2013.
14. McGee KJ, Burkett LN. The national football league combine: A reliable predictor of draft status? J Strength Cond Res 17: 6–11, 2003.
15. Mayhew JL, Houser JJ, Briney BB, Williams TB, Piper FC, Brechue WF. Comparison between hand and electronic timing of 40-yd dash performance in college football players. J Strength Cond Res 24: 447–451, 2010.
16. Moore AN, Decker AJ, Baarts JN, DuPont AM, Epema JS, Reuther MC, Houser JJ, Mayhew JL. Effect of competitiveness on forty-yard dash performance in college men and women. J Strength Cond Res 21: 385–388, 2007.
17. Podkaminer B. Fully automatic conversion. In: 2014/2015 NCAA Men's and Women's Track and Field and Cross Country Rules. Seewald R., ed. Indianapolis, IN: National Collegiate Athletic Association, 2014. p. 86.
18. Robbins DW. The National Football League (NFL) combine: Does normalized data better predict performance in the NFL draft? J Strength Cond Res 24: 2888–2899, 2010.
19. Robbins DW. Relationships between National Football League combine performance measures. J Strength Cond Res 26: 226–231, 2012.
20. Robbins DW, Goodale TL, Kusmits FE, Adams AJ. Changes in the athletic profile of elite college American football players. J Strength Con Res 27: 861–874, 2013.
21. Robbins DW, Young WB. Positional relationships between various sprint and jump abilities in elite American football players. J Strength Cond Res 26: 388–397, 2012.
22. Sierer SR, Battaglini CL, Mihalik JP, Shields EW, Tomasini NT. The National Football League Combine: Performance differences between drafted and nondrafted players entering the 2004 and 2005 draft. J Strength Cond Res 22: 6–12, 2008.
23. Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. J Strength Cond Res 19: 231–240, 2005.