Timing light systems are becoming increasingly popular as a means of assessing acceleration and speed (2) by measuring the time to cover certain distances. A timing light consists of a light source (transmitter-see Figure 1) and an optical pick-up (receiver). Visible, infrared, standard, or monochromatic laser lights are the usual light source used. The light source is aimed across the track at the reflector and bounced back to the receiver (4). This forms a “gate” that the athlete will pass through. The timing light systems can vary from single, dual, and three-beam reflector units. Research by Swift Performance Equipment (3) has shown that single beam systems can be triggered early by a swinging arm, and may produce measurement error of up to 80 ms. The research has also proven that a three-beam system is not necessary, as accuracy improvement is less than the system resolution of 0.01 second.
Although the use of timing lights is wide spread, the descriptions of their use, in particular methodologies, are generally very limited. For example the start distance of the athletes behind the lights, foot position, first step strategy and height of timing light gates are all factors that will affect the sprint assessment. However, most research fails to describe these factors in any great detail, which affects the ability to compare results between studies. In terms of timing light height, two heights (shoulder and hip height) are commonly used (5). The effect of timing light height on sprint times has not yet been detailed. A better understanding of the magnitude of such effects should improve standardization of sprint assessment protocols and the comparability of results between studies. The purpose of this study therefore was to examine the reliability and effect of timing light height on sprint times.
Experimental Approach to the Problem
A sample of convenience was used to investigate whether the height at which timing lights were set affected sprint times. The subjects were tested simultaneously over 2 distances (10 and 20 m) at 2 heights (60 and 80 cm) and reliability statistics were estimated.
The sample consisted of 9 male and 6 female sport and recreation university students (age 22.7 ± 3.6 yrs, height = 1.72 ± 0.09 cm, mass 71.8 ± 12.2 kg). The subjects were active, of an athletic background and were recruited from a sample of convenience. The Human Subject Ethics Committee of the Auckland University of Technology approved all procedures undertaken and all subjects signed an informed consent form.
A dual-beam (Swift Performance Equipment, Lismore, Australia) modulated visible red light system with polarizing filters was used to measure sprint times to 1/100th of a second. Two hand held computers recorded the times of 2 sets of timing lights simultaneously.
Following a standardized warm-up, including 10 minutes of light aerobic exercise (e.g., running), dynamic stretching, and familiarization with the sprint protocol, the athletes performed 6 trials. For the first 3 trials, the gates were set at 10 m at the low (60 cm) and high (80 cm) heights (5). For the second 3 trials the gates were shifted to 20 m and times recorded at both heights (Figure 1). The participant was required to start with 2 feet on a line 30 cm behind the first gate. In their own time, the participant sprinted through the timing gates' ensuring the first step of the take off was forward. The participant was instructed to sprint past a cone placed 5 m after the last gate. Rest intervals between sprints were approximately 120 seconds.
The data were analyzed using SPSS 11.0 statistical software (SPSS, Chicago, IL). The mean of 3 trials was used for analysis of all variables. Data are presented as mean values and standard deviations. Within-trial reliability of both the high and low sprint assessments was assessed using a coefficient of variation (CV = SD/Mean × 100). The differences between the low and high gates at 10 and 20 m were calculated (Average [low-high score]). Data was log-transformed to determine the differences and typical errors between the low and high gates at 10 and 20 m. Paired t-tests were used to determine whether significant differences existed between means at the 2 different heights at the 10 and 20 m distances. An alpha level of 0.05 and 95% confidence limits (CL) were used for statistical procedures where appropriate.
The mean and standard deviations for times at the 2 gate heights can be observed in Table 1. The 20 m times were approximately 1.69 times greater than their 10 m times indicating an increase in velocity over the second 10 m. The within trial variability (CV) for each of the gate heights ranged from 0.69 to 1.2%, with the greatest variability found over the shorter distances.
Mean differences in the time between the 2 light heights at both distances were found to be 0.07 seconds (95% CL = 0.05-0.10 second). The difference between the means for the log-transformed data was greater for the 10 m distance (3.4%, 95% CL = 2.5-4.5%) as was the typical error represented as a coefficient of variation (1.3%, 95% CL = 0.9-2.0%). The low gates were found to record significantly faster times than the high gates at both distances.
To determine the measurement error associated with the timing light assessment a number of reliability measurements were used. A coefficient of variation (CV) was calculated to determine the stability of measurement between the trials on the same day. Such an approach is common in the literature. Some scientists have arbitrarily chosen an analytical goal of the CV being 10% or below, but the merits of this value are the source of conjecture (1). Nevertheless, the within-trial variability across all conditions was small (1.2%) with less variation associated with the longer distances. The practical application of such findings for sprint assessment is that in cases where large numbers of subjects/athletes are being assessed and time is a constraint, only a small number of trials (1 or 2) are required to gather reliable information. Furthermore in research paradigms where many conditions are being compared, order, fatigue, and motivation effects may confound results. In such circumstances, the use of a small number of trials would appear acceptable to gather reliable information. Finally, given the difficulty in making even small improvements in acceleration and speed, having measures this consistent is beneficial.
Greater percent changes between light height times were found at the shorter distance (3.4 vs. 0.8%). The typical error was also greater (1.3%) for the 10 m distance. As subjects were timed simultaneously with the exact same timing light systems, the percent change and error can no doubt be attributed to technological error (timing light height) rather than biological error. The lower typical error and percent change at the longer distance can most probably be explained by the greater variation in body position over the shorter distance, running posture becoming more consistent as the body becomes more upright at the 20 m distance.
The difference in means between the high and low timing lights at 10 and 20 m was 0.07 second, which was found to be significant. As intimated previously, sprint times from the low and high timing gates were measured simultaneously, eliminating the possibility of biological variation due to the subjects varying performances. Therefore the differences must be attributed to different body parts breaking the lights at different times. It was thought that the timing lights set at shoulder height would result in the faster times due to the lean of the body in the acceleration phase. However, this was not the case: the lower timing light resulting in quicker times. It would seem that the legs break the beams prior to the upper body, this difference decreasing with distance as the body becomes more upright.
Small within-trial variation and typical errors would suggest that acceleration and speed assessment using low and high timing lights heights are reliable. However, the lower gate height was found to have significantly quicker times. Practitioners using timing lights need to be aware of such methodological issues especially when comparing results between studies. It is suggested that a standardized protocol using timing lights be developed if comparison of athletes is of interest to researchers, sport scientists, and coaches.
1. Atkinson, G and Nevill, AM. Statistical methods for assessing measurement error (reliability
) in variables relevant to sports medicine. Sports Med
26: 217-238, 1998.
2. Graham-Smith, P. How to use timing lights. www.ukathletics.net
. Accessed May 9, 2005.
3. Swift Performance Equipment. Available at: http://www.spe.com.au/swift_bu/speed-li.htm
. Accessed May 15, 2005.
4. Westenburg, TM. Timing light accuracy in sport. In: The engineering of sport: design and development: Blackwell Science
. Haake, S, ed. Oxford, U.K.: Blackwell Publishing, 1998. pp. 291-299.
5. Yeadon, MR, Kato, T, and Kerwin, DG. Measuring running speed using photocells
. J Sport Sci
17: 249-257, 1999.