The study sample consisted of 102 professional male basketball players (height: 194.92 ± 8.09 cm; body mass: 89.33 ± 10.91 kg; 21.58 ± 3.92 years of age) from Bosnia and Herzegovina. During the study course, players were members of the teams competing in the two highest national divisions (i.e., first (n = 58) and second divisions (n = 42)). The players were classified, as reported by coaches, based on their primary playing position, into three groups: guards (n = 48), forwards (n = 22), and centers (n = 32). Testing was organized at the beginning of the basketball season, after the preseason period, and only players without any injuries or illness in the 30 days preceding the testing participated in the study (players completed a health history questionnaire before testing). All players were practicing basketball for at least seven years and their usual weekly training regime consisted of 5 basketball-specific technical–tactical sessions (1–4 h per day), 2–3 strength and conditioning sessions, and 1 or 2 competitive games. Eligibility criteria included: minimum of 8 years of systematic basketball training, no injuries and/or illnesses for 20 days before the testing, and regular participation in training/games during the last two weeks. The health status was based on the team physician’s opinion/report. No players were taking substances that might be expected to affect their performance on study tests. Approval for the research experiment was provided by the Ethical Board of the University of Split, Faculty of Kinesiology (No: 2181-205-02-05-14-001).
Variables in this study included (i) participants’ general information (age, playing position, and performance level), (ii) anthropometrics, and measures of (iii) one-leg jumping and (iv) agility capacities.
The anthropometric variables included body height (BH), maximal reach height, body mass (BM), and body fat (BF). Standardized stadiometers, scales (Seca, Birmingham, UK), and a skinfold caliper (Holtain, London, UK) were used for measurement. Both height assessments were conducted barefoot, with participants extending their dominant arm as high as possible in maximal reach height MRH. Biceps, triceps, subscapular, and suprailiac skinfolds were collected and used in the formula for body density (BD), which was later used for the purpose of the calculation of BF% as follows [21
BD = 1.162 − 0.063 × log Σ4skinfolds; BF% = (4.95/BD − 4.5) × 100
The same investigator (observer) measured skinfolds for all tested participants in order to minimize unsystematic measurement errors. All measurement procedures were obtained in accordance with the International Biological Program, with the observer’s reliability levels’ ICCs ranging from 0.78 (for suprailiac skinfold) up to 0.99 (for body height) [22
Jumping and agility testing was conducted in an indoor gymnasium with a wooden floor. Conditions were similar for all participants and included one day of rest before measurements, a standard basketball floor, temperatures of 20–25 °C, a self-preferred type of footwear, and the time of day (between 9 and 11 AM). Before measurement, all players completed a 15 min warm-up protocol that consisted of jogging (5 min), mobility exercise (5 min), dynamic stretching (3 min), and activation (2 min), which included skipping and light jumping. Before both jumping and agility tests, players were given three trials in order to familiarize themselves with all tasks. Participants first performed an assessment of one-leg vertical jumping capacities that included: (i) maximal running vertical jump (running jump) with take-off from the right and left leg, and (ii) two-step approach vertical jump (lay-up jump) with take-off from the right and left leg. All tests were conducted with a VERTEC apparatus (Vertec, Sports Imports, Hilliard, OH, USA).
In running jump, participants were instructed to perform a self-chosen running start in the 5 m marked space, to take off from their left or right leg, and to try reaching a maximal height with their extended arm. The technique of the jump was not predetermined, as participants performed it in the subjectively most appropriate way, imitating the real game situation. The final result of the test was calculated as the difference in centimeters between the reached height after the jump and the participants’ standing vertical reach. The lay-up jump was measured in a similar way but with participants performing a typical basketball two-step approach (lay-up) before the jump. All participants conducted three trials for each jump with a one-minute rest between the trials, and the highest result was taken as final in each particular test. The leg with the better achieved final result was noted as dominant. The reliability and validity of the jumping tests used in this study were previously reported to be appropriate to high (ICC: 0.80–0.85 and 0.86–0.88 for intertesting and intratesting reliability, respectively) [13
After jumping assessments, agility capacities were measured with a basketball-specific RAG test performed on the dominant and nondominant sides, and a CODS basketball-specific agility test performed on the dominant and nondominant sides. The RAG and CODS were created with the aim of testing basketball-specific agility, and the test design is presented in Figure 1
The measurement system consisted of an ATMEL microcontroller (model AT89C51RE2; ATMEL Corp, San Jose, CA, USA), a photoelectric infrared (IR) sensor (E18-D80NK), and LEDs placed in 30 cm high cones. The system was connected to a laptop PC with a Windows 7 operating system. The reliability and validity of the tests have previously been evaluated and confirmed on similar participants with intertesting reliability obtained by ICCs ranging from 0.81 to 0.90 [16
], so in this study, we reported only intratesting reliability for the observed measurement (see below for details).
In the RAG tasks, the participant runs from a starting position and lights up one of the cones after crossing the IR signal positioned 1 m from the start, as shown in Figure 1
. After recognizing which cone lit up, the participant needs to react in the correct direction and run to that one and rebound the ball placed at the top of the cone with his arms. After that, in order to complete a successful attempt, the participant needs to return to the starting position as quickly as possible, and the timing stops after crossing the IR signal on the way back. Participants had to perform five attempts in each of three trials. In the CODS tests, participants had the same task, but participants knew in advance which cone would light up and performed two attempts, one on the right and one on the left side, and they repeated it in three trials. The shortest time was registered as the final result in all tests, and the shorter time needed to complete the task was characterized as the dominant side performance.
There are multiple methods used to calculate symmetry levels and the selection depends on a number of factors [23
]. Although the percentage difference method has recently been suggested as most appropriate to estimate asymmetry levels, we used Limb symmetry index 1, which is calculated by dividing the performance from the nondominant leg/side (note that in RAG and CODS testing, we speak about the “dominant side,” and in jumping performances, we speak about the “dominant leg”) by the corresponding performance from the dominant leg/side [23
]. Actually, limb symmetry index 1 is more of a measure of limb symmetry than the asymmetry, and the results of the percentage difference method, although seemingly different, show values at the opposite end of the asymmetry spectrum [25
]. This allowed us to clearly present the interlimb difference for jumping, RAG, and CODS performance. In all cases, the value closer to 1 presented a better symmetry of performance (note that for jumping, we divided the nondominant performance by the dominant performance, and vice-versa for agility). Theoretically, the ideal symmetry of performance on the dominant and nondominant sides was noted by a numerical value of 1 (100%).
2.4. Statistical Analyses
Normality was checked via the Kolmogorov–Smirnov test, and descriptive statistics included means and standard deviations.
While the observed tests of one-leg vertical jumps (lay-up jump and running-jump) and agility performances (CODS and RAG) were previously extensively studied for reliability and validity on similar samples of participants, and results are presented elsewhere [13
], herein, we calculated and interpreted only the intrasession reliability of the tests via the intraclass correlation coefficients (ICCs) and coefficients of variation (CVs).
Factorial analysis of variance with playing position (guards, forwards, and centers) and performance levels (1st division vs. 2nd division) as main effects, and the interaction (playing position × performance level) with Scheffe’s post-hoc analysis was calculated to identify the association between the symmetry in performance of running jumps and agility, with playing positions and performance levels. Partial eta squared (η2) was calculated to identify the effects size (ES) and was interpreted accordingly (small ES: >0.02; medium ES: >0.13; large ES: >0.26).
For all calculations, Statistica 13.5 (TIBCO Software Inc., Palo Alto, CA, USA) was used. The significance level was set at p < 0.05.