One of the primary exercise modalities used in the fitness industry since the new millennium has been the elliptical cross-trainer (ECT), and it has been estimated that the total number of individuals who use ECTs has increased 220% between 2000 and 2007 (16). One of the major reasons for the increased use of ECTs is the notion that these exercise modalities offer the user a low-impact option to traditional treadmills (TML), stairclimbers, and overground running in which increased training periods can occur with less exertion placed on the low back and lower-extremity joints (i.e., hip and knee). These perceptions are commonly directed toward individuals who suffer from low back pain, lower-extremity joint pathology (i.e., osteoarthritis) or are overweight/obese. Past research investigating ECTs have found comparable physiologic responses at maximal exercise (3) and submaximal exercise intensities (2) when compared with exercising on a TML. With reduced impact at the hip, knee, and ankle joints, users of the ECT likely perceive that increased workout durations can be accomplished with less joint discomfort, which would result in positive training adaptations. However, an investigation by Green et al. (5) confirmed the opposite because greater ratings of perceived exertion (RPE) for the legs during ECT exercise was observed when compared with TML exercise. Moreover, a more recent study (12) demonstrated greater hip flexor and knee extensor moments to exist when exercising with an ECT compared with walking on a TML. These results were believed to occur because of the smaller loading rates evident at heel strike as well as the greater slope of the pedal as it traveled along its elliptical path.
Cybex International, Inc., has recently manufactured a new exercise modality, the Cybex Arc Trainer (ARC), that uses a low-impact mechanical design. Unlike traditional ECTs that use an elliptical path of motion during exercise, the ARC uses an arc path of motion in which the focus of the design is to more closely simulate lower-extremity joint kinematics that are more natural to human walking/running gait. In addition, the manufacturer claims that this design also allows for less shear forces to be placed on the lower-extremity joints. Thus, these design characteristics should result in improved exercise efficiency during maximal and submaximal exercise while subjects perceive less whole body exertion and experience less discomfort at the low back, hip, and knee joints. Because of the popularity of the ECTs (16), we wanted to determine how the ARC compared with the more popular ECT and the with regard to cardiorespiratory variables and perception of discomfort at commonly prescribed exercise intensities for the general population (55%, 65%, and 75% of maximal oxygen uptake [O2max]) and during a O2max test. Therefore, the purpose of this psychophysiologic study was to 1) compare the oxygen cost and heart rate (HR) responses at maximal exercise intensities between the ECT, ARC, and TML exercise modalities and 2) investigate the cardiorespiratory responses and perceived discomfort of the low back, hip, and knee joints when exercising at 3 different submaximal intensities.
Experimental Approach to the Problem
Two separate repeated measures designs were used to guide this investigation. For the first purpose of this study, a 1 × 3 design was used in which the single independent variable was exercise modality with 3 levels: ECT, ARC, and TML. The dependent variables compared between each exercise modality were O2max (ml·kg−1·min−1), maximal HR (HRmax, bpm), and time to attain O2max (min). For the second purpose of this study, we used a 2 × 3 fully repeated measures design where the first independent variable was exercise modality with 2 levels (ECT and ARC), and the second independent variable was exercise intensity with 3 levels (55%, 65%, and 75% of TML O2max). The dependent variables compared between each exercise modality and exercise intensities were oxygen consumption (O2, ml·kg−1·min−1), HR (bpm), minute ventilation (VE, L·min−1), respiratory exchange ratio (RER, O2/O2), whole body RPE (6-20), and perceived pain and discomfort (numerical rating scale-11 for pain [NRS-11], 0-10 = none to extreme discomfort) at the low back, hip, and knee joints.
Eighteen apparently healthy subjects (10 male and 8 female; age = 24.7 ± 2.6 yr, height = 172.2 ± 10.3 cm, mass = 69.8 ± 14.9 kg, %fat = 22.5 ± 8.1%) recruited from the campus community participated in this investigation. All subjects were free of any cardiovascular, pulmonary or musculoskeletal pathology that would have negatively affected their participation in this study. Each subject was required to visit the Exercise Physiology Research Laboratory at approximately the same day of week and same time of day for 5 consecutive weeks to complete the study. Before enrolling in this study, all subjects were explained the study requirements for participation and giving time for any questions to be answered. Afterward, written informed consent was obtained pursuant to law from the subject, and a medical health history questionnaire was completed. All study procedures were performed according to the human experimentation policy statement outlined by the American College of Sports Medicine. In addition, this study was approved by the university's institutional review board.
Subjects performed a graded maximal exercise test using an industry standard TML (Quinton 1000 Treadmill, Cardiac Science Corporation, Bothell, WA, USA). Subjects performed maximal and submaximal exercise tests on the Cybex ARC (Cybex International, Inc., Medway, MA, USA) and on a LifeFitness X9i ECT (LifeFitness, Schiller Park, IL, USA). To measure respiratory gases and metabolic data, subjects wore a noseclip and mouthpiece connected to a 1-way breathing valve, with all expired gases analyzed by a ParvoMedics TrueOne Metabolic System (ParvoMedics, Sandy, UT, USA). Last, all body composition assessments were made using a BodPod system (Life Measurements Inc., Concord, CA, USA).
On the first visit to the laboratory, each subject completed the medical health history questionnaire and informed consent form, after which the subject performed a body composition assessment and TML O2max test (Figure 1). During the next 2 visits, all subjects were randomly assigned to perform a maximal exercise test on the Cybex ARC and the LifeFitness X9i ECT. For the last 2 visits, the subjects were randomly assigned to perform 2 30-minute submaximal exercise tests on the ARC and ECT.
Maximal Exercise Testing Procedures
Before any data collection, the subject was provided a warm-up period to acclimate for the maximal exercise test. Maximal oxygen uptake was measured during incremental exercise with the subject exercising until volitional exhaustion. The incremental maximal exercise test was performed by each subject on all 3 exercise modalities (TML, ARC, and ECT). All incremental maximal exercise tests were designed similarly, with the first 6 to 8 minutes of the test comprising 2-minute stages. Thereafter, each test stage was 1 minute in duration, with the goal of each test to attain O2max in approximately 8 to 12 minutes (11). Each subject was connected to the metabolic system by way of a 1-way breathing valve, and all subjects wore a Polar Heart Monitor that was interfaced with the metabolic system to allow for measurement of HRmax. In addition, RPE (Borg's 6-20 scale) was monitored at the end of each incremental stage and at the end of the test. All data were analyzed in 30-second segments, with O2max determined by averaging the 2 highest consecutive 30-second values.
Submaximal Exercise Testing Procedures
Before any submaximal test data collection, the subject was provided a warm-up period to acclimate to the exercise modality. Each submaximal test (ARC and ECT) comprised 3 consecutive 10-minute stages at 55%, 65%, and 75% of the subject's TML O2max. Oxygen consumption and HR were measured continuously throughout the 30-minute exercise test using the same metabolic and HR monitoring system as described above. Rating of perceived exertion (Borg's 6-20 scale) was measured during the last 30 seconds of each 10-minute stage for the perception of overall body exertion. Also, both knees, both hips, and low back discomfort were assessed during the last 30 seconds of each 10-minute stage using the NRS-11. Subjects were first asked to point to the rating of perceived overall body exertion on the Borg 6-20 scale. Subjects were then asked to point to the rating of pain and discomfort on the NRS-11 scale. Joint pain and discomfort was described using a 0 to 10 scale, with 0 suggesting “no pain or discomfort,” and 10 suggesting “worst pain possible.” The NRS-11 was chosen for the current study because of its ease of use during exercise trials, with a mouthpiece inserted for gas exchange measurements, and because it is highly preferred by other study populations, is correlated with other pain scales, and is widely used in the clinical and rehabilitation community (6,17). No intraclass correlation coefficients for reliability with the NRS-11 were determined for the current study. All HR and metabolic data were analyzed by averaging the last 5 minutes of each 10-minute stage.
All maximal exercise testing data were compared between modalities (TML, ARC, and ECT) with the use of a 1-way ANOVA with repeated measures. Tukey multiple comparison tests were performed post hoc to identify differences between modalities. Submaximal exercise testing data were compared between exercise modalities (ARC and ECT) and exercise intensities (55%, 65%, 75%) with the use of a 2-way ANOVA with repeated measures. An alpha level of 0.05 was used to establish statistical significance. Power analyses were amde with the perception of discomfort data at the hip during submaximal exercise, with a power = 0.80 and p = 0.05 and n = 15. Power analyses were made with the perception of discomfort data at the lower back, using a power = 0.80 and p = 0.05 and n = 23. All data are expressed as means ± SD. The JMP 7.0 statistical program (SAS Institute, Inc., Cary, NC, USA) was used for all statistical analyses.
Maximal Exercise Tests
A total of 18 subjects (10 male, 8 female) (Table 1) completed the 3 maximal and 2 submaximal exercise tests. Maximal exercise testing data are presented in Table 2. A significant main effect was observed for exercise modality on O2max (p = 0.007) (Figure 2A). Post hoc testing revealed a greater O2max value with the TML (53.8 ± 12.3 ml·kg−1·min−1) than the ECT (50.2 ± 11.1 ml·kg-1·min-1; p = 0.003). No significant difference was observed between the TML and ARC modalities (51.2 ± 12.1 ml·kg−1·min−1, p = 0.23).
A significant main effect was found for exercise modality HRmax (p < 0.0001) (Table 2 and Figure 2B). Post hoc testing revealed a greater HRmax value with the TML (188.9 ± 7.8 bpm) than the ECT (184.2 ± 8.4 bpm; p = 0.0006) and ARC (185.0 ± 8.0 bpm, p = 0.004). No significant difference was observed between the ECT and ARC modalities (p = 0.76).
Time to Attain O2max (min)
Time to attain O2max exhibited a main effect for exercise modality (p = 0.0001) (Table 2 and Figure 2C). Post hoc testing revealed a lesser time to achieve O2max with the ECT (8.8 ± 1.5 min) when compared with the TML (10.3 ± 1.0 min; p = 0.005) and ARC (10.9 ± 2.3 min, p = 0.0006). No significant difference was observed between the TML and ARC modalities (p = 0.71).
Submaximal Exercise Tests
All data (means ± SD) for the submaximal exercise tests are displayed in Table 3. In comparing the physiologic responses between the ARC and ECT across 3 exercise intensities (55%, 65%, and 75%), no significant main effect was found for oxygen cost between the 2 modalities (p = 0.20) (Figure 3A). No significant main effect was observed for exercise intensity (p = 0.19). Last, there was no interaction for modality and exercise intensity (p = 0.96).
Heart Rate (bpm)
Heart rate exhibited a main effect for exercise modality (p = 0.003) (Figure 3B), with exercise on the ECT eliciting a greater HR than ARC exercise. In addition, there was a significant main effect for exercise intensity (p < 0.0001). There was no interaction for modality and exercise intensity (p = 0.86).
RPE and Discomfort Measures
There was a significant main effect for rating of overall body perceived exertion (RPE) with exercise modality (p = 0.0009) (Figure 3C), with greater ratings of exertion during ECT exercise. The RPE values exhibited a main effect for exercise intensity (p = 0.0001). There was no interaction for modality and exercise intensity (p = 0.76).
There was no main effect for exercise modality (p = 0.68) (Figure 4A) or exercise intensity (p = 0.37) with perception of knee discomfort. In addition, no interaction was noted for modality and exercise intensity with knee discomfort (p = 0.44). Perception of hip discomfort exhibited a significant main effect for exercise modality (p = 0.021) (Figure 4B), with perceptions of discomfort being greater during ECT exercise. There was no main effect for exercise intensity with perception of hip discomfort (p = 0.75). Last, there was no interaction effect for modality and exercise intensity with perception of hip discomfort (p = 0.80). We observed no main effect for exercise modality with perception of low back discomfort (p = 0.077) (Figure 4C). There was no main effect for exercise intensity with perception of low back discomfort (p = 0.24). Also, there was no interaction for modality and exercise intensity with low back discomfort (p = 0.51).
To our knowledge, this is the first study to investigate various physiologic responses during exercise on a device (Cybex ARC) that uses an arc pathway of motion compared with popular fitness devices with which an elliptical path of motion is used. Although similar physiologic comparisons of ECT and TML exercise have been performed (2-5,15), comparing the physiologic responses of the ECT and TML with a newer exercise modality is certainly warranted given the widespread use and popularity of ECTs in health and fitness facilities (16). The primary purpose of this work was to determine whether differences exist between specific cardiorespiratory variables and lower-body joint discomfort during submaximal exercise intensities commonly prescribed to the general population (55%, 65%, and 75% of O2max) and during a maximal exercise test performed on an ARC, an ECT, and TML. We found the ARC exercise modality elicited O2max values similar to the TML, whereas the ECT values were lower than those obtained on the TML. It appears that the ARC is just as effective of an exercise modality in achieving O2max compared with the gold-standard TML during a graded exercise test. Similarities between the O2max derived while exercising on the ARC and the TML suggests the ARC may prove to be an additional modality for the purposes of exercise testing individuals with various conditions (e.g., lower-body osteoarthritis) who may have limitations to exercising on a TML (7).
In an investigation comparing the physiologic effects of exercising on a TML and ECT (3), similar values for O2max was observed between the 2 exercise modalities. This result is not in agreement with our work in which a lower O2max was achieved on the ECT in comparison with TML exercise. In the Dalleck et al. (3) investigation, the researchers used a graded exercise protocol for the ECT that progressively increased the step cadence and altered the incremental changes in resistance depending upon the sex, perceived fitness of the individual, and the time period of the test protocol. This protocol is quite different from the methodology we used in our work. In the current study, the graded exercise protocol for the ECT used a constant step cadence (140 steps/min), with the resistance increasing incrementally according to a standardized metric (2 units/2 min) for all subjects. Moreover, we used a similar protocol for the TML and ARC exercise modalities in which cadence or pace was able to be maintained throughout the testing, whereas the slope and resistance were increased according to a standardized, constant value. In addition, Dalleck et al. (3) found no difference in the time to obtain O2max with the ECT and TML, whereas, in our study, the time to obtain O2max was significantly less when exercising on the ECT compared with the TML and ARC. This may also, in part, explain the lower O2max values observed for the ECT. Although we did not monitor perception of lower-body joint discomfort during the maximal exercise tests, our findings of significantly greater perception of discomfort when performing submaximal exercise on the ECT suggests that lower-body limitations may influence maximal exercise testing on the ECT.
Another important finding in this study was that a greater perception of joint discomfort was observed at the hips, along with a trend toward significance (p = 0.077) at the lower back, when subjects exercised on the ECT compared with exercise on the ARC at similar submaximal exercise intensities. Our power analysis, as noted above, suggests a significance at the lower back would be observed with a sample population of 23 subjects. Although the ECTs are a widely popular exercise device for training at various submaximal intensities, it appears that the ARC induces less discomfort at the low back and hip joint than the ECT at similar exercise intensities. This finding is most likely attributed to the arc path of motion used by the ARC compared with the traditional elliptical path commonly used by the ECTs in the industry and certainly has its implications for individuals who suffer from low back discomfort when choosing an exercise device.
The submaximal exercise tests for the ARC and ECT were matched for oxygen cost between modalities, with the ventilatory responses (VE and RER) exhibiting no differences between exercise modalities at the 3 submaximal intensities. However, the HR responses for each of the 3 exercise intensities tested (55%, 65%, and 75%) were greater during the ECT trials. Because the O2max for ECT exercise was lower than what was measured during TML exercise, it could be suggested that the HR difference observed during submaximal exercise could be caused by the subjects exercising on the ECT at higher exercise intensities. However, this suggestion is unlikely because of the similarities in HRmax when performing on the ECT and ARC modalities. Perhaps a more likely reason for the greater HR response during the submaximal ECT exercise is the greater discomfort noted by the subjects. Perceived discomfort was significantly greater at the hip joints during ECT exercise, and it was also shown that perceived discomfort for the low back trended toward being significant (p = 0.077). Thus, it appears likely that the perception of discomfort exhibited by the trunk and hip region during exercise may be the stimuli for an increased HR response. Research investigating the influence of superficial and deep pain on sympathetic drive has produced different results (8,9). Pain stimulation has been observed to stimulate the insular cortex, within the lateral sulcus, of the cerebral cortex (9). The resultant autonomic response appears to depend upon the signal intensity and location of stimuli. More recently, Henderson et al. (10) have shown repeated valsalva maneuvers effectively influences the stimulation of the insular cortex, along with the cerebellum and dorsal brain stem, activating the sympathetic nervous system and decreasing parasympathetic drive. These researchers observed an increase in HR over time with repeated muscle maneuvers. Last, Henderson, et al. (8) observed a subset of subjects who exhibited no alteration in cardiovascular response to cutaneous and muscle pain stimuli. It is interesting to note that significant changes in technology over the past decade appear to be providing a mechanism for the advancement of research regarding pain activation of the autonomic nervous system. However, knowledge regarding the ability of joint and muscle discomfort to activate the autonomic nervous system during exercise is not conclusive at this time.
Specific to the differences in perception of lower-body discomfort, Porcari et al. (15) monitored ground reaction forces, O2, RPE, and HR during TML walking, TML running, ECT, cycle, and step exercise. In their investigation, the authors required subjects to exercise at a self-selected pace, in which whole body RPE was similar, as was the O2 and HR responses for TML run and ECT exercise. However, the O2 and HR responses were greater for TML walk, cycle, and step exercise. The authors suggested that oxygen consumption and HR relative to RPE do not appear to be closely linked for performing exercise on an ECT, with lower O2 and HR responses occurring at a given RPE (15).
Green et al. (5) noted higher RPE for the legs during moderate exercise with the ECT when compared with similar HR-matched intensity on the TML. No differences, however, were found for perceptions of exertion between TML and ECT at the chest or overall body. Previously, these researchers noted a difference in perceived exertion at the legs between cycle and ECT exercise (4). Importantly, the higher perceptions of exertion were noted specifically for the legs during ECT and cycle exercise compared with work performed on the TML. Collectively, these findings (4,5,15) are in agreement with our results in which we observed a greater perception of joint discomfort when exercising on the ECT compared with the ARC, as well as for whole body RPE. Past research regarding conscious and subconscious cues influencing one's perception of exertion suggests the exerciser compares different physiologic and perceptual cues during the exercise with resting and exercise expectation cues to provide input for the determination of exhaustion (1,14). Thus, different perceptions of exertion early in the exercise bout will provide feedback influencing the determination of one's onset of exhaustion. Our findings suggest that a localized perception of discomfort may occur early in the exercise bout at the hip region that may translate to whole body exhaustion. Ratings of perceived exertion typically reflect the integration of influences throughout the body and not any specific physiologic or psychological influence. Therefore, it appears that our 10-minute submaximal exercise periods were sufficient to influence discomfort at the hip region with exercise on the ECT that, over extended periods of time, would result in greater perceptions of exertion and an earlier onset of exhaustion.
Another factor that is expected to affect the specific localized area discomfort and whole body RPE is the path of motion used by the exercise modality, which ultimately dictates the lower-extremely joint mechanics and thus the loads experienced at the specific joints. With respect to this study, there is a stark contrast between the path of motion exhibited by the ARC and the ECT. The ECT tested in this study uses a traditional elliptical path for excursion, whereas the ARC tested in the current study uses a true arc pathway. By virtue of these fundamental differences, it is speculated that the dynamics at the hip, knee, and ankle would expect to be different, thus placing less stresses on these joints while theoretically providing exercise at the same intensity. Recent research by Lu et al. (12) has investigated joint mechanics when exercising on a traditional ECT versus other exercise modalities. In their work, they observed greater hip and knee exertion during ECT exercise when compared with other modalities. Moreover, this group demonstrated that greater hip flexor and knee extensor moments occur with exercising on an ECT when compared with walking. These findings may be attributed, in part, to smaller loading rates occurring at heel strike and to the increased slope of the pedal ellipse through the path of motion on the ECT. One would expect that the differences in joint moments would influence physiologic efficiency when exercising over extended exercise durations. Past research with ECTs (13) have noted large variability in O2 measured at predetermined stride rates and resistances, suggestive of inefficient exercise mechanics. In addition, we observed large variability in the perception of discomfort in the lower body during ECT, suggesting some individuals do not experience a large amount of lower-body discomfort with ECT exercise. Thus, individuals who use ECTs may experience discomfort because of inefficient mechanical loading that could possibly influence the onset of fatigue.
In summary, TML O2max values obtained during a traditional TML protocol were greater than those measured during exercise with the ECT but were not different from the values measured while exercising on the Cybex ARC. However, HRmax values were found to be greater during maximal TML exercise when compared with both the ARC and ECT. Moderate intensity, submaximal exercise with the ECT was found to exhibit greater HR responses and perceived discomfort at the hip joints when compared with similar oxygen costs while exercising with the ARC. Our findings suggest that different responses to submaximal exercise with the ECT result in greater perceptions of discomfort, which may be attributed to altered hip flexor and knee extensor moments. The use of ECTs as an exercise modality for individuals with, or at risk for, lower-body joint dysfunction may result in an increased perception of lower-body discomfort, which may not be beneficial in comparison with the ARC, which uses a different path of motion.
Fitness participants regularly exercise on TMLs and ECTs when training indoors. Our findings suggest individuals who are experiencing lower-body discomfort, or are at-risk for lower-body discomfort (e.g., individuals suffering from issues involving overweight, obesity, osteoarthritis, etc.), may gain relief from discomfort and improve their fitness if exercising on the Cybex ARC. When exercising at the same level of effort, the RPE responses were lesser with the ARC. In other words, perception of effort and discomfort were lesser with the ARC at similar training intensities. Any likely reasons for differences in discomfort between the 3 modalities are unknown at this time. It should be kept in mind that these differences in discomfort are likely to influence the fitness participant's program adherence and goal accomplishment.
The authors thank Cybex International Corporation for their financial support to complete this work. None of the authors are affiliated with either of the manufacturers of the exercise modalities tested in this study. Furthermore, the results of this study do not constitute endorsement of either product by the authors or the National Strength and Conditioning Association. The authors would also like to acknowledge the exercise testing efforts of Natalie Fulton.
1. Coquart, JBJ and Garcin, M. Knowledge of the endpoint: effect on perceptual values. Int J Sports Med
29: 976-979, 2008.
2. Crommett, A, Kravitz, L, Wonsathikun, J, and Kemerly, T. Comparison of metabolic and subjective response of three modalities in college-age subjects. Med Sci Sports Exerc
31: S677, 1999.
3. Dalleck, LC, Kravitz, L, and Robergs, RA. Maximal exercise testing using the elliptical cross-trainer and treadmill. JEPonline
7: 94-101, 2004.
4. Green, JM, Crews, TR, Bosak, AM, and Peveler, WW. Overall and differential ratings of perceived exertion at the respiratory compensation threshold: effects of gender and mode. Eur J Appl Physiol
89: 445-450, 2003.
5. Green, JM, Crews, TR, Pritchett, RC, Mathfield, C, and Hall, L. Heart rate and ratings of perceived exertion during treadmill and elliptical exercise training. Percept Mot Skills
98: 340-348, 2004.
6. Hartrick, CT, Kovan, JP, and Shapiro, S. The numeric rating scale for clinical pain measurement: a ratio measure? Pain Pract
3: 310-316, 2003.
7. Haug, RC, Porcari, JP, Brice, G, and Terry, L. Development of a maximal testing protocol for the NordicTrack cross-country ski simulator. Med Sci Sports Exerc
31: 619-623, 1999.
8. Henderson, LA, Bandler, R, Gandevia, SC, and Macefield, VG. Distinct forebrain activity patterns during deep versus superficial pain. Pain
120: 286-296, 2006.
9. Henderson, LA, Gandevia, SC, and Macefield, VG. Somatotopic organization of the processing of muscle and cutaneous pain in the left and right insula cortex: a single-trial fMRI study. Pain
128: 20-30, 2007.
10. Henderson, LA, Macey, PM, Macey, KE, Frysinger, RC, Woo, MA, and Harper, RK. Brain responses associated with the valsalva maneuver revealed by functional magnetic resonance imaging. J Neurophysiol
88: 3477-3486, 2002.
11. Howley, ET, Bassett, DR Jr, and Welch, HG. Criteria for maximal oxygen uptake: review and commentary. Med Sci Sports Exerc
27: 1292-1301, 1995.
12. Lu, T-W, Chien, H-L, and Chen, H-L. Joint loading in the lower extremities during elliptical exercise. Med Sci Sports Exerc
39: 1651-1658, 2007.
13. Mier, CM and Feito, Y. Metabolic cost of stride rate, resistance, and combined use of arms and legs on the elliptical trainer. Res Q Exerc Sport
77: 507-513, 2006.
14. Parvizi, J and Damasio, A. Consciousness and the brainstem. Cognition
79: 135-159, 2001.
15. Porcari, JP, Zedaker, JM, Nawer, L, and Miller, M. Evaluation of an elliptical exerciser in comparison to treadmill walking and running, stationary cycling, and stepping. Med Sci Sports Exerc
30: S168, 1998.
16. Sporting Goods Manufacturer's Association. 2008 Sports & Fitness Participation Report
. Washington, DC: Sporting Goods Manufacturer's Association. 2008. p. 3.
17. Williams, A, Davies, H, and Chadury, Y. Simple pain rating scales hide complex idiosyncratic meanings. Pain
85: 457-463, 2000.