Running economy (RE) is generally defined as a measure of steady-state oxygen consumption (V̇o2) per unit body mass (BM) at various submaximal running speeds (27). Compared to persons with poor RE, in general, individuals with good RE consume less oxygen per unit BM, exhibit greater reliance on aerobic means of energy production, display an enhanced ability to use fat as a fuel, rely less on fast-twitch muscle fiber contribution, and accumulate lower concentrations of lactic acid while running at the same velocity (27). Many factors have been shown to affect RE, including muscle fiber type, training status, gait mechanics, muscle stiffness, tendon length, and flexibility (27).
Another variable that has been shown to influence RE is footwear. Two studies have found that for every 100 g of mass added to a shoe, V̇o2 increases by approximately 1% across a range of submaximal running speeds (9,11). In addition, well-cushioned shoes have been reported to lower the aerobic demand of running by up to 2.8% when compared with less-cushioned shoes (9,12). Taken together, these findings provide support for athletes to use lightweight cushioned shoes to improve RE.
Many athletic shoe companies have taken advantage of recent advances in shoe material technology to create lightweight shoes of varying cushioning characteristics. Recent advances in shoe design have also led to the development and marketing of minimal shoes, which are extremely lightweight and lack a cushioned elevated heel, arch support, and a stiff sole. The recent popularity of running in minimal shoes (i.e., minimal or natural running) can be attributed to the training practices of the Tarahumara Indians, who run in thin sandals with little material (22). Also, shoe companies have begun manufacturing lightweight, low-cushioned minimalist shoes that mimic running barefoot.
Advocates of minimalist running believe that this form of locomotion may aid in decreasing the incidence of chronic running injuries (19,26). However, an initial consequence of running barefoot or in minimal-type shoes is to increase stride frequency compared with that of running in standard running shoes (25). This adjustment is possibly because of an increased sense of footfall awareness, which causes the runner to moderate their stride frequency and, in turn, stride length, to reduce impact discomfort (25). A reduction in discomfort would therefore decrease overall effort and therefore reduce overall oxygen consumption and improve or maintain RE (20).
However, despite the evidence supporting an increase in stride frequency because of changes in footwear, the effect of wearing minimalist shoes on RE is sparse and inconclusive. Hanson et al. (14), for instance, demonstrated that while running at 70% of velocity at V̇o2max (vV̇o2max), barefoot running was more economical than running shod during both over ground and treadmill conditions. Conversely, Franz et al. (9) found that running barefoot at 3.35 m·s−1 offered no energetic advantage over running shod, when energy usage was assessed from a metabolic power perspective. Moreover, these authors reported that running shod provided metabolic savings of 3–4% compared with running barefoot.
Given the ambiguity of previous research on this topic, the purpose of this study was to quantify differences in RE at 50 and 70% of vV̇o2max while running in 3 different footwear conditions (barefoot, minimal shoes, and normal running shoes) among recreational distance runners. It was hypothesized that running in a minimal shoe would increase step frequency and lead to better RE than in a standard running shoe or barefoot at speeds of 50 and 70% of vV̇o2max. We also hypothesized that running barefoot would increase step frequency and produce a lower aerobic demand of running compared with running in a standard shoe.
Experimental Approach to the Problem
A repeated-measures multivariate design was used to evaluate the effects of running barefoot versus that of running in a minimal and standard running shoe, on RE (oxygen consumption) and step frequency. Each subject was asked to visit the laboratory on 4 separate occasions. During the first visit, assessment of velocity at V̇o2max was performed. During the second, third, and fourth running sessions, each subject ran in a different randomized footwear condition (barefoot, minimally shod, and standard shod) from the previous visit. Each session began with a 6-minute warm-up to familiarize the subject with the treadmill. This was followed by a 5-minute RE assessment at a speed of 50% vV̇o2max, a 5-minute standing rest to allow heart rate (HR) to return to baseline, and then another 5-minute RE assessment at a speed of 70% vV̇o2max. Each RE trial was video recorded to allow for later analysis and hand-counting of step frequency. The final 2 minutes of each 5-minute trial was used for analysis of both RE and step frequency.
Participants were healthy, injury-free males (N = 9) aged 18–40 years (mean age = 26.8 ± 6.8 years, BM = 81.2 ± 11.9 kg, and height = 177.6 ± 7.0 cm). Before testing began, university Institutional Review Board approval was obtained and all participants provided written informed consent. Participants completed 4 laboratory testing sessions and refrained from liquids (aside from water), food, and caffeine 4 hours before testing. Furthermore, the subjects were told not to engage in strenuous exercise the previous 24 hours and in general aerobic work the previous 8 hours of each visit. To account for any confounding effects from wearing minimal shoes for the first time, participants were required to have worn and trained (≥5 miles a week) in a minimal-type shoe for a minimum of 3 months before the start of this investigation (26).
Maximal and Submaximal Oxygen Consumption
Maximal oxygen consumption (V̇o2max) and submaximal V̇o2 were measured using a metabolic cart with an open-circuit spirometry gas collection system (MOXUS Modular V̇o2 System; AEI Technologies, Pittsburgh, PA, USA). Concentrations of expired oxygen and carbon dioxide were determined using an S-3AOxygen Analyzer sensor and CD-3A Carbon Dioxide analyzer, respectively. Inspired volumes were measured by a turbine attached to a Hans Rudolph mouthpiece, and the turbine was calibrated before each testing session using a certified 3-L calibration syringe. Gas analyzers were calibrated before each treadmill test to room air (20.93% O2 and 0.03 CO2) and known standard gas concentrations (4.03% CO2 and 16.02% O2; NexAir, Memphis, TN, USA). All running bouts were conducted on a FITNEX high-speed treadmill (Dallas, TX, USA).
Step frequency was obtained from a video recording using a Sony Handy Cam with 16 GB internal memory. A large digital clock (Big Time Clocks Model #116; New York, NY, USA) with a stopwatch function was placed in view of the video camera to determine the number of steps taken during each 5-minute RE trial. Step frequency was determined from the video recording by manually counting the number of steps taken between minute 4 and 5 of each RE trial. Each step frequency video was viewed, and steps counted, twice to obtain a reliable measure. In the instance that 2 measures were greater than 2 steps apart, a third measure was conducted, and the nearest 2 measures were averaged.
Height and BM were measured twice in minimal clothing and barefoot, to the nearest 10th of a centimeter (SECA 222-stadiometer; SECA North America, Chino, CA, USA) and 10th of a kilogram (Health O′ Meter mass scale), respectively. If height or BM varied by more than two-tenths of a centimeter or kilogram, respectively, an additional measure was taken and the nearest 2 measures were averaged.
To eliminate footwear accommodation time, subjects were not provided with any type of footwear. Instead, subjects were asked to arrive with and wear their own minimal and standard running shoes during each respective RE trial. Minimal shoes are described as lacking those features found in standard running shoe (a cushioned elevated heel, arch supports, and a stiff sole). As minimal-type shoes often vary between manufacturers, a heel-to-toe drop of less than 4 mm, a mass of less than 255.1 g, and stack height of less than 20 mm (29) were required for each pair of minimal shoes. Stack height (shoe heel height) was determined by measuring the participant's height in each of the 2 respective shoe conditions (standard shod, minimal) and subtracting the participant's barefoot height, with the calculated difference deemed as each shoe's respective heel height. In cases when stack height varied by more than two-tenths of a millimeter, an additional measure was taken and the nearest 2 measures were averaged. Assessment of shoe mass for the standard and minimal shoes was done using a scientific grade gram scale (American Weigh Scales, Norcross, GA, USA). Standard running shoes are defined as having a cushioned elevated heel, arch supports, and a stiff sole (25).
Heart rate was recorded using a Polar Electro (Model #N2965; Polar Electro, Lake Success, NY, USA) HR strap and sensor synced to the metabolic cart and measured throughout the V̇o2max test and each RE trial.
Maximal Oxygen Consumption
Maximal oxygen uptake was measured during the first laboratory visit using procedures applied by Hanson et al. (14), and each participant ran in his respective standard shoe. The testing protocol began at a speed of 80.5 m·min−1 and increased by 26.8 m·min−1 every 2 minutes until HR reached 170 b·min−1. Once a HR of 170 b·min−1 was attained, running velocity was increased by 13.4 m·min−1 every minute until volitional exhaustion occurred. Maximal oxygen consumption was deemed to have been met if participants satisfied 2 of 3 criteria, which include a plateau in oxygen consumption with an increase in workload, an respiratory exchange ratio of 1.10 or higher, or attainment of a HR value within 10 beats of age-predicted max (220-age) (1). A protocol designed by Hanson et al. (14) was used to determine velocity at V̇o2max. Velocity at V̇o2max is the speed at which the athlete is running when V̇o2max occurs, as long as this speed is sustained for at least 1 minute. If an athlete achieved a V̇o2max during a workload that was not sustained for 1 minute, the speed of the previous workload was recorded as vV̇o2max (2).
Running Economy Trials
Within 5 days of the assessment of V̇o2max, participants completed the first of 3 footwear-specific RE trials (barefoot, minimal shoe, and standard running shoe) in the participant's own footwear (i.e., footwear not provided by researchers). Each laboratory visit began with the measurement of BM and was followed by a treadmill accommodation period of 6 minutes at a self-selected pace in the randomly assigned footwear condition (18). Following the accommodation period, runners were then prepared for the RE assessment, which consisted of a 5-minute trial at a speed of 50% and also at 70% of vV̇o2max. Additionally, to simulate over ground running, the treadmill was set at 1% grade for each trial (16). Running economy was calculated as the average V̇o2 over minutes 3 to 5 of each trial with a steady-state condition confirmed as maintenance of an RER of less than 1.0. Using a procedure followed by Hardin et al. (15), a 5-minute standing rest period was provided, in which HR was required to fall to below 120 b·min−1 before beginning the second RE trial. Heart rate was measured as the average beats per minute during minutes 3 to 5 of each RE trial, whereas rate of perceived exertion was measured using the Borg scale (6–20) during the last 15 seconds (4:45–5:00) of each trial.
Data analysis was performed using the International Business Machines Corporation Statistical Packages for the Social Sciences (version 19.0) software. Descriptive statistics for participants and shoe characteristics (e.g., mass, stack height) are expressed as mean ± SDs. A 2-way repeated-measures multivariate analysis of variance (MANOVA) was used to evaluate RE and step frequency under 3 footwear conditions (barefoot, minimal shoe, standard running shoe) and 2 running speeds (50 and 70% of vV̇o2max) as within-subjects factors and statistical significance defined with an alpha level of p ≤ 0.05. If a significant multivariate effect was detected, separate univariate ANOVAs were used to identify group differences (alpha level of p < 0.025) for each dependent variable (RE, step frequency). Intraclass correlations were calculated to evaluate the test-retest reliability of the dependent variables and can be found in Table 1.
Descriptive statistics for participants, including shoe characteristics, can be found in Table 2. The 2-way repeated-measures MANOVA indicated a significant multivariate effect for footwear (Pillai's F = [4, 32] = 7.79, p < 0.001, partial η2 = 0.49), intensity (Pillai's F = [2, 7] = 188.67, p < 0.001, partial η2 = 0.98), and the interaction between footwear and intensity (Pillai's F = [4, 32] = 4.96, p = 0.003, partial η2 = 0.38). Because of the significant interaction effect, each level of intensity (50 and 70%) was analyzed separately by 1-way repeated-measures MANOVAs using an adjusted alpha level of 0.025. The effect of footwear was significant at both 50% of vV̇o2max (Pillai's F = [4, 32] = 3.91, p = 0.01, partial η2 = 0.33) and 70% of vVO2nax (Pillai's F = [4, 32] = 10.02, p < 0.001, partial η2 = 0.56). When results for RE and step frequency were considered separately, step frequency was statistically significant at both 50% of vV̇o2max (G-G F = [1.5, 12.1] = 14.78, p = 0.001, partial η2 = 0.65) and 70% of vV̇o2max (G-G F = [1.6, 12.6] = 65.29, p = 0.001, partial η2 = 0.89). Running economy was not significant at 50% of vV̇o2max (G-G F = [1.6, 12.4] = 0.07, p = 0.89, partial η2 = 0.01) or 70% of vV̇o2max (G-G F = [1.7, 13.9] = 2.50, p = 0.126, partial η2 = 0.236). Descriptive statistics for RE and step frequency measures can be found in Table 3. At 50% of vV̇o2max, step frequency was higher when running barefoot than when running in the minimally shod condition (p = 0.007). At 70% of vV̇o2max, step frequency was also higher in the barefoot condition than in the minimal (p < 0.001) and standard conditions (p < 0.001), with the minimal condition also exhibiting significantly higher step frequency than the standard condition at 70% vV̇o2max (p = 0.007).
The current study was conducted to determine the effect of footwear on RE at 2 submaximal running speeds in recreational distance runners. The hypothesis that running in a minimal shoe would be more economical than running barefoot or in a standard shoe was not supported at either speed. However, the second hypothesis was partially supported, in that running in minimal footwear significantly increased step frequency compared with running in the standard shoe condition at a speed of 70% vV̇o2max but not at 50% vV̇o2max. Moreover, while running barefoot, step frequency was significantly higher when compared with running in the minimal and standard shoe conditions at 70% vV̇o2max, but only significantly higher than the minimal shoe condition at 50% vV̇o2max. Despite these differences in step frequency, however, RE did not vary significantly across footwear conditions at 50 and 70% vVV̇o2max.
The results support the notion that the association between gait descriptors and the aerobic demand of running is complex and multifaceted (11,12,28). At stable speeds, step frequency and step length exhibit a negative and interchangeable relationship, such that, as stride length is increased, stride frequency is relatively reduced and the opposite is also true (27). Although not statistically measured, conversion of step frequency to a stride length measure (meters per minute ÷ 1/2 number of steps taken per minute) yielded worthwhile information when comparing means. For instance, at 50% of vV̇o2max, running in minimally shod shoes led to a 2.1% longer stride length compared with running barefoot, whereas running in standard shoes resulted in a 2.4% longer stride length compared with the barefoot running condition. At 70% of vV̇o2max, minimally shod runners adopted a stride length that was 3.2% longer than barefoot runners, whereas running in standard shoes produced a 2.3% greater stride length than running in the minimally shod condition and a 5.6% longer stride length than barefoot running. Previous research (6,7,9) has also indicated an increase in stride length from barefoot to a standard shoe running conditions among recreationally fit runners with significant barefoot running experience. Similarly, in a group of highly trained runners, Bonacci et al. (3) demonstrated that stride length and stride frequency were significantly longer and slower, respectively, while running in minimal shoes compared with running barefoot. In the current study, running in the shod condition yielded a measurably longer stride length and significantly lower step frequency (at 70% vV̇o2max) than running in both the minimal and the barefoot conditions.
Although step frequency varied across footwear conditions, no significant changes were observed in RE at either speed. In a small number of published studies, reductions in the aerobic demand of running have been demonstrated when running barefoot or in a barefoot-similar condition (e.g., thin diving sock or Vibram 5-finger shoe) versus shod running (7,8,14,28), whereas other research has indicated that running barefoot is less economical than running shod (9). Although no single explanation can be offered as to why RE has been shown to vary among different footwear conditions, variations in shoe mass may provide some insight relative to this question.
In this regard, multiple authors have reported that the oxygen demand of running is increased by approximately 1% per 100 g of mass added per shoe (9,11). Current literature suggests that a threshold mass may exist, beyond which RE becomes worse. A number of studies, for instance, have shown that a mass of ∼350 g or greater may induce statistically significant changes in RE. Specifically, Divert et al. (7) assessed oxygen consumption in runners wearing a thin diving sock loaded with either 150 or 350 g and in standard running shoes weighing either 150 or 350 g. The authors found that RE in the 150 g barefoot conditions (diving sock and completely barefoot) did not significantly vary from either the unweighted diving sock or the unweighted barefoot conditions. However, both 350 g conditions (diving socks loaded with 350 g and 350-g standard running shoes) exhibited higher oxygen costs compared with the 150 g conditions (diving socks loaded with 150 g and 150-g standard running shoes). Similarly, Flaherty (8) reported that running in shoes with a mass of 356 g (per shoe) resulted in 4.6% greater oxygen consumption compared with running barefoot. In contrast, Franz et al. (9) found that adding 300 g to a lightweight cushioned shoe (∼450 g per foot) was only 1.2% less economical when compared with running barefoot. In our study, 2 subjects ran in standard running shoes with a mass of greater than 350 g (362 and 381 g), with 1 subject exhibiting his lowest oxygen consumption and the other subject exhibiting his highest oxygen consumption when compared with running in the barefoot and minimal conditions. Therefore, a certain shoe mass may indeed lead to a noticeable alteration in oxygen consumption while running, but the extent of the change is likely to vary on an individual basis.
Although shoe mass has been proven to be inconclusive in explaining variations in RE, shoe design or cushioning characteristics may help explain the fluctuation in energy demands caused by running in different types of footwear. Nigg et al. (24) tested the effects of 2 shoe midsole materials on various measures, including RE, in 18 runners. These authors found that 5 runners exhibited a more economical running style while wearing a more elastic and viscous heel, whereas 8 participants failed to demonstrate consistent results in either condition. These findings (13,17) suggested that the influence of shoe materials on RE are personalized and support previous research (13,17). Thus, runners will modify their running patterns, including altering footfall patterns from a rearfoot to a midfoot strike, to maintain constant external loads and minimize changes in metabolic energy demands while running. The current study did not specifically analyze footfall patterns, but visual inspection suggested the presence of a variety of foot placement styles between participants and within footwear conditions.
Previous studies (7,13) indicate that runners will ultimately rely on the running patterns that are most economical for them. Additionally, numerous studies have demonstrated that aerobic demand is increased at a constant speed when stride length is either lengthened or shortened beyond that which is the freely chosen for an individual (21). Indeed, Cavanagh and Williams (5) suggested that freely chosen stride length is at or near optimal RE. Therefore, as suggested by Nigg and Enders (23), runners may automatically adapt to a change in running condition or strike pattern to maintain the most economical running performance.
Significant differences in RE across footwear conditions were not found, yet individual RE between footwear conditions sometimes varied substantially. For example, one runner exhibited a 3% difference in RE among footwear conditions at 50% of vV̇o2max, whereas another exhibited a 12% difference in RE across footwear conditions. At 70% vV̇o2max, one runner exhibited a 2% difference in RE across footwear conditions, whereas another exhibited a 12% difference across footwear conditions. The implications of these differences may be more valuable when considering their influence on running performance. Burkett et al. (4) demonstrated that a 1.7% increase in absolute V̇o2 is equal to roughly a 5 m·min−1 decrease in speed. Using the findings of Burkett et al. (4) and Hanson et al. (14), a 1.0% increase in V̇o2 would result in a 2.94 m·min−1 decrease in running speed. Based on these data, a 3–12% difference in V̇o2 would translate to a speed increase of approximately 8–35 m·min−1 or 0.30–1.31 m·h−1. In regards to performance, a change of 1 m·h−1 could potentially improve performance by ∼3 minutes and 25 seconds over the course of a 5-km race. Therefore, a change in footwear (or running barefoot) could prove useful for a given individual in improving RE, subsequent training, racing performance, or a combination of the two.
The results of the current study support the theory that runners ultimately adapt and are most economical in conditions specific to that individual's characteristics. Although RE was relatively constant among shoe conditions, large interindividual differences in RE could be readily distinguished. At 50% of vV̇o2max, for instance, 4 of the 9 runners exhibited their highest oxygen consumption (i.e., least economical) while running in a minimal shoe. Alternatively, another 4 subjects were most economical in the standard shod condition. At 70% of vV̇o2max, 5 of the 9 subjects exhibited their highest oxygen consumption in the minimal shoe condition, whereas another 4 of the 9 participants were most economical while engaged in barefoot running. One potential explanation for these disparate findings could be related to the comfort level experienced by each participant while running under each shoe condition. These findings are similar to those of Luo et al. (20) who demonstrated that RE was significantly improved when running in a shoe deemed most comfortable compared with shoes that were labeled less comfortable.
Future research considerations should include determining the impact of shoe heel-to-toe drop on RE and the possibility of a mass threshold at which RE may be significantly affected. In addition, researchers may wish to test RE at speeds greater than 50% of vV̇o2max for a low-end speed because many runners found it to feel very slow. Only runners on the lower end of the vV̇o2max continuum seemed to be affected by the slower running speed, but analysis of RE at speeds of greater than 60% of vV̇o2max may improve external validity.
In conclusion, RE was not significantly different across barefoot, minimally shod, and standard shoe conditions in a group of recreational distance runners while running at low and moderate intensities. Stride length and step frequency differences were observed between footwear conditions, thereby providing a possible explanation for the mechanisms responsible for maintenance of RE despite footwear differences. Finally, these findings suggest the presence of a potential individual physiological preference for a particular running condition that could ultimately determine metabolic effort rather than the presence of an actual barefoot or shoe effect on RE.
No significant differences in RE were observed in recreationally trained male distance runners with experience running in minimal footwear. However, measurable differences in RE can be seen between individual runners and between footwear conditions. These findings indicate athletes should be wary of switching footwear conditions without having spent time training and adjusting to the footwear condition because it can affect individual RE. Thus, runners attempting to improve RE through running barefoot or changing footwear type should seek the footwear choice that most adequately conforms to their physical needs and comfort-level preferences.
1. Astorino TA, Robergs RA, Ghiasvand F, Marks D, Burns S. Incidence of the oxygen plateau at VO2
max during exercise testing to volitional fatigue. J Exerc Physiol Online 3: 1–12, 2000.
2. Billat VL, Hill DW, Pinoteau J, Petit B, Koralsztein J-P. Effect of protocol on determination of velocity at VO2max
and on its time to exhaustion. Arch Physiol Biochem 104: 313–321, 1996.
3. Bonacci J, Saunders P, Hicks A, Rantalainen T, Vicenzino BT, Spratfor W. Running in minimalist and lightweight shoe is not the same as running barefoot: A biomechanical study. Br J Sports Med 47: 387–392, 2014.
4. Burkett LN, Kohrt WM, Buchbinder R. Effects of shoes and foot orthotics on VO2
and selected frontal plane knee kinematics. Med Sci Sports Exerc 17: 158–163, 1985.
5. Cavanagh PR, Williams KR. The effect of stride length variation on oxygen uptake during distance running. Med Sci Sports Exerc 14: 30–35, 1982.
6. De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech 33: 269–278, 2000.
7. Divert C, Mornieux G, Baur H, Mayer F, Belli A. Mechanical comparison of barefoot and shod running. Int J Sports Med 26: 593–598, 2005.
8. Flaherty RF. Running economy and kinematic differences among running with the foot shod, with the foot bare, and with the bare foot equated for weight. Doctoral dissertation. Springfield, MA: Springfield College, 1993.
9. Franz C, Wierzbinski CM, Kram R. Metabolic cost of running barefoot versus shod: Is lighter better? Med Sci Sports Exerc 44: 1519–1525, 2012.
10. Frederick EC. Measuring the effects of shoes and surfaces on the economy of locomotion. In: Biomechanical Aspects of Sport Shoes and Playing Surfaces. Nigg B.M., Kerr B.A., eds. Calgary, AB: University of Calgary, 1983. pp. 93–106.
11. Frederick EC, Clark TE, Larsen JL, Cooper LB. The effect of shoe cushioning on oxygen demands of running. In: Biomechanical Aspects of Sport Shoes and Playing Surfaces. Nigg B.M., Kerr B.A., eds. Calgary, AB: University of Calgary, 1983. pp. 104–114.
12. Frederick EC, Daniels JT, Hayes JW. The effect of shoe weight on the aerobic demands of running. In: Current Topics in Sports Medicine. Bachl N., Prokop L., eds. Vienna, Austria: Urban & Schwarzenberg, 1984. pp. 616–625.
13. Hamill J, Russell EM, Gruber AH, Miller R. Impact characteristics in shod and barefoot running
. Footwear Sci 3: 33–40, 2011.
14. Hanson NJ, Berg K, Deka P, Meendering JR, Ryan C. Oxygen cost of running barefoot vs. running shod. Int J Sports Med 32: 401–406, 2011.
15. Hardin EC, Bogert AJ, Hamill J. Kinematic adaptations during running effects of footwear, surface, and duration. Med Sci Sports Exerc 36: 838–844, 2004.
16. Jones AM, Doust JH. A 1% treadmill grade most accurately reflects the energetic cost of running. J Sports Sci 14: 321–327, 2007.
17. Kong PW, Candelaria NG, Smith DR. Running in new and worn shoes: A comparison of three types of cushioning footwear. Br J Sports Med 43: 745–749, 2009.
18. Lavcanska V, Taylor NF, Schache AG. Familiarization to treadmill running in young unimpaired adults. Hum Mov Sci 24: 544–557, 2005.
19. Lieberman DE, Venkadesan M, Werbal WA, Daoud AI, D'Andrea S, Davis I, Pitsiladis Y. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 463: 531–535, 2010.
20. Luo G, Stergiou P, Worobets J, Nigg B, Stefanyshyn D. Improved footwear comfort reduces oxygen consumption during running. Footwear Sci 1: 25–29, 2009.
21. Martin PE, Morgan DW. Biomechanical considerations for economical walking and running. Med Sci Sports Exerc 24: 467–474, 1992.
22. McDougall C. Born to Run: A Hidden Tribe, Super Athletes, and the Greatest Race the World Has Never Seen. New York, NY: Alfred A. Knopf, 2009.
23. Nigg B, Enders H. Barefoot running
—Some critical considerations. Footwear Sci 5: 1–7, 2013.
24. Nigg BM, Stefanyshyn G, Stergiou P, Miller J. The effect of material characteristics of shoe soles on muscle activation and energy aspects during running. J Biomech 36: 569–575, 2002.
25. Perl DP, Daoud AI, Liebermann DE. Effects of footwear and strike type on running economy. Med Sci Sports Exerc 44: 1335–1343, 2012.
26. Robbins SE, Hanna AM. Running-related injury prevention through barefoot adaptations. Med Sci Sports Exerc 19: 148–156, 1987.
27. Saunders PU, Pyne DB, Telford RD, Hawley JA. Factors affecting running economy in trained distance runners. Sports 34: 465–485, 2004.
28. Squadrone R, Gallozzi C. Biomechanical and physiological comparison of barefoot and two shod conditions in experienced barefoot runners. J Sports Med Phys Fitness 49: 6–13, 2009.
29. Winn Y. Barefoot/minimalist running shoes: How to choose. 2013. Available at: http://www.rei.com/learn/expert-advice/how-to-choose-barefoot-minimalist-running-shoes.html
. Accessed June 10, 2014.