University of Colorado Study Determines that “Running in lightweight, cushioned shoes is equally as efficient as running barefoot.”
Posted on March 08 2012
The big news in the world of running science over the past week was the release of a study from a locomotion lab at the University of Colorado. The study, titled “The metabolic cost of running barefoot vs. shod: is lighter better?”, was headed up by Jason Franz, and was co-authored by Corbyn Wierzbinski and Rodger Kram.
I received a press release about the study from one of the co-authors last week, and the quote in the title of my post is taken directly from the first paragraph of the release (I have now also read the full text). In the press release they describe their experiment as follows:
“In the study, 12 subjects with substantial barefoot running experience ran at 7.5 MPH with a mid-foot strike pattern on a motorized treadmill, both barefoot and in lightweight cushioned shoes (~150 g/shoe, 5.4 oz). In additional trials, they attached small lead strips to each foot/shoe (~150, ~300, ~450 g). For each condition, they measured the subjects’ rates of oxygen consumption and carbon dioxide production as an index of metabolic cost.”
Here is their report of the key findings:
1. For every 100g (3.5oz) (the average weight of a deck of playing cards) added per foot, energy cost increases by approximately 1% whether running barefoot or shod.
2. Running barefoot and in lightweight shoes do not significantly differ in energy cost.
3. When controlling for shoe/foot mass, running in lightweight shoes requires ~3-4% less energy than running barefoot.
And here is the graph depicting their results (via Alex Hutchinson of Sweat Science):
What you’ll notice in the above graph is that all of the shod datapoints (dark circles) fall on the bottom line, whereas all of the “barefoot” data points fall on the upper line, indicating greater oxygen consumption and thus lower economy when “barefoot.” However, the stars indicate that the only significant differences were when the lightweight 150g shoe (Nike Mayfly, photo at top of post) was compared to a “bare” foot with an equal amount of weight attached, and when the weight-matched “bare” foot with 300g attached was compared to the shoe with 150g attached.
When I read a study like this, I tend to think immediately in terms of what these results might mean from a practical standpoint to runners. What can we learn and apply from this?
I see two major take-home messages:
1. Adding weight to the foot reduces economy at a rate of roughly a 1% increase in energy cost per 3.5 oz added. This means that trading your 12 oz trainers for an ultra-light flat like the sub 4 oz Mizuno Universe on race day could improve your economy quite a bit. But, presumably you’d have to also be adapted enough to run the race distance in the ultra-light flats (i.e., you should probably do some amount of training in the shoes you plan to race in so that you are used to them).
2. The study shows that if you are already a regular barefoot or minimalist runner and are not a heel-striker (the authors specifically controlled for footstrike to avoid comparing heel striking in shoes to midfoot/forefoot when barefoot), running in a lightweight racing flat will probably not significantly change your economy (of the 12 runners, 8 were more economical in the shoe, 4 were more economical “barefoot,” and when pooled there was no significant difference economy between when the runners were shod vs. unshod). Thus, if you feel that lightweight shoes are desirable in a race situation, go for it – they might even improve your economy a bit.
Both of these findings are interesting and of practical value, and the study raises a ton of questions that will likely be examined in future research. Running Times went so far as to run a headline saying “Here’s Proof Barefoot Isn’t Better” – well, barefoot was in fact better for one-third of the subjects, but we don’t really know why. Given the tendency to jump to interpretations like this, it’s important to recognize that like every scientific study, there are some limitations that should be mentioned related to the interpretation of the results – below are some things to think about:
1. None of the study subjects were actually barefoot, which is why I have been using quotation marks whenever I refer to the “barefoot” condition. In the study Methods section it mentions that “For the duration of the experiment, subjects wore very thin, slip-resistant yoga socks for safety and hygienic purposes.”
Would this have any effect on the results? I don’t really have any idea (what do you barefoot runners think?). Certainly, wearing a sock does add a layer of material between the foot and the ground that could alter sensory perception and plantar abrasion/shear (reducing the latter could actually improve economy relative to barefoot when running on a treadmill – truly barefoot treadmill running can create friction with the belt and resulting discomfort for some people). Just as people get sensitive when the phrase “barefoot-shoes” gets thrown around, I think it’s important to be clear with terminology – barefoot means nothing on the foot.
2. For similar reasons to the above, it’s debatable whether adding weight to the bare foot tells us anything of practical relevance about barefoot running. Here is the description of how weight was added to the “bare” foot:
“To add mass to each foot during barefoot trials, we modified the uppers of a different model of running shoes to allow for easy attachment of small lead strips while still simulating barefoot running (Figure 3). We removed the outsole, midsole, and the entire front portion of the shoe upper, anterior to the mid-shaft of the fifth metatarsal, leaving only the heel counter, thin fabric arch section, tongue and laces.”
So, the runners were essentially wearing a chopped shoe upper with no sole or forefoot, with lead weights attached. I do understand the difficulty of doing an experiment like this, but running in a sock and a shoe upper with attached lead weights gets pretty far away from the barefoot condition, and should not be referred to as such (again, barefoot means nothing on the foot).
Rather, I think the more relevant description of what this comparison showed is that running in a “shoe” (and sock) of equivalent weight without a sole is less economical than running in a shoe with a cushioned sole – still an interesting finding. This demonstrates that presence of a shoe sole can improve economy (possibly due to cushioning effects reducing work needed to be performed by muscles, increased longitudinal stiffness, etc.). It would be interesting to compare a lightweight cushioned shoe like the Mayfly to an equal-weight, minimally cushioned shoe like a Merrell Road Glove or Inov-8 Bare-X 180 and see how they compare (the reality is that more people are running and racing in shoes like this than are running or racing truly barefoot anyway).
3. There appears to be some amount of variation among the study subjects in barefoot/minimalist running experience. One of the requirements for inclusion was running “8 km/week barefoot or in minimal running footwear (e.g. Vibram Five Fingers) for at least 3 months out of the last year.” So, the runners did not even have to have true barefoot experience. However, this study does a much better job than most previous studies in attempting to control for level of barefoot/minimalist experience when looking at unshod running (i.e., they weren’t asking a bunch of people who had only ever run in traditional trainers to run with their shoes off for the first time).
I don’t really have a problem with this level of variation as it can be very difficult to recruit subjects for this type of study who meet all of the required criteria, but I am intrigued to know if there is any relationship between amount of barefoot experience and whether or not the individual was more economical when “barefoot” or in the Mayfly’s (remember, one-third of the runners had higher economy when barefoot). I’m curious as to whether individuals who run more barefoot miles (or perhaps who are more fit) demonstrate soft tissue adaptations that might improve their economy in the barefoot condition.
4. In the press release, the authors state that “Extrapolating their data indicates that running in only slightly lighter shoes (~129 g) would offer a statistically significant 2.5% energetic advantage over running barefoot.”
Extrapolating requires following the dotted line in the above graph downward to the left, which I don’t personally feel is statistically valid. There are no data from this region of the graph, and one cannot simply assume that the relationship will remain linear – as cushioning is reduced with decreasing shoe weight, it is entirely possible that linearity will be lost if economy is reduced. In fact, if you look at the three dark circles on the bottom line, it looks like a non-linear relationship might be a better fit for the existing points (but it’s hard to say with only three data points). One of the authors indicated to me that this region of the graph will be the subject of future study, which is fine, but extrapolating based on data provided in the present study is speculative.
6. In his analysis of this study on Sweat Science, Alex Hutchinson makes an excellent point:
“One of the proposed advantages of barefoot-minimalist running is that it automatically helps to correct overstriding — an extremely common problem among inexperienced runners. The fact that all these runners were already forefoot strikers suggests that none of them were likely overstriding, which would make them less likely to benefit from barefoot running. It’s possible that a truly “random” group of runners might have been less efficient in the shod condition, because more of them would have been dramatically overstriding.”
Alex’s point raises an interesting question – what happens if we take a traditionally shod, overstriding runner and have them begin to incorporate some barefoot training. Does their economy improve over time if they are able to modify their stride in a positive way? This would be a very interesting study to see.
All of the above being said, this study does provide some interesting and useful results. To be quite honest, I would not be surprised in the least if most people are more economical when running in a lightweight racing flat than when barefoot. Personally, I would always opt for a lightweight shoe over barefoot in a race if for no other reason than shoes allow me to cut loose and doesn’t require that I pay attention to where I am stepping. I find the ability to be a bit more reckless in a race to be of benefit. In this sense, this study actually supports my personal approach and beliefs (and yes, though I would not call myself experienced, I have run barefoot and enjoyed it on multiple occasions). However, I also suspect that most barefoot runners aren’t doing so out of an expectation of improved economy or performance (How many barefoot runners do you see in a typical race? Answer: Very few.). Enjoyment and escaping injury seem far more often cited reasons, so if that’s why you sun sans shoes, by all means keep it up.
Even more importantly, and as is the case with any good scientific research, the limitations of this study raise some excellent questions. What happens if we continue to reduce shoe weight – does the relationship remain linear? How do equal weight shoes with cushioned vs. non-cushioned soles compare (e.g., compare the Merrell Road Glove to the Merrell Bare Access). How does true barefoot running differ when compared to running in socks on a treadmill? I’m also very intrigued by the individual variation observed. Why were four of the individuals more economical when barefoot? Why were the others more economical shod? That’s what I’d really like to know.
In the end, it’s always important to keep in mind that people are different, and not everybody is going to be best barefoot, and not everybody is going to be best wearing shoes (as was the case here). Factors like fitness, training history (e.g., number of weekly barefoot running miles), and peculiarities of individual anatomy and physiology could all have a bearing on what works best at an individual level. A study like this can reveal general patterns, but it cannot tell an individual runner what is best for them. It does not “prove” that barefoot running isn’t better for you. That comes instead from knowing your own body, experimenting with different footwear conditions, and finding out what works for you as an individual.