Going the extra mile doesn’t always make things better (but then again, it might)

I picked this study for two reasons: 1) It’s actually not a half-bad study, and 2) It addresses a significant fitness issue that has been plaguing athletes, trainers and coaches for decades–to stretch or not to stretch. However, despite the study’s many strengths, it falls just short of making it truly useful in helping active people make the decision whether or not to perform static stretching.

Kokkonen J, Nelcon AG, Eldredge C, Winchester JB. Chronic static stretching improves exercise performance. Medicine and Science in Sport and Exercise. 39(10): 1825-1831, 2007.


I really liked this introduction. The authors did a great job of presenting the literature, including why the previous literature doesn’t really give us any definitive answers on whether non-pre-event static stretching has detrimental effects on performance. They define their boundaries early on and don’t dispute the clear literature base on pre-event stretching. This is not the issue here. However, they make a great case as to why we are still in the dark with regular static stretching and where it should be placed (if at all) in a training program. Clearly, it is not before significant events (which, for many active people, includes their “workout”–whether that be lifting weights, going for a run or otherwise), but is there a place for stretching on say, rest days, or post-workout? And if so, what kinds of benefits might we see from it?

The authors go on to review the literature on performance benefits derived from static stretching, most of which are strength related. Yes, that’s right. Stretching might make you stronger. Well, some of you anyways, as we will see.

The purpose of this study therefore, was to determine whether a static stretching program could have an impact on strength, muscle endurance and power. (For those of you who are scratching your heads at “static stretching”, this refers to the “stretch and hold” variety of stretching–as opposed to “bounce” or “move” varieties, also referred to loosely as “dynamic stretching”)


The authors recruited 40 students (why 40, I have no idea, as there is no sample size justification in this paper) who were attending Brigham Young University in Hawaii (Damn it, I should have gone to university in Hawaii). The authors describe these students to be either inactive or recreationally active. They further narrowed this vague term down to exclude anyone who did specific endurance or strength training on multiple days during the week; as well as anyone who did more than 60 minutes of physical activity more than 3 times per week. This basically excluded anyone who lifted weights more than once a week, anyone who went for a run more than once a week (strength and endurance training, respectively), anyone who played basketball for more than 60 minutes, more than 3 times per week, and anyone who went for a walk for more than 60 minutes, 3 times per week. They also excluded people who did any sporting activities more than 6 times per month. Basically if they thought you were doing any kind of structured or planned physical activity, you got booted from the study. And from the sounds of the description, they were fairly strict about it.

[The authors’ selection of subjects is both one of the biggest strengths and biggest weaknesses of this study. In order to isolate the effect of stretching on strength, muscular endurance and power, they chose to study people who were basically sedentary. This has two advantages: 1) If you study a group of people who do basically no physical activity other than stretching, then improvements in physical activity testing can be attributable to the stretching; and 2) you are most likely to see the biggest changes in those people who are furthest away from their “athletic potential”. You are least likely to see an improvement in reading ability if you study lawyers than if you study third graders (debate what you will about lawyers, but they read good). The same principle applies here.

The drawback to this approach, is naturally, the extent to which we can generalize these results to anyone but basically sedentary people. If you study third graders’ reading, you can’t go and claim that your revolutionary program is going to improve the reading ability of the lawyers–or even fifth graders. BUT, given that we know so little about static stretching and performance in this context, this study IS the most appropriate starting point, as it gives us the best chance of detecting a benefit. If these researchers had gone with “trained college men” and NOT found a difference, we would still be in the dark because that population is in the middle of the road. Given that the study finds that static stretching improves performance in the least-trained individual, we can then start working our way up to see if this effect persists with more-trained people.]

Subjects were “randomly assigned” to either the stretching group or the control group. The control group were asked to refrain from any stretching activities, but otherwise, the groups were told to maintain what little physical activity they were doing. The researchers also made sure that there were the same number of men and women in each group. Every had to keep a log of their activities.

[I really wish they had gone into just a little more detail about randomly assigning people to groups.]

In addition to all of this, the authors report that all local recreational facilities were monitored by research staff and, “…the presence of any subjects was noted and collaborated with their exercise logs.”

[Holy big brother, Batman. This is by far, the furthest extent to verifying non-activity I have ever encountered. A little creepy perhaps, but from a scientific perspective, one of the more rigourous ways of ensuring subject compliance (or non-compliance, as the case may be) ]


The authors tested the following things:

1) Sit and reach (flexibility)
2) 20m sprint time (power)
3) Vertical jump height (power)
4) Standing long jump length (power)
5) Knee extension and flexion 1RM (strength)
6) Knee extension and flexion endurance (endurance)
7) VO2 peak (endurance)

Testing was done over 3 days. The details of this aren’t that important. If you’re really interested, read the article.

[We are not told who is blind to what in this article. So, for all we know, everyone know what everyone is doing (i.e. subjects, evaluators, group-assigners, the gym spies…I still can’t get over that.) ]

The stretching program:

The stretching program was a 10-week program with 15 different stretches. Each stretch was held for 15 seconds and each stretch was repeated 3 times. All stretches were performed non-assisted, and 12 of the 15 stretches were performed again, but with assistance. Stretching sessions lasted about 40 minutes each.

[I won’t go into each of the individual stretches. They are described though. Forty minutes is a long time to stretch though, even for a static stretching program. I would say that this is perhaps one of the bigger weaknesses of the study. How feasible would this be if this were you?]


This is by far the least liked part of the study for me. The authors decided to treat the data as four separate experiments, in order to avoid adjusting their alpha-level to oblivion with 7 multiple tests. So, instead of having to meet an alpha-level of 0.007 (which is 0.05 divided by 7), they had to meet an alpha-level of 0.05 for the “flexibility experiment” (only one test there), 0.017 for the “power experiment” (3 tests), etc. There is an argument that can be made for this kind of strategy–certainly the “experiments” are grouped logically and not haphazard, but given that it’s all on the same population, it’s still ONE experiment. However, this point is relatively moot when we get to the results.

Otherwise, statistical testing was appropriate with a two-way ANOVA with the protected Tukey as a post-hoc test.

[I’m always a bit conflicted with respect to using the two-way ANOVA in these studies. Basically, you’re trying to see if a significant difference exists either between the two groups (the first way), or within each group (the second way), or both. However, in the context of a randomized controlled trial, we are really not interested in whether one group did better compared to itself (i.e. the stretching group improved their, say, strength, when compared to their pre-testing), because whether the stretching group did better compared to itself is irrelevant, if the control group ALSO did similarly well (significantly, or not). So why bother seeing if the groups do better compared to themselves?]


Thirty-eight subjects finished the study. One subject from each group dropped out because they wanted to exercise. Of the remaining subjects, no one exercised for more than 18 days of the 10 weeks. No one in the stretching group missed more than 2 stretching sessions.

[Overall, I did not like the reporting style, nor the statistical reporting of this paper. Everything was reported and interpreted in percentages, despite the authors providing the actual numbers in a table. This leads me to wonder whether the statistical tests were done with the percentages as well, as opposed to the actual number results–which has implications for the likelihood of uncovering a significant difference, since differences may become magnified once converted to percentages. This includes confidence intervals, which were used, and I think any authors who use them should get some kudos. I just wish they had done them on the actual numbers]

Flexibility: The stretching group was more flexible than the control group after 10 weeks. I’m sure you’re all floored. Fortunately, the control group’s flexibility did not change (on average, it got worse), which is a good thing because even if they DID stretch on the sly (away from the piercing eyes of the gym spies), they didn’t get more flexible, so we can still make statements about the effects of increasing flexibility on other things.

I could go through all of the outcomes, but in the end, the stretching group always did better than the control group from a percentage point of view.

However, when I look at the all-important results table with the actual numbers, what I see is very very small improvements, with not a whole lot of convincing numbers to show that the “significant differences” detected through percentages carry through with the raw performance numbers. For example, the stretching group ended up with an average knee extension 1RM of 82.0 kg, but a standard deviation of 25.8, while the control group had an average 1RM of 71.0 kg, but a standard deviation of 20.8! While the difference is 11 kg, on average, the variance is SO large that I would be surprised if that turned out to be statistically significant. Less dramatic would be something like the 20m sprint time: Stretching group average 3.8 (SD 0.51), control group average 3.68 (SD 0.31), both post- values. Even if that was statistically significant, I’m not sure that a difference of 0.18 seconds is anything to write home about–remembering that we are not talking about Olympic athletes upon which tenths of a second are important.

So, in the end, what we have is a bit of a mixed bag of a study. It has some REALLY strong points (I mean, KUDOS for spying on all your subjects at the gym!), and a few weak points, but overall, fails to produce data that shows us that static stretching is a worthwhile pursuit, unless your life is judged on percentages alone. Given that one would have to stretch for FORTY minutes, 3 times a week, I can think of several things I would rather do for 40 minutes, 3 times per week, which might produce the same (i.e. possibly very little) _absolute_ benefit in strength, power and muscular endurance.

Still, I suppose if I was unable to do any activity other than stretching, I could be convinced that it might prevent me from deteriorating too much. And that’s not necessarily a benefit to sneeze at, given that there are many athletes who are sidelined into inactivity due to injury or other circumstances, not to mention several patient groups in inactivity for health reasons. So, in a way, it all depends which side of the fence you’re approaching the problem.

The bottom line:

If you can’t do anything else, static stretching for 40 minutes, 3 times a week could keep you from deterioration from a muscular strength, power and muscular endurance perspective.

Click Here to view the Full Version of our Website