The report Dave was referring to is pretty easy to find using a database
such as PubMed.  I've included the citation and abstract below.  More
importantly, the issue of having 1538 subjects and difficulty controlling
all variables, and Henry's contention that this may make it a poor study
should be addressed.

If you use a small, homogenous sample, you may be able to control many
confounding variables, and find a treatment effect if one exists.  This
concept is referred to as "internal validity", and in essence means that the
researcher can attribute the treatment effect to the treatment applied with
confidence.  However, the price paid for high internal validity is low
external validity, or the ability to generalize the results outside the
experimental setting or to subjects who are not similar to those included in
the study.

The clear advantage to an RCT is the first word; randomized.  The process of
randomization is the absolute best means to control for the effects of ALL
confounding variables, even those whose existance the researcher is ignorant
to.  With a small sample, randomization is not very effective, but with
larger samples, randomization works.  Therefore, if 1538 subjects were
randomized to either a control or a treatment group, even without rigorous
supervision of the stretching, randomization would likely have taken care of
the confounding variables.

One cannot make any definitive statements about research by reading only the
abstract.  Because the research by Pope et al did not find a statistically
significant difference between the control and stretch groups, the issue of
statistical power should be addressed.  If the power was sufficiently high,
then the researchers can state with confidence that a type II statistical
error (failing to reject a false null hypothesis) is improbable.

Frank Underwood



Pope RP, Herbert RD, Kirwan JD, Graham BJ.  A randomized trial of
preexercise stretching for prevention of lower-limb injury.  Med Sci Sports
Exerc 2000 Feb;32(2):271-7 

Physiotherapy Department, Kapooka Health Centre, New South Wales, Australia.
[log in to unmask]

PURPOSE: This study investigated the effect of muscle stretching during
warm-up on the risk of exercise-related injury. METHODS: 1538 male army
recruits were randomly allocated to stretch or control groups. During the
ensuing 12 wk of training, both groups performed active warm-up exercises
before physical training sessions. In addition, the stretch group performed
one 20-s static stretch under supervision for each of six major leg muscle
groups during every warm-up. The control group did not stretch. RESULTS: 333
lower-limb injuries were recorded during the training period, including 214
soft-tissue injuries. There were 158 injuries in the stretch group and 175
in the control group. There was no significant effect of preexercise
stretching on all-injuries risk (hazard ratio [HR] = 0.95, 95% CI
0.77-1.18), soft-tissue injury risk (HR = 0.83, 95% CI 0.63-1.09), or bone
injury risk (HR = 1.22, 95% CI 0.86-1.76). Fitness (20-m progressive shuttle
run test score), age, and enlistment date all significantly predicted injury
risk (P < 0.01 for each), but height, weight, and body mass index did not.
CONCLUSION: A typical muscle stretching protocol performed during
preexercise warm-ups does not produce clinically meaningful reductions in
risk of exercise-related injury in army recruits. Fitness may be an
important, modifiable risk factor.


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%