[Update 1: Minor text updates and added write up to this post instead of just linking to the original post]

Recently there was a study by Dr. David Carrier on the efficiency of different foot positions while walking, most notably a heel-strike vs a forefoot strike.  The results showed that heel striking was more efficient than forefoot striking while walking.  Many, particularly in the barefoot community, disagreed with some of his methods and results.  In particular the choice of a forefoot strike that did not have the heel drop fully to the ground struck many as different than the forefoot strike that they might use with the heel dropping completely to the ground.

I did a little experiment to begin looking at these questions that have been raised.  The initial experiment was very small, just a cursory look at the question, but it replicated Carrier’s results and gave an initial answer to the question of whether an alternate footstrike would have shown different results.  In this preliminary experiment, a heel strike walking gait was more efficient than a forefoot strike with or without the heel contacting the ground, and there was not a significant difference between efficiency of a forefoot strike with and without the heel contacting the ground.

This article at present is placeholder for one that I intend to expand upon with a less hurried experiment, analysis, and write-up than I post here, but for now here is a cleaned up copy of post where I quickly describe it, after the jump.

“Some new science about how we evolved to walk”

Very quick write up of a very quick experiment.

Short version:  In my own little experiment heel strike was more
efficient than ball-heel-ball or the low-digitgrade from Carrier’s
paper, while the ball-heel-ball and low-digitgrade did not
significantly differ.  One subject with three measurements on three
gaits.

Long version:

Heart rate monitor used to measure energy expenditure on treadmill set
to 3.5mph.

Heel-strike (plantigrade), Ball-heel-ball, and Forefoot strike (low-
digitgrade) compared.

3 minute warmup on each, one minute of observation with 3 data points
at 30 second intervals, 3 minute seated rest between observations.

ANOVA comparison of bpm by gait, bpm was significantly affected by
gait.

Mean(sd) of each:
Heel strike:  102.33(1.53)
Ball-heel-ball:  115.33(2.08)
Forefoot strike:  117.67(2.08)

T-test for pairwise comparisons.
Heel strike v. ball-heel-ball:  p=0.0014
Heel strike v. Forefoot:  p=0.0008
Ball-heel-ball v. Forefoot:  p=0.2417

Heel strike was significantly lower than either ball-heel-ball or
forefoot, while ball-heel-ball and forefoot did not significantly
differ.