Normalization of Running Speeds

[Petcu Daniel]

Re: Normalization of Running Speeds

Maybe will be of help https://www.ncbi.nlm.nih.gov/pubmed/22019141
Daniel

[Douglas Powell]

Re: Normalization of Running Speeds

Hi Josh,
I believe Cris Stickley at the University of Hawaii was working on a normalization paradigm for lower extremity kinetics during running based on speeds a couple of years ago. You may wish to reach out to him at cstickle@hawaii.edu.
DP

[Ted Andresen]

Re: Normalization of Running Speeds

Josh,

Sorry for the late reply. Perhaps this small dataset may help answer your question.

Six years ago I was interested in the relation between gait parameters (speed, step frequency/rate), anthropometrics (height, BM, age, gender) and training schedules for average road runners. About 40 out of 60 runners in a local 5K run volunteered. From a video at the 3K point I was able to estimate each participant's step frequency. I had wanted to capture their leg stiffness, so I could estimate their Mass Specific Power, but that didn't work out.

If you can work with Excel, you might examine the graphs to the right of Column T -- set the Zoom to 50% -- and see if the data might be of use. (The lower graphs are the participant's running velocity normalized to their Froude velocity.) Using sort and delete, you could pull out different sub-categories and examine their running biomechanics.

At another time I became interested in understanding how the Spring Mass Model connected with the step frequency across a runner's velocity range from a very low speed (v=0.5 m/s), through their walk-run transition speed and up to their max. I was interested in the "signature" of the SF(v) curve. If you're interested, I'll post that graph.

Hope this helps,

Ted

[Morgan Sangeux]

Re: Normalization of Running Speeds

Hi Josh,

Interesting question. Two years ago I would have replied non-dimensional scaling is the way to go, but I am now a bit more uncertain.
If you have a sizable sample with good variability within the sample for these covariates (sex and size), I would probably run a multiple regression to investigate any relationship, and then residualise to remove the statistical effect of the covariates.
I would still perform non-dimensional scaling according to At Hof 1996 to have a look, and check if there is a residual relationship with size, because it could be interesting to (try to) find out why.

Good luck,

[Josh Isherwood]

Hello,

We are looking to compare different sub categories of runners for their running biomechanics (gender and body dimensions). In order to meaningfully compare these groups would there be a need to normalize running speed between people in order to be able to compare bigger and smaller runners. If so, how would one do that?

Thanks

Josh