An Interesting Comparison

[pubrunner]

Has anyone else seen this interesting road test ?

http://www.bmj.com/content/341/bmj.c6801.full

I'll think I'll hold on getting a carbon bike and buy another 6 steel frames

 

[PpPete]

This is a very damaging and thoroughly offensive study.
My wife has read it and has added it to armoury of reasons why I should not buy N+1

 

[rockyraccoon]

Expensive light bicycles do not get you to work faster

 

[Tollers]

Sample size is way too small to establish statistical significance.

Also, there is no control in this 'multivariate' test. Despite establishing an error rate and confidence interval saying his study is not statistically significant, this data includes external factors and hasn't isolated carbon vs steel as the only factor influencing his test. It needs an effective control to understand the natural variation in timings and errors by utilising one of the bikes again as another sample to understand what the natural error rate is due to external factors. Once that's done....keep riding to improve sample size.

The telegraph article adds no weight here. This really reminds me of the book 'bad science' although it's not as bad as Gillian McKeith's plain wrong understanding of science.

 

[Fiona N]

Also the issue of bike fit was never explored. The author merely notes that the carbon-framed bike was more uncomfortable - how much of this was due to poor fit?

My own less substantial, admittedly, experiment suggests that other things (weather in particular) being equal, my carbon-framed, perfectly fitted bike is about 3 minutes (< 4%) faster on average than my less well-fitted Al/C-framed bike over a regularly ridden, undulating 40km circuit. There is only about 1 kg difference between the bikes and no detectable difference in my weight over the period of the experiment.