"Wattage"-keskusteluryhmässä tätä asiaa hiljan puitiin ja itse Andrew Coggan vastasi seuraavasti.

Ryhmän osoite https://groups.google.com/forum/#!forum/wattage access pitää anoa, mutta kyllä sen joku myöntää kohtuu pian.

Kysymys:
"Anyone here tested Typical Error of Measurement of 20 min tests?"

Andrew Coggan:
"I am not aware of any studies focused specifically on 20 min, but the CV of TT-type performance is typically 2-5%, e.g.,:

https://www.ncbi.nlm.nih.gov/pubmed/3325488


https://www.ncbi.nlm.nih.gov/pubmed/16767608



https://www.ncbi.nlm.nih.gov/pubmed/11528349


The CV of time-to-exhaustion tests tends to be greater, e.g., in one of my dissertation studies I found a CV of almost 10%:


https://www.ncbi.nlm.nih.gov/pubmed/3325488


However, when studying an intervention the effect size also tends to be larger, such that both approaches have similar sensitivity:


https://www.ncbi.nlm.nih.gov/pubmed/18379223


Since this sort of research is relatively easy to conduct, and since there has long been a strong emphasis on research methods in many old-school PE programs, the above references really only represent the tip of what is a very large iceberg."