Deep phenotyping of post-infectious myalgic encephalomyelitis/chronic fatigue syndrome, 2024, Walitt et al

you can test for normality and test for equality of variances and then triage each feature to different tests depending on whether parametric/nonparametric or variance equality, can give you more statistical power while still being rigorous

is covered for excel here in figures 6 & 7 https://pubs.rsc.org/en/content/articlelanding/2020/mo/d0mo00087f

 

Thanks. I'll see if I can parse that, but these concepts are getting a bit advanced for my beginner stats knowledge.

 

If you install realstats addin for excel and use the formulae as in the paper, it will do everything for you automatically so you need not worry about the theory being over your head to execute it. In terms of concepts the embedded image is from the paper and is amazing at helping to understand the logic.

If you'd like some help feel free to email me D.Missailidis@latrobe.edu.au

You've been given good advice anyway (when in doubt, mann-whitney U test! no assumptions made) not trying to micromanage etc

 

Something that seems to happen a lot in ME research is that the patient group shows much more variance than the healthy controls (as in the red group compared to the blue group here: most of the red lines slope sharply up or down while most of the blue lines stay nearly horizontal), but then the results are averaged to cancel out any differences and the conclusion is that there are 'no differences'.

I'm sure it's statistically sound but I can't help wondering if something's being missed!

 

I have forgotten my statistics, but if someone said that a family where everyone was either four foot tall or eight foot tall was statistically indistinguishable from a family where everyone was six foot tall is not doing the right statistics.

 

Difference might be a little more visually apparent if normalizing based on fat free mass (order is baseline total, baseline sleep, day 1 total, day 1 sleep, day 2 total, day 2 sleep, day 3 total, day 3 sleep)



Here's the fat free mass for each group:

(Edit: I accidentally posted the wrong mass chart at first.)

I did notice the study's figure in the post above has one feature that is statistically significant for the metabolic chamber study, respiratory quotient on the 3rd day (p=0.01). (MECFS: 0.87 +- 0.04, HV: 0.89 +- 0.01)