John Slough II

Data Scientist & Statistician

[John Slough II @JohnSloughII]

@DrEricRodgers @theproof @ItIsMikeFitz @YouTube @realDaveFeldman That was a small contextual part of the podcast, 30 seconds out of a 3.5 hour podcast, has nothing to do with the main critique in the podcast or prior video. Not worth spending any time on it.
I actually agree that the Cleerly dataset is anomalous.

[John Slough II @JohnSloughII]

@realDaveFeldman Our critiques in the prior video:
PAV Y-axis range relabeled from 0-20% to 0-10%
TPS Y-axis relabeled from 6 to 4
linear model assumptions violated
weak modeling, overstated results
No reporting of TPS models
Bayes factor prior wider than ultra-wide
and more

[John Slough II @JohnSloughII]

@realDaveFeldman Our critiques are:
misdefined central “exposure” variable
False/fabricated “IQR” metric
No ApoB model reported
Incoherent TOST procedure
Incomplete reporting on revascularization
post-hoc power justification
crude, fragile statistical modeling
and more…