234 Comments
тна Return to thread

Please start looking at proportions.

I really don't know why nobody is doing this, it's so frustrating to see you turning circles.

https://pervaers.com

Expand full comment

thanks, but without an explanation of the statistic it is hard to understand the importance of delta p. Also, I see nothing to compare with, no other ailment, no basis to understand.

Expand full comment

Oh thank you for pointing this out. I fixed the link to the help section which was dynamically generated.

Here is an image of the section in question: https://substack.pervaers.com/misc/PRD.png

Full help section, where I also explain how I created the "pseudo-placebo control report cohort": https://help.pervaers.com

I changed the confidence interval since I created this, but the math is still the same.

Mind you I am just a physical therapist who picked up coding as a hobby and had never created a website before this, so please excuse any design flaws. :D

I am pretty weak on the mathematical terminology, so please excuse that as well. Whatever math knowledge I needed for this website, I learned it on the fly.

Russ Wolfinger said I did a good job, but I don't know how thoroughly he reviewed the methods.

I don't think J. Haldane's method for confidence interval calculation of proportional differences is an ideal pick, but the differences to other methods are probably only marginal, at least for common medical concepts. It's not easy for me to translate the math from scientific articles into code, so I picked the one that I could translate. ;)

Expand full comment

Proportion analysis from pervaers.com is also available as a quick link from the list of symptoms in the results table of matchyourbatch.org. It gives a good comparison between the difference between males and females as well.

Expand full comment

Thank you! <3

Expand full comment