That's not an evidential appraisal
-
-
-
Replying to @DrPaulND @auscandoc and
Good to know you don't base your assessment of homeopathy on evidence. I imagine that's how most homeopaths keep practicing
1 reply 0 retweets 1 like -
Also, I'm interested to know what methodology was used in that meta-analysis that you are citing, because I re-ran the code in Stata 15 using the metan command with their exact odds ratios and CIs and it was entirely non-significantpic.twitter.com/V5FCEAHGVz
2 replies 0 retweets 1 like -
Replying to @GidMK @auscandoc and
I'd advise discussing that with mathie himself.
1 reply 0 retweets 0 likes -
Replying to @DrPaulND @auscandoc and
It's something that should really be reported in the study itself. Given the general quality of this MA, it's very odd that they didn't do so, unless I'm wrong and there's a report buried in the supplementary materials somewhere
1 reply 0 retweets 0 likes -
-
Replying to @DrPaulND @auscandoc and
Here we go, this is much more similar. Re-ran the model exponentiating differently. Looks like their basic analysis probably wasn't an error, although in the process I've found at least 2 minor numerical errors in the paper so it's a very mixed bagpic.twitter.com/HaxipIKC8F
2 replies 0 retweets 0 likes -
Replying to @GidMK @auscandoc and
Were thes minor numerical errors sufficient to affect the results?
1 reply 0 retweets 0 likes -
Replying to @DrPaulND @auscandoc and
Hard to tell without doing a truly annoying amount of work. Sadly, I definitely don't have time. At a guess, I'd say probably not
2 replies 0 retweets 0 likes
Although one thing that was truly weird is that they took studies with nonsignificant mean differences and instead used the significant risk ratios. Again, without doing an enormous amount of work, hard to know what impact this might've had
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.