Mind just throwing out here the names of the ones that, by your analysis, have no issues, or are honest if imperfect?
-
-
Replying to @alexandrosM @diviacaroline
The studies I would personally consider as having low/no risk of fraud (although potentially other issues) at this point would be Mahmud, Mohan, Vallejos, Babalola, Ravirkirti, Lopez-Medina, and Biber. Not everyone agrees, but that's my assessment so far
1 reply 1 retweet 2 likes -
Together is very low risk for fraud overall, but have not yet assessed as it is not published. I personally think that the Okumus study is unlikely to be fraudulent simply because no one would fake a study that bad
3 replies 1 retweet 0 likes -
Replying to @GidMK @diviacaroline
IIUC, you're saying that several studies with strongly positive results look reasonably solid by the analysis you and your collaborator(s?) did. Other than the Vallejos study, the others in your first tweet are all strongly positive, some of them showing results in the 90% range.
1 reply 1 retweet 1 like -
Replying to @alexandrosM @diviacaroline
Not really. Most of them had null results for their primary endpoint, although some did find some positive stuff on subgroup analyses. It varies
1 reply 0 retweets 0 likes -
Replying to @GidMK @diviacaroline
Can you please be specific? Which ones had null results and which didn't?
1 reply 0 retweets 0 likes -
Replying to @alexandrosM @diviacaroline
Mohan, Vallejos, Lopez-Medina had null results for their primary outcomes (as did Together). Biber reported positive results, but this may be due to protocol violations. Babalola and Ravikirti have issues, but were positive. Mahmud is positive and a solid study.
1 reply 0 retweets 1 like -
Replying to @GidMK @diviacaroline
So, what you're doing here is fascinating. 1. You use the arbitrary cutoff threshold of p<0.05 to throw out good studies. Anyone versed in statistics knows you don't throw out evidence like that. Much literature on this, but let's go with sth accessible:https://fivethirtyeight.com/features/statisticians-found-one-thing-they-can-agree-on-its-time-to-stop-misusing-p-values/ …
2 replies 0 retweets 1 like -
2. Even though you've mentioned you looked at 30 studies, the ones you mentioned here were "not fraud", that were also significant, somehow all but one have "issues".
1 reply 0 retweets 0 likes -
3. The only one you deemed acceptable combines IVM with doxycycline, so I assume your next move would be to say "gotcha!" on that one too.
2 replies 0 retweets 0 likes
I never said that
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.