Invisible Data

 ... means no contradictions!

Normally, I try to ponder the meanings and implications of things. In this case, I don't understand it at all, so I need to buy some time. 

Long time friend, Ohioan, did some research on Ivermectin testing and found out some interesting stuff. First, Google is hiding the research.

I did a Google search on "Ivermectin (IVM) and CoViD-19". I found only negative papers. Or more precisely, I didn't find any positive ones on the first page of the results (as I'd expected). I kept looking into the later pages, mostly because I consider it a given that Google ("Be Evil"™) will be 'protecting' me from what they consider 'misinformation' (a.k.a. anything that does not line up with the official woke/left beliefs). Anyways, the difference between me and a person who would make a statement such as the above is that I distrust all of the media, while they seem to trust the media that tells them are smarter than everyone else. (Why does that sound familiar?) 

It wasn't hard to find two studies that are positive (and I stopped after the first two because my intent was to READ them so I could make up my own mind and hey, two, and specifically the first two from Google, disproved the statement that there is only one, and they seem like as good a good place to start as any). 

Why does Google try and shove you to the negative papers? Why not show them all and not bias the search results? Hmm.

Once Ohioan made it to the positive results, here's what he found.

Now I'm no expert on meta-analysis, nor would I claim to understand the details of the analyses, but I do understand statistics well enough to know that a study with 1,131 subjects may well be too small to be confident that they see an effect, while 3,428 might be big enough. I also note that the effect in the T+ve paper is given as "ivermectin reduced risk of death compared with no ivermectin (average risk ratio 0.38, 95% confidence interval 0.19–0.73)".  Note that a 95% confidence level (a.k.a. a p-value of 0.05) means they are reporting a 2σ error bound. That means if you want to go to the 99% confidence level (3σ), the result would not differ from the null hypothesis. It also means that a sample only 1/3 the size would almost certainly not report a result different from the null hypothesis (as σ would be bigger by roughly a factor of √3).

Conclusions: In any case, while I'm not yet willing to say that IVM is clearly an effective treatment, I would insist that at the barest minimum the book is still open as to that possibility, I see no reason to claim huge problems with the larger study (as the difference may be nothing more than the march of science), and that further studies are absolutely necessary and should be underway NOW!

I'll let you go read the whole thing, which I recommend.

I'm trying to figure out the point of the concealment exercise. I can understand the wokesters hiding sites like the Daily Wire and so forth, but why take a side on Ivermectin? I don't recall the Bad Orange Man pushing it, although maybe he did and they're reflexively banning it because of that. Something to ponder, that.

Cleaning Up The Comments

I have to admit, I stopped reading One Brow's comments a long time ago. He rarely addressed the point of my posts and his interactions with the regulars here were frustrating, to say the least. His most recent comment scrap with another long-time friend, WC Varones, is a perfect illustration.

This blog is an exploration of ideas and I'm grateful for the guidance, corrections and conversations, but recently, it's become exhausting to scroll through acres and acres of ever-shifting contradictions. This isn't a democracy, it's a blog and I don't have the time or energy to police it constantly.

Sorry, OB, but you'll need to go somewhere else.

Original Python skit here.