While I don't consider the target of your critique to be a scientist, his biased mindset and dishonest methods are far from uncommon among academic scientists. A couple of anecdotes from my 30 years in academic biology.
Shortly after I retired, I was eating at a crowded cafe next to the university that generously gave me a doctorate. At the next table were a middle-aged man and a young man, the former apparently being the graduate advisor and the latter his grad student. The conversation consisted of the advisor recounting with evident pride how he had blocked an academic rival's publication and his participation in a symposium. The man was as physically repulsive as morally, and I easily imagined that his only pleasure in life was dirty dealing.
As an evolutionary ecologist, I often took up projects in subareas that were new to me. After all, evolution intersecting with ecology spans a huge range of phenomena. Usually in publishing I had some trouble with territorial reviewers, but always managed to satisfy editors. The worst reviewer corruption, however, was appalling. That fellow's review stated that if he were put on as an author and rewrote one section, then the paper could be accepted for publication.
An editor of a journal bragged to a group of grad students that he had rejected a manuscript because it used a one-sided t-test instead of a two-sided. Rejected, not requested a recalculation and resubmission of the manuscript.
There are plenty of people in science who, while as pacific as can be, display attitudes that would be attributed to anti-social personality disorder if they were physical acts rather than attitudes and tricks of the trade.
Haha! How can you expect to be taken seriously when there are two "Bad level 6" paragraphs? Someone who can't count up to 7 is not worth reading! <GMK-mode off>
One thing, though : it could be useful to explicit somewhere that every time you say "confidence interval", you specifically refer to the "95% confidence interval", i.e the interval inside which we're 95% confident the actual value of the risk ratio stands. A 90% confidence interval would be narrower, for example, and a 99% confidence interval much wider. (And a 100% confidence interval would go from 0 to infinite.)
The Ivermectin debacle is another in a series of public questions that have filled me with epistemological doom.
It's almost impossible for me to discover the truth and to make matters worse, in the last 6 years, Fiat censorship has reared its ugly head on the internet.
As a pest control guy, my newest approach is to simply guess that only truth that pushes back at centralized power is censored.
Therefore, it's easy to conclude that Ivermectin works, the jabs don't work and cause death, and that Covid was always overdiagnosed.
Maybe I'm wrong but I can't see how my epistemolgical approach is any worse than Scott Alexander's.
At first I thought this was going to be a beating on a dead horse. But I'm surprised there's a lot of attention to how to work and to details. I enjoy that the review of Scott work brings me gives me humility when reading someone else work. Since most of my hubris comes from my lack of knowledge.
Great article and this whole experience has been very enlightening because it really makes me question how gullible I can be when reading scientific-sounding people (as I felt when originally reading Scott's easy).
One quibble though on this:
> If we’re being strict about it, any evidence that emerged after this tweet should be considered tainted, given that this is a pretty direct threat for the careers of any perceptive scientist who sees it:
The linked FDA article does note, "the FDA has received multiple reports of patients who have required medical attention, including hospitalization, after self-medicating with ivermectin intended for livestock."
So the simpler explanation of that tweet is that some non-scientist that runs the FDA's social media was emphasizing that point in a catchy way.
I think the better argument about unscientific bias would be the media lying about Joe Rogan after Rogan mentioned Ivermectin, although that doesn't have nearly as direct a line to the government and/or Pharma companies.
I'm afraid that as with everything related to the pandemic response this response to Alexander's piece is as far as its going to get. That for those that want to believe in the FDA and CDC and mainstream stances on all the issues, they get their studies and think pieces which are published and promoted full force. While those of us convinced that there was something to early treatments have works like this one, which spell it all out and affirm that there is more than enough evidence to to back our heterodox views. But this won't go mainstream and will only be circulated among those of us that already believed. At least this has helped me with my own sense making. Thanks.
I feel much the same. There is a disbelief that the system could be this broken, ergo, the dissidents are wrong.
We have had to put up with some serious craziness during this pandemic. Investigators seriously overdosing a drug in a trial and then trying to say it doesn't work. Fraud allegations by the BMJ against the Pfizer vax trials, and then Pfizer trying to hide their raw data for 75 years. Tess Lawrie's recorded conversation with Andrew Hill where he admitted being pressured by a sponsor to change the outcome of his analysis. The Together Trial still refusing to supply trial data for their results. Papers being retracted without a reason given.
I do not know the truth of some of these matters, but what I do know is that Tess Lawrie's assertion that pharmaceutical trials need to be quadruple-blinded, rather than double-blinded, and that other means to prevent games being played with methodologies, needs to be implemented henceforth. That's a start.
I think its a bit crazy the manufacturers run the trials of their own products. Its seems off from a check and balance standpoint but I say that with complete ignorance as to how the process was setup in the first place. I think there needs to be formal institution that will advocate for generic drugs. Perhaps such a project could also look to advocate for the public domain with an open-source mentality.
Definitely agree on the lack of checks and balances here. This would be less of a problem if the big players did not have the most money in the system with, thus, the greatest ability to run large trials. The independents can run only smaller trials which are then handwaved away.
I do worry that if you set up an institution, it will eventually suffer the same problems of capture that everywhere else seems to be suffering. But, yes, generics need somebody who is invested in going into bat for them, and who can also negotiate the costs and bureaucracy around the approval process. Even more than generics, we have studies on exercise, diet, and various supplements that looked promising and were never really followed up on. (And exercise and diet would have so many other beneficial effects outside of COVID.)
What would happen if you could somehow set up a system that blinded the regulators? I have no idea how you could make that happen though. "Here are Drug A's results from multiple studies. Here are Drug B's results. Here are Drug C's results. Please evaluate."
While I don't consider the target of your critique to be a scientist, his biased mindset and dishonest methods are far from uncommon among academic scientists. A couple of anecdotes from my 30 years in academic biology.
Shortly after I retired, I was eating at a crowded cafe next to the university that generously gave me a doctorate. At the next table were a middle-aged man and a young man, the former apparently being the graduate advisor and the latter his grad student. The conversation consisted of the advisor recounting with evident pride how he had blocked an academic rival's publication and his participation in a symposium. The man was as physically repulsive as morally, and I easily imagined that his only pleasure in life was dirty dealing.
As an evolutionary ecologist, I often took up projects in subareas that were new to me. After all, evolution intersecting with ecology spans a huge range of phenomena. Usually in publishing I had some trouble with territorial reviewers, but always managed to satisfy editors. The worst reviewer corruption, however, was appalling. That fellow's review stated that if he were put on as an author and rewrote one section, then the paper could be accepted for publication.
An editor of a journal bragged to a group of grad students that he had rejected a manuscript because it used a one-sided t-test instead of a two-sided. Rejected, not requested a recalculation and resubmission of the manuscript.
There are plenty of people in science who, while as pacific as can be, display attitudes that would be attributed to anti-social personality disorder if they were physical acts rather than attitudes and tricks of the trade.
I don't think Scott Alexander is lying. I think he is doing a half assed analysis. Unfortunately his blog is being used to push a false assesment
Thank you for all this great work. I just commented over at Scott's substack, urging him to read this and correct his article.
He should buy Alexandros dinner and apologist profusely for the work he's made him do
He should republish Alex's series on his own blog.
Haha! How can you expect to be taken seriously when there are two "Bad level 6" paragraphs? Someone who can't count up to 7 is not worth reading! <GMK-mode off>
One thing, though : it could be useful to explicit somewhere that every time you say "confidence interval", you specifically refer to the "95% confidence interval", i.e the interval inside which we're 95% confident the actual value of the risk ratio stands. A 90% confidence interval would be narrower, for example, and a 99% confidence interval much wider. (And a 100% confidence interval would go from 0 to infinite.)
Fixed and fixed. Thank you sir!
The Ivermectin debacle is another in a series of public questions that have filled me with epistemological doom.
It's almost impossible for me to discover the truth and to make matters worse, in the last 6 years, Fiat censorship has reared its ugly head on the internet.
As a pest control guy, my newest approach is to simply guess that only truth that pushes back at centralized power is censored.
Therefore, it's easy to conclude that Ivermectin works, the jabs don't work and cause death, and that Covid was always overdiagnosed.
Maybe I'm wrong but I can't see how my epistemolgical approach is any worse than Scott Alexander's.
At first I thought this was going to be a beating on a dead horse. But I'm surprised there's a lot of attention to how to work and to details. I enjoy that the review of Scott work brings me gives me humility when reading someone else work. Since most of my hubris comes from my lack of knowledge.
Many have made a similar comment, and to be honest I have also been surprised with the depth and breadth of discussion topics this review opens up.
I feel like you already destroyed and buried his reputation. Now you're just dancing on his grave 🤣
That's reasonable to me. 👍
Great article and this whole experience has been very enlightening because it really makes me question how gullible I can be when reading scientific-sounding people (as I felt when originally reading Scott's easy).
One quibble though on this:
> If we’re being strict about it, any evidence that emerged after this tweet should be considered tainted, given that this is a pretty direct threat for the careers of any perceptive scientist who sees it:
The linked FDA article does note, "the FDA has received multiple reports of patients who have required medical attention, including hospitalization, after self-medicating with ivermectin intended for livestock."
So the simpler explanation of that tweet is that some non-scientist that runs the FDA's social media was emphasizing that point in a catchy way.
I think the better argument about unscientific bias would be the media lying about Joe Rogan after Rogan mentioned Ivermectin, although that doesn't have nearly as direct a line to the government and/or Pharma companies.
I'm afraid that as with everything related to the pandemic response this response to Alexander's piece is as far as its going to get. That for those that want to believe in the FDA and CDC and mainstream stances on all the issues, they get their studies and think pieces which are published and promoted full force. While those of us convinced that there was something to early treatments have works like this one, which spell it all out and affirm that there is more than enough evidence to to back our heterodox views. But this won't go mainstream and will only be circulated among those of us that already believed. At least this has helped me with my own sense making. Thanks.
I feel much the same. There is a disbelief that the system could be this broken, ergo, the dissidents are wrong.
We have had to put up with some serious craziness during this pandemic. Investigators seriously overdosing a drug in a trial and then trying to say it doesn't work. Fraud allegations by the BMJ against the Pfizer vax trials, and then Pfizer trying to hide their raw data for 75 years. Tess Lawrie's recorded conversation with Andrew Hill where he admitted being pressured by a sponsor to change the outcome of his analysis. The Together Trial still refusing to supply trial data for their results. Papers being retracted without a reason given.
I do not know the truth of some of these matters, but what I do know is that Tess Lawrie's assertion that pharmaceutical trials need to be quadruple-blinded, rather than double-blinded, and that other means to prevent games being played with methodologies, needs to be implemented henceforth. That's a start.
I think its a bit crazy the manufacturers run the trials of their own products. Its seems off from a check and balance standpoint but I say that with complete ignorance as to how the process was setup in the first place. I think there needs to be formal institution that will advocate for generic drugs. Perhaps such a project could also look to advocate for the public domain with an open-source mentality.
Definitely agree on the lack of checks and balances here. This would be less of a problem if the big players did not have the most money in the system with, thus, the greatest ability to run large trials. The independents can run only smaller trials which are then handwaved away.
I do worry that if you set up an institution, it will eventually suffer the same problems of capture that everywhere else seems to be suffering. But, yes, generics need somebody who is invested in going into bat for them, and who can also negotiate the costs and bureaucracy around the approval process. Even more than generics, we have studies on exercise, diet, and various supplements that looked promising and were never really followed up on. (And exercise and diet would have so many other beneficial effects outside of COVID.)
What would happen if you could somehow set up a system that blinded the regulators? I have no idea how you could make that happen though. "Here are Drug A's results from multiple studies. Here are Drug B's results. Here are Drug C's results. Please evaluate."
I've been finding Jenny Holzer's admonition more and more applicable these days: "Abuse of power comes as no surprise."