> Beware that semi-educated guesses are sometimes worse than uneducated
> ones, since they are more plausible and not necessarily more correct.
This is one of the most insightful comments I've read on the internet ever.
This is why we shouldn't take medical advice from people who have read a few Wikipedia articles. This is why years of experience in e.g. software are necessary to even begin to understand how deeply hidden security issues can be. This is why self-driving cars still elude us, but will be so much safer than human drivers when they finally do arrive.
I was thinking the opposite point: sometimes the experienced doctor or researcher made a mistake, and the guy who posted something on Reddit is correct. Usually not, but sometimes.
Generally speaking, you should trust someone whose professional or experienced more than some random stranger. Especially if it's not one person but a large group of experienced professionals (most of the controversial "anti-science" issues today).
But it's actually important to "do your own research", and not just listen to authority. Provided you actually know how, i.e. you can find accurate sources and distinguish fact from fiction. Because even the most qualified, unbiased, genius authority are occasionally wrong. The big issue today is that a lot of people don't know how to research and distinguish lies, but (although it's probably impossible for most) finding a way to teach someone how to find the truth, will always be more effective having some authority tell them the truth.
Key example: doctors often miss symptoms and diagnoses that a patient can discover on their own. It's not that the doctor is stupid or unqualified, in fact they have way more medical knowledge than you. But the doctor only has a series of questions and maybe a few tests to make a diagnosis, whereas you have the full experience of symptoms and an internal understanding of your "normal". IMO people actually should "play Internet doctor" and research their symptoms, not to reach any 100% conclusions, but to suggest diagnoses and tests to their real doctor.
> IMO people actually should "play Internet doctor" and research their symptoms, not to reach any 100% conclusions, but to suggest diagnoses and tests to their real doctor.
I agree, and do this. But I never tell the doctor "I suspect that I have X", rather I tell the doctor "I've observed J, K, and L under Foo conditions" and let the doctor decide where to go from there. In short time we can learn to give the doctors better information, but we cannot replace their breadth of knowledge.
Yeah sometimes people say "doctor I have X, I'm trying to get diagnosed with X". You're biasing the doctor to give you a diagnosis so you don't drop them, and you're biasing the doctor to ignore your symptoms because you're just saying them to get the diagnosis and maybe you just want to get certain medication.
Even if you don't outright say it, you should never convince yourself you have a diagnosis. The human body is extremely complex and some particularly weird issues need patient research, doctor's research, and experimentation to solve.
I recently had an appointment with specialist for my son. My son has ADHD but was never diagnosed within Kaiser and so Kaiser won't prescribe medication.
The appointment started with, "My son has ADHD, was first diagnosed at 7 by _____, here are volumes of paperwork demonstrating that he had this diagnosis, has continued to have symptoms, and documenting what we have done for him. He wants to go back on medicine despite the side effects from last time. Here is more documentation that he continues to have symptoms."
There really ARE times when you're coming with enough background that the diagnosis is never in doubt.
For a less recent but more trivial example, I once showed up to the doctor and said, "I broke this bone in my hand while playing volleyball, what should I do about it?" Again, the diagnosis was never in doubt.
If you manage to hit the exact diagnostic questions they were trained for, that's the same, actually worse, than saying you have X, because you'll make them think they had the idea that you have X, when in reality you saw the list of symptoms, thought "I guess I have those," and then repeated them to the poor unsuspecting doctor.
>This is why we shouldn't take medical advice from people who have read a few Wikipedia articles.
>sometimes the experienced doctor or researcher made a mistake, and the guy who posted something on Reddit is correct
Both of those points are true, and we can reconcile them thusly:
Broadly, the doctor is the fully educated, while the layman-interested-in-subject is making semi-educated guesses. However this relationship inverts on narrow subjects, like a particular rare disease: the doctor ends up making a semi-educated guess (based on his general medical education), whereas a layman with intense interest in this particular rare disease can easily be the fully educated one - again, on this particular narrow subject.
> I was thinking the opposite point: sometimes the experienced doctor or researcher made a mistake, and the guy who posted something on Reddit is correct. Usually not, but sometimes.
If a researcher made a mistake and if 50 people on reddit said 50 different things, at random, one might be correct by chance.
Being correct once means nothing unless you are correct most of the times, but by then you are the researcher in the scenario above.
>Generally speaking, you should trust someone whose professional or experienced more than some random stranger. Especially if it's not one person but a large group of experienced professionals
Except if the large group is mostly a handful of experts with non-scientific private interests (say, to publish papers, to get grants, paid by companies with a product to sell), plus a large mass of professionals who seldom/never question the information they receive and just follow everything they read from their research and protocol-setting breathen.
When non-reproducible papers with BS mistakes get top citation counts and are held as gospel without any researcher/peer reviewer bothering verify them for decades, does anyone really think the average "professional doctor" really questions what they're told to follow?
> But it's actually important to "do your own research", and not just listen to authority. Provided you actually know how, i.e. you can find accurate sources and distinguish fact from fiction.
And to know if you’re only seeing narrative-approved sources, when contra-narrative sources have been suppressed.
Lots of folks got fooled by Iraq WMD because you couldn’t say anything to the contrary without being fired. Other examples, some more recent, are available.
> IMO people actually should "play Internet doctor" and research their symptoms
You should probably not play internet doctor if you’re a hypochondriac or prone to anxiety. The least probable cause of almost any symptoms is cancer or another deadly condition and you can probably find a gruesome description of someone’s experience online somewhere. If your nervous system can’t grasp that, stay away from googling your symptoms!
> IMO people actually should "play Internet doctor" and research their symptoms, not to reach any 100% conclusions, but to suggest diagnoses and tests to their real doctor.
I do this regularly, and having a hard sciences background it's usually an educated and well-thought research. Unfortunately some doctor can get quite upset if you dare to suggest a diagnosis, especially if it's an evidently correct one they didn't think about. You really have to be very careful, allude to things, stress key symptoms and try to make them reach the same conclusion without explicitly telling them how to do their job.
> some doctor can get quite upset if you dare to suggest a diagnosis
A well known idiom is that when the client says he has a problem, he's right. When the client tells you what the cause is, he's wrong.
Don't tell the doctor your diagnosis. You didn't go through a decade of medical training. Tell the doctor what you've observed, let him conclude the diagnosis.
> When the client tells you what the cause is, he's wrong.
using a software analogy, I've seen plenty of bug reports where the reporter did some research on his own that turned out to be determinant in the issue resolution. You have training, inside knowledge of the system that's failing, but you see plenty of issues everyday, cannot dedicate to each one the time they would deserve. The reporter instead is motivated to reach a solution and has plenty of time to research it properly. Discarding a potentially fundamental input just because it comes from someone supposedly untrained would be a shame.
Indeed. My hypothyroidism was ultimately diagnosed by my mother's work colleague who recognized the (frankly quite textbook) symptoms because her dog had it.
Doctors are like everyone else: 90% of them are kinda shit at their job.
> This is one of the most insightful comments I've read on the internet ever.
On the other hand, if I’m in the market for a guess, I’ll probably go for the educated guess rather than the uneducated one, because even though I recognize it’s not necessarily more correct, it probably is less wrong.
Sometimes it's better to know you don't know than to think you know. If it's midnight and you're standing at the edge of a cliff you're better off thinking that you have no idea what's ahead of you than if you listen to a guy with "better than average night vision" telling you that there's a sidewalk ahead.
If you are, as you say, "in the market for a guess" that would imply that you are actively looking at different guesses and comparing the credentials of the guessers. Then you would _probably_ be right to choose eg. the guess from the virologist over the proctologist or the mechanic when you need a guess on how the Covid situation will unfold. But the issue might be that an "educated guess" in itself could be worse because you wouldn't base any major decisions on a mechanic that present an "uneducated guess" that "the virus will die down in a month" but if you have a crackpot chiropractor presenting themselves as a "certified doctor with an expertise in everything concerning the body" that says the same thing, you might be worse off in saying that this "educated guess" is worth basing your travel plans on.
Medical advice on the internet can be n=1, but read enough to get to n=100+, while still limited, can be quite valuable. I can't ask 100 doctors what’s wrong with me, but I can ask my question to hundreds in an online forum. They might lead me astray, but so can a single doctor’s misdiagnosis. A doctor can prescribe treatment that helps or harms, but I can ask a forum of people, with the same problem I have, if I should take his advice. I can't, however, ask 100 of his colleagues or peers if I should trust his advice, and that's why we end up with so many charlatans in medicine. But he's a doctor. Of course you can trust him! I don't know about the rest of you, but I've found a fair number of doctors to be mostly useless.
>This is why self-driving cars still elude us, but will be so much safer than human drivers when they finally do arrive.
Isn't that just a by definition thing?
My Tesla can already drive itself, I don't trust it and I don't think it's safe. When I get one that I trust and is safe then it will be safer than me, and by extension other humans.
The problem is that the current generation of algorithms are brittle.
The way I state the above point is: "The smartest people in the room are the most skilled at convincing people (including themselves) of a narrative that sounds logical but is wrong".
I am afraid many times when talking about educated guesses we mean schooled
guesses. The problem is that, more than once, we learn per Mark Twain's
famous quote ...that schooling has a way of interfering with education...:-)
Specially for subtle but critical concepts:
"Researchers misunderstand confidence intervals and standard error bars"
This is why we shouldn't take medical advice from people who have read a few Wikipedia articles. This is why years of experience in e.g. software are necessary to even begin to understand how deeply hidden security issues can be. This is why self-driving cars still elude us, but will be so much safer than human drivers when they finally do arrive.