This was longer than I'd hoped but I guess that's what happens when you don't write in advance. Feeling a bit stressed out that I went over time. I need to relax. Does anyone know any calming mantras that I could chant? Edit: The chant request was just a silly joke about the cult bit in the middle but to the people answering seriously, just so you know I appreciate you. Thanks! 🙏🏽🙏🏽
Here's a fun selection bias story from my dad, retired General Practicioner. He was among the first generations of doctors in the Netherlands where becoming a GP was treated as a separate specialization, instead of just being a doctor without specialization. With it came new clinical research in the GP's office, and as a result a number of illnesses were reclassified from very dangerous to mildly annoying with the occasional complication. Why? Because before GPs were included in the medical research community, the only people *writing* about these illnesses were the specialists at the hospital. Bu the problem was that these experts only got to see the patients who were so ill that the GP could not treat them any more, leading to a selection bias that introduced the false belief that these illnesses were all very dangerous.
I'm afraid naso-centric stabilization is now your gimmick, and you are obligated to enable it for all future videos. Don't blame me, I don't make the rules...
Please don't. I had to stop the video after you mentioned the naso-centric image stabilization because I couldn't focus on anything else anymore, couldn't stop giggling, and didn't register a word you said.
@@preussischblau The view of his nostrils didn't bother me, it was the fact that I was getting queazy. Your nose was steady, but the room behind you seemed to undulate.
When I first found him from that Tom Scott cross-over, I was a bit unsure of his style, but now I absolutely love this channel and can't wait for each new video!
Just FYI mate, although I barely ever comment on RU-vid at all, I don't watch your videos and think "eh, that was okay". Everything I can think of that I've ever seen from you on RU-vid has been amazing, (and often very funny!) I feel so grateful to you for sharing your knowledge and perspective. I just don't comment because I don't have anything to add of significant enough value to be worth taking time gambling and throwing my thoughts into the bizarre, terrible and often toxic youtube comment section. Oh, also, I'm Scottish and I loved your deep-fried mars bar / phone book joke/example. Although (surprisingly enough) I've never had one myself, I certainly don't think the light-hearted stereotypes about us are offensive, as long as they don't come from genuine ill-will. Just another piece of rather neutral feedback that you probably don't get very often because of the type of selection bias that you were talking about.
You deserve to know that your use of spacing around the slashes, at the beginning of that second paragraph, is seen, appreciated, and quite effective. 👍🍍
@Heloise O'Byrne pretty much that. They also take observational studies, which can't determine cause. But they'll give relative risks of say 1.35 and say "35% increased risk" not state relative risk, nor that it's all just noise and correlations at that level. Once relative risk exceeds 2x(so 200%} risk hazard ratio, it warrants further research, as there's probably something there, but they often overstate the significance or imply causal relations when none exist. Another is when they pick a metric for whatever reason based on the above methids.. make a drug to "fix it" and the drug does what it should... But in the end it doesn't improve outcomes, or does slightly, butnot because of it's effect on the metric.. And is potentially harmful to many, as reducing this metric is bad... *Looks with disdain at statins and the cholesterol hypothesis* Related ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-o4ePp0DmYnY.html
@@StoutProper The Behemoth is a ship on the show 'The Expanse' (On Amazon Prime) which has 3 groups of people: Earth, Mars and belters. The belters call each other Beltalowda
Every single time I watch a Medlife Crisis video I have a few good laughs and learn something new. You are quickly becoming one of my favorite channels.
Thank you. I searched through the comments for this specific comment without any evidence of its existence, and found it just when I was about to give up 😁
On the other hand, one can not assume that the opposite is true, or that there is no causal link. I would say "Correlation is not necessarily causation."
Eugenics is still alive and well as syphillis and most docs are either now not taught much about it or in their eyes, there's a test for that. The problem is with the cheap unreliable testing. It has the ability to mutate or damage the genes so PCR testing is the new standard. If its spread primarily by mucous or hereditarily, it makes sense to test it that way. It disseminates quickly to the organs and does not linger in blood, especially infants. Too many antibodies from mom and baby create a prozone phenomenon and creates false negatives. It is common for almost everyone to recieve a false negative due to incubation timing. It has now caused "autism" and "leukemia" and any congenital organ defects and ranks second highest cause besides malaria for stillbirths, miscarriages, infertility. Peter Buxton said untreated syphilis will surely cause blindness, heart defects, mental illness and death and let anyone and everyone he could know about it. Why are more 30 year olds having heart attacks and strokes compared to 20 years ago? Dating sites, more partners, access exposure infedility...logic. It should be a doctors first question. Could this be an STd? Treat preventatively. Confounding
Another issue with publication bias is that replication studies often don't get published, and as a result, researchers are doing replication studies less and less often.
blame the software all you want: we know the casual parking in the no-go MRI zone in 11:37 has endowed your nose w/magnetic focus powers, time-warping comments and distorting all BG paintings into unspeakable psychodelic horrors. THE CORRELATION SPEAKS FOR ITSELF AND THE SCREAMS ARE DOPPLER-SHIFTED.
How many paused to read the fake article at the beginning? 🤣🤣 I washed my eyes with turmeric and now I can see through time! This guy never misses opportunity to crack a joke.
THANK YOU! I teach graduate level statistics and I am sharing a link to this video with all my students. Excellent job describing the various types of bias!
.. though it's often a good tell of the both factors having a common underlying cause. Also, papers rarely go into causation. Causation is something researchers talk about during their lunch breaks, and almost never something they put into their papers
Eugenics is still alive and well as syphillis and most docs are either now not taught much about it or in their eyes, there's a test for that. The problem is with the cheap unreliable testing. It has the ability to mutate or damage the genes so PCR testing is the new standard. If its spread primarily by mucous or hereditarily, it makes sense to test it that way. It disseminates quickly to the organs and does not linger in blood, especially infants. Too many antibodies from mom and baby create a prozone phenomenon and creates false negatives. It is common for almost everyone to recieve a false negative due to incubation timing. It has now caused "autism" and "leukemia" and any congenital organ defects and ranks second highest cause besides malaria for stillbirths, miscarriages, infertility. Peter Buxton said untreated syphilis will surely cause blindness, heart defects, mental illness and death and let anyone and everyone he could know about it. Why are more 30 year olds having heart attacks and strokes compared to 20 years ago? Dating sites, more partners, access exposure infedility...logic. It should be a doctors first question. Could this be an STd? Treat preventatively. Confounding
People in my lab are also very concerned remarks about how quiet things are, so I also tell them as much as possible, reminding them that our lab is a temple of science and we do not tolerate beliefs in jinxes. I also like to remind people when I drive with them that I have never been in a car accident. Freaks them out.
I'm loving all of the uploads recently. I've convinced myself that you will never upload becuase of your busy job but you keep putting out more and more content. You must have no spare time!
An interesting correlation causation problem is the link between police officers and crime. An increase in police numbers correlates to increased crime because as crime goes up you need more police to deal with it, as a result its really difficult to judge how much increaseing police budget would reduce crime. Also meh, it's ok
Also more policing resources often = more detection of crimes and more prosecutions (ie investigation leads not being pursued otherwise due to low resources, more investigation of 'routine' or low level offences like burglaries, etc that would otherwise be lower priority so go unsolved)
@@NFMorley aye and if you're cutting a police budget as an austerity measure then the likelyhood is that you're cutting social services and/or the countries in a recession and as poverty is a big cause of crime its difficult to isolate policing as a factor. I think usually they use reporting as an indicator rather than convictions which has its problems like people are more comfortable today reporting racisism and sexism but its unlikely that society is worse now than in the 70s
Watching these videos increases my conversational pub skills by 10%. Video suggestion: please discuss the average amount of years there are breakthroughs in medicine regardless of their impact.
Another problem with your phone book example: even if you had selected people with last names A through Z, instead of just Mac... You'd still end up with a sample of people that are relatively old, since anyone under, say, 40 is much less likely to even have a home phone, let alone have it registered in the phone book.
This is one of the best overviews of clinical study mistakes I've ever seen! And it was by far the most entertaining! Good show! A note on confounding (6:15) - I'm in nuclear engineering, and I notice that one of the hallmark reports that dictates how the US government handles radiation risk (the BEIR report) follows those who have been exposed (bomb victims, those living near nuclear accidents, etc). They have a really hard time removing confounding variables (the fact that Russians generally use far less salt in their food and therefore more readily absorbed radioiodine, or smoking/drinking/other bad health habits, for example). I would love for you to do a video about the hazards of nuclear and the medical body of knowledge behind procedures that employ radiation sources (both isotopes and electronically generated).
Another excellent video! As someone who used to work in the pharmaceutical industry as a rep, I think this video should be compulsory viewing for all reps so they can have healthy cynical attitude to the view their companies try to brainwash into them.
@Medlife_Crisis I was all set to comment with "video was okay" in reference to the selection bias portion, but I can't because I got to the part about memory, realized there were _not_ five things mentioned, tried very hard to remember the first four things, then had to go back and rewatch the preceding section to figure out that the reason I couldn't recall the fourth one (btw, examples were very helpful in cementing recall, so 👍 for memorable examples) was that the section on recall bias _was_ the fourth one. All of which is a very long-winded way of saying 👏 well-played, sir, and a sneaky way of getting people to review and retain info - wish some of my college profs had done this 😝
"The inconvenient truth about Gwyneth's PHD in immunology" So I've disabled my adblocker, turned off my CSS mod's, enabled annotations and given the page a hard refresh/'cache clear' and I still can't click on this 'link' @ 19:10 Google is blissfully unaware of any such video and RU-vid itself won't admit it's existence either - _Are you pulling my leg doctor?_
Although correlation is not *necessarily* causation, causation is *always* correlation! The issue is a conditional statement (if causation, then correlation) and its converse (if correlation, then causation) are not logically equivalent. This means that although the first statement is true, the second is not necessarily true. I had a roommate once who fought me about this point because he didn’t understand this!
I swear you'll be shooting a video and at some point there will be a cut to you in a tomb with "Well I kinda died between that last point and this one, but as I was saying-" You should take a good break between doing these videos and work, they're still so informative and interesting.
regarding your selection bias example: you did fall for it already when you decided to use a phone book: the sample population reachable through the phone book differs already greatly from the target population (assuming mobile numbers are not added to them without active involvement of the respective person). ;)
I lost my mum and my sister to cancer, so thanks for this video. I feel for anyone suffering a medical condition, I don't blame the desperate to turn to anything that might help. But at the same time please be careful.. there are those that just see you as a paycheck. Be honest with yourself, and always consult your doctor & family.
Oh regression to the mean is a good one that I'd never really thought about. I can see how that could be the cause of a TON of superstitions. It reminds me slightly of the anthropic principle from cosmology, which is basically a counter to how coincidental it is that the universe is fine-tuned for life -- any laws of physics that we observe must necessarily be perfect for life, otherwise we wouldn't be there to observe them. Similarly, any observations of "Wow, it's quiet" are necessarily observed when the day has the most potential to become busy again.
I got myself a masters degree in statistics, and I stopped believing in anything I didn't perform myself (if that even). I work as a statistical consultant and I've had to just abandon telling people how I really feel because people just really want to feel certainty in what they are doing... even when I don't.
Man, as a scientist I have to just say: Keep this GREAT work! Science is polluted, we all know that, but this doesn't mean that we have to loose trust in it, it means that we need more scientists and people that know how to approach scientific papers doing science communication before everything else. If the public loose trust in us the world will fall into a sea of antivaxers wondering what's underneath our flat Earth while eating chocolate in order to loose weight and cure cancer.
I feel as if I just sat through an Intro to Statistics class. I would recognize this because my son is Statistics professor at a local University. Just recently found your channel and love it. I live in the US where even with health insurance we pay out of pocket approximately $300 to spend 5 minutes with our GP who spends more time looking at a computer screen than looking at the patient. They then insist on running every test they can think of so as to maximize their profits across the entire medical community. God forbid that we actually get sick!!!!!
Sir! With all due respect, I report, that to my ears and ability to digest your brilliant video, the background music was something I could live without.
@@MedlifeCrisis If you do decide to keep background music, I suggest dropping the volume on the music between 3dB - 6dB. Another thing you can try is using an EQ to knock down 2-5kHz on the music while there is speech; this should help with intelligibility.
This should be written by every student in every field, each one with their own blood, so they will remember! Btw. I hope you will answer my question I have written in your previous video.
when i clicked on this video i wasnt expecting to laugh so hard at the nose-centric stabilization and the comments people leave about it that i couldnt even get up from the floor for five entire minutes thank you, all your videos have been amazing so far!
I just want to say that the four people who complained of smelling like cabbage were definitely not compliant with the protocol. Who puts cabbage in minestrone?!?
Your regression to the mean example made me laugh so hard! My girlfriend is also a doctor working shifts at the hospital and she forbids me from asking here over text messaging if the shift is quiet or no 😂
Haha joke's on you I remembered there hadn't beed 5 types of bias and could name 3 of them only because naso-centric gyro stabilisation made me laugh so hard that I forgot the whole point there. 10/10 would like more stabilisation in the future.
Thank you once again for putting out such informative and humorous content. I have just finished watching this one for the third time. Just to let you know, though I imagine that you already know this, but these reasons, or biases are useful to be aware of when developing algorithms. I looked forward to your videos. While, I will never be a Doctor, it will make me a better patient (hopefully, when the time come, let’s face it, I could just drop dead in my tracks), and help me to make better choices. Speaking of which, where is my cup of Schrodinger’s coffee. Cheers! 😎
I wanted to comment about bias but realized that it's not really in research so I will comment under the next video if you'll forget about that one. No pressure. But the real reason to comment is - after one of the recent videos I decided to give Brilliant a try, after checking it months ago and being afraid that it's going to be either too trivial or too difficult. Ended up getting the annual subscription and really enjoy it every day. So even though there are a lot of those ads around RU-vid - seriously - give it a try, you might be positively surprised. Or am I biased because I want to convince myself that I didn't waste my money on it... Goddammit!
this is so good! :) thank you! This is applicable not only to medical research, but also to other fields of science - so me watching this does not even count as procrastination :D
I interviewed for a mental health study once, and the interviewer asked me questions pertaining to bipolar, which I have been diagnosed by several doctors and treated in such manner. The questions I hadn't seen before, and they were all open - ended questions. I can understand this was an attempt to eliminate bias. She asked my symptoms and I gave them a best I could. It was hard because I have had many over 2 decades past, and I wasn't sure whether it was for any symptoms I ever had or only currently with treatment. So subject - bias was given, though not intentionally. The next thing that happened was she said, "That sounds more like anxiety than bipolar!" and her tone wasn't ingenuous. Trying to control my anger at her method, I asked if she were a doctor, psychiatrist, psychologist, a licensed clinical social worker, and mate a psychiatric nurse practitioner (if I didn't the last, I should have). She said no. So I lambasted her for trying to diagnose me over the phone in 5 minutes. I said she should take the answers down and let the doctor decide whether I'm eligible or not. I chewed her out good. Then I called up the Psychologist (PhD) and chewed him out for her conduct. He was solicitous, but I did not let up. I told him what she should have done and then withdrew my candidacy. The interviewer would have introduced another bias had I continued. I was steamed, an because of that I will never volunteer for any more studies. I was so incensed! And the Psychologist should have been, too. This is the example that sticks with me whenever I hear research bias. It still sticks in my craw.
Not arguing the veracity of your claim, but is it possible they were studying the effect of resistance to treatment on those with mental illness? "Are you sure you have X instead of Y?" Is a common response to sufferers, and its a problem that many people acquiesce to the premise.
Nice endcard. That music is from "Narcissistic Bitch", the Dr Phil meme. My God, are you very cultured. It had to be played during the sponsorship part...Sneaky Bastard.
One can observe: - 930 liked this video vs 2 who did not. So far.... A bias of people who like a certain type of British accent? Or perhaps it's the background music? Or perhaps the mention of Minestrone. I like Minestrone - usually.
I hate the way my research so far has been reported by the media. E.g. 'Zinc may be a treatment for urinary tract infection'. No, no that's not what we showed At all. We only looked at cells in a dish, a million miles away from patients and treatments. Also, if anything we showed that zinc wouldn't likely work because the bacteria we studied evaded the cell response which harnessed zinc. Just a huge misinterpretation. Bah.
Thanks for commenting. Hey at least the media is interested in your work! That's a positive. But yeah very annoying how things are twisted. Did you or your establishment issue a press release? I've seen the press office at my university give quite the wrong information, so sometimes journalists are not to blame
When you tell someone something they don't want to hear, like it's quiet, is disturbing their peaceful mind and stressing them. If you don't like the heat get our of the kitchen, but also don't turn up the temperature on purpose. ;) I also have a particular disdainful attitude towards superstitions.
I enjoyed your explanation of what a phone book was. It's a simple but non-trivial and not immediately obvious piece of information. It reminds me of a RU-vid video I saw where two teenage boys tried -- and failed -- to figure out how to dial a number on a rotary phone. It was a lot less clear to them than I would have expected.
I think what I like about this channel is the fact that you're an actual doctor working at f** off in the morning and talking about real science and problems that you find in the real world.
I was going to send you money, but you would just spend it on cheese. Statistics show that cheese-eaters are more likely to create cheesy youtube videos than non-cheese-eaters. Cheese is bad m'kay?
Hey Medlife Crisis guy! This was informative and downright enjoyable! Had me thinking and laughing (not hysterically but more than a chuckle) throughout. Then (dant dant daaaaa) "oh no he dint" you said the "q" word! I will say that in my lifetime of double, triple shots of espresso, not blind but wearing my shades, studies in the fields of finance and retail groceries if someone says the "q word", "s word" or the "it's not 'b word" as the case may be, it will in fact get absolutely bat guano crazy "b word". That being said...lovely video, you looked like you could just fall asleep if you stopped talking. You are quite the dab hand. My very best to you, Kathi😎 in AZ🇺🇸
will you do a video on the study that was published a few weeks ago on nutritional health? the study was called "Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake" and published in cell metabolism. I found it interesting, I would like to hear your thoughts
The answer to that last question is obvious if you think in logic terms, but assuming that normally spoken english translates 1 to 1 to logic terms is not reasonable. The way most people use "or" is not like a logical "or" operator but more like an "xor" - exclusive or, which is what people who chose the wrong answer would assume was implied. So, the question states ("A") vs( "A" and "B"), but many people would interpret it as ("A" and not "B") vs ("A" and "B"), so the only probability that would matter in that interpretation would be wether B or not B is more likely. Does this show bias or just a mismatch between spoken language and written logic?
10:18 Trick question, this is only the fourth bias. So far we had confirmation and selection bias, and confounding in that order. It is 4.50 a.m. where I live, I am 22 years old and have a messed up sleep cycle.
how dare you to say that your beard without shaving every day is scruffy? It is looking good, my favourite doctor. Good vid, it's worth to telling people about that, as well as generally about VERYFIYNG sources instead of taking everything as granted.
my personal favorite bias, survivor bias, hasn't shown up. It describes basically that you only look at the survivors and hence miss actual weak-points. The famous examples given were WW2 bombers: the returning aircraft shown alot of hits to wings and none to the tail or engines. A easy mistake would have been now to strengthen the wings, because they get hit all the time. But those hits actually didn't matter, because the planes survived them. What matters where the tail and the engines because getting hit, meant that the planes crashed and didn't come back to review the damage.
Wow. With this new Canon feature, you no longer need to strap yourself to a frame with a 1950's 16mm film camera running at 8 frames per second to look like a 1990's Toronto alt-rock music video. You too can believe your nostrils are the centre of the universe.