To me the primary concern is that the ratio of vaccinated:unvaccinated cases is very separate from the probability of catching the disease when exposed. With a highly vaccinated population, it is quite likely that 50% of cases will be among the vaccinated.
Even higher, i would suspect... hypothetically 100% in a completely-vaccinated population. And i agree, the probability of infection given exposure is a very important factor to consider when describing vaccine effectiveness.
Wage gap doesn't exist when you factor marriage, prior marriage, having children, and the overall fact that men have a higher likelihood of working longer hours
In your plot on the left at 40sec, the red and blue for over 60 should be swapped otherwise the numbers don’t work. You actually do this when you stack them at 45 sec.
@@MrJosiahCochran You're welcome. It's a very easy thing to miss, and if this were a 5 minute video, I would explain it in much more detail. Thank you for checking it out!
@@sarqf212 Think of it like this... even though more old vaccinated people died, it represented a smaller proportion (of old vaccinated people) than their unvaccinated counterparts. But when you plug that number (vaccinated deaths) into the whole population, it gives the impression that the rate is relatively higher than it is. Does this make sense?
@Michael Tan yes, that was my point: the compound statistics for deaths is misleading and makes people believe the disease is more deadly than what it really is
@@sarqf212 Oh, I think i've misunderstood your question. Are you saying that the compound statistics for all deaths is misleading, because it's mostly old people who die? This is only misleading in the context of claiming that a 5% chance of death applies equally to people of all ages, which I think is easy to explain as not the case. That is, when I say a disease has an average 5% mortality, most people will not simply assume that a healthy 20 year old is at equal risk as a 70 year old with health issues. This video is specifically looking at whether the effectiveness of VACCINATION can be born out by looking at the compound data.
Very interesting. People these days are not yet educated enough in a general sense to be able to understand statistics properly. That's why it's so easy to mislead the public with sensationalist but false headlines. And as long as this is allowed to remain profitable, people are gonna get hurt because of it.
@Steve Gracy Thanks for the contribution... I've written in the description that these data are hypothetical (while remaining plausible and mathematically sound) to illustrate the paradox :)
The timing of the video is the factor that made me write this. If it hadn't been done during this pandemic then I'd agree with it, because it clearly summarizes your point. The over simplification occurs due to the fact that the word "vaccine" in the video will make the viewer think all vaccines are the same. This just isn't true, live and killed virus vaccines are far more effective in the real world than say vector vaccines. Also the maker of the same type of vaccine can give very different results, eg: Russia vs China. There are different strategies for different vaccines. Some vaccines try to stop the spread (HPV vaccine), others prevent severe illness and death (small pox vaccine) The general health of the population is a big factor. Currently the West has 60% of its population with unhealthy BMI(60%). The healthy(40%) are a minority. Health plays a major factor in how the immune system responds to a virus. And today, politics, the media and money determine what is and isn't a good vaccine and what is and isn't a good strategy. Eg: New Zealand vs the UK. I guess if you hadn't used vaccines as an example it wouldn't have bugged me. Oh well, my fault.
He said the people in the example "had been vaccinated against the disease". At no point did he mention which disease and if there is more than one vaccine for it. The fact that you latched onto this show pretty clearly what you are like.
@@Jacquobite hahaha, the irony of your statement also tells us what type of person you are. My point is still very valid. This is because viruses affect different people differently. Using vaccinations as an example for Simpson's Paradox is an over simplification of vaccines and their effect on people. Eg: Covid, the WHO states that 90 to 99% of people are naturally immune to the virus. Therefore, regardless of whether a vaccine is used or not, 90 to 99% of people won't suffer from it because they are naturally immune, and the death rate is even smaller. You need to prove that it is the vaccine that is preventing the death and suffering and not the natural immune system that is preventing the death and suffering first. This FACT is not being taken into account in the video. Does the example explain Simpson's Paradox? YES, is it a realistic example of vaccination? NO.
See problem here is you're explaining very basic systems, only the people who don't understand them wont want to understand them. You can lead a horse to water but you can't force it to drink.
@@JonathanXLindqviust ah i understand. this is true! having said that, this concept is very counterintuitive and took me some time to fully understand too. but, i follow the data and stats, not an ideology :)
when you asked the question ,how in 1 million deaths half are vaccinated,,,,,maybe becz fairly large part of population has got vaccinated,isnt it so ?
@@MichaelTanMusic From the observation i would tell that people dont get that if they are SICK vaccination should wait :v (old people dont care, and rest are not clever enough to care about it or dont belive in this anyway :v)
I don't get it. Let's say for simplicity 100 people died, 55 were vaccinated and 45 were unvaccinated. We need to split these 100 people in 3 groups, so that in each and every group the unvaccinated are more than the vaccinated, and I don't think this is possible! Can anyone find such a combo? Either I or the video missed a critical piece of information!
Thank you for the comment. If I had 5 minutes, I would elaborate more :) The numbers in these graphs do all add up, and your example is completely plausible. It's very counter-intuitive, and even the simplest example ends up with a lot of numbers! I think the thing you are missing is that, when the unvaccinated deaths are shown to be higher than the vaccinated deaths (in each of the age groups at 0:33), we are looking at a PERCENTAGE within each group. It doesn't necessarily mean that *more* unvaccinated people died; in fact, more OLD VACCINATED people died than any other group. OK, here's a scenario similar to yours! (It's tricky to reverse-engineer your numbers, but these are similar). Let's start with the 60+ group; and say that 58 deaths were in vaccinated people, 34 were in unvaccinated. And let's say this corresponds to 12.5% of those vaccinated, and 25% of those unvaccinated; creating a vaccinated pool of 464 people, and an unvaccinated pool of 136. And puts 600 people in this age group. In the middle group, say it's 8 and 9 deaths (vax, unvax) corresponding to 2% and 4%. 625 people in the age group. And the youngest group, say it's 1 and 4 deaths corresponding to 0.5% and 1% of each group. 600 in this group. So, this adds up to 67 vaccinated deaths and 47 unvaccinated deaths. And within each age group, the vaccinated people die less frequently. And, add it all up, and the vaccinated deaths are 67/1064 = 6.3%, unvaccinated 47/761 = 6.2%. Spooky huh!!
@@MichaelTanMusic Ah thanks for the response. This makes sense now, I though the percentages were on the total population, not separately within each group. I believe this is exactly the same thing that happens with gerrymandering and votes right? But from the opposite perspective.
The RUclips copyright protection is making it very difficult to post these videos any more, it was a lot of trial-and-error to know which songs would result in the video being blocked, or forcing me to make sections of it silent. heartbreaking for me :(
Well done - very nicely illustrated.
This is so incredibly important. Thank you.
To me the primary concern is that the ratio of vaccinated:unvaccinated cases is very separate from the probability of catching the disease when exposed. With a highly vaccinated population, it is quite likely that 50% of cases will be among the vaccinated.
Even higher, i would suspect... hypothetically 100% in a completely-vaccinated population.
And i agree, the probability of infection given exposure is a very important factor to consider when describing vaccine effectiveness.
Fascinating and well put together
Interesting analysis!
science for the win
This is simply magnificent. Subscribed!
Very Nicely done. Ran across this from Veritasium.
awesome, thanks for checking it out :)
I’ve also seen a correlation of morbidly obese people dying more often as well.
maybe there’s a Simpson’s Paradox within a Simpson’s Paradox !
@@MichaelTanMusic We'll just call that one Homer Simpson's paradox.
@@WhoopityDoo brilliant
Saving to favorites. I know I'm going to need to need to use this in a debate at some point
This is such a cool fun fact thank you
Love it!
Was it an epidemic of greyscale?
Same for the wage gap?
Interesting question, but definitely a political mine field!
Wage gap doesn't exist when you factor marriage, prior marriage, having children, and the overall fact that men have a higher likelihood of working longer hours
nice. where could i find a spanish version? autotranslate generates subtitles but they cover the graphs
In your plot on the left at 40sec, the red and blue for over 60 should be swapped otherwise the numbers don’t work. You actually do this when you stack them at 45 sec.
No, the data are correct. One shows PERCENTAGE of the cohort, the other shows NUMBER OF DEATHS. This is the engine of Simpson's Paradox!
@@MichaelTanMusic I see thanks for the clarification!
@@MrJosiahCochran You're welcome. It's a very easy thing to miss, and if this were a 5 minute video, I would explain it in much more detail. Thank you for checking it out!
Do more statistical errors!
Wow.. 😮
Doesn't the same apply for the statistics related to the mortality of the disease?
not sure what you’re asking! the example i’m using is looking at the vaccine’s effect on mortality :)
@@MichaelTanMusic I'm talking about the effects of different age groups in the overall mortality rate of the disease
@@sarqf212 Think of it like this... even though more old vaccinated people died, it represented a smaller proportion (of old vaccinated people) than their unvaccinated counterparts.
But when you plug that number (vaccinated deaths) into the whole population, it gives the impression that the rate is relatively higher than it is.
Does this make sense?
@Michael Tan yes, that was my point: the compound statistics for deaths is misleading and makes people believe the disease is more deadly than what it really is
@@sarqf212 Oh, I think i've misunderstood your question. Are you saying that the compound statistics for all deaths is misleading, because it's mostly old people who die? This is only misleading in the context of claiming that a 5% chance of death applies equally to people of all ages, which I think is easy to explain as not the case. That is, when I say a disease has an average 5% mortality, most people will not simply assume that a healthy 20 year old is at equal risk as a 70 year old with health issues.
This video is specifically looking at whether the effectiveness of VACCINATION can be born out by looking at the compound data.
Very interesting. People these days are not yet educated enough in a general sense to be able to understand statistics properly. That's why it's so easy to mislead the public with sensationalist but false headlines. And as long as this is allowed to remain profitable, people are gonna get hurt because of it.
Great video! If I didn't know any better and saw that data I would agree and say "The vaccine does nothing"
And you wouldn't be alone! It's an easy mistake to make, and everyone from News reporters to Senators have been making it.
@Steve Gracy Thanks for the contribution... I've written in the description that these data are hypothetical (while remaining plausible and mathematically sound) to illustrate the paradox :)
@Steve Gracy which graphs?
Well it definitely doesn't "vaccinate" people.
The timing of the video is the factor that made me write this.
If it hadn't been done during this pandemic then I'd agree with it, because it clearly summarizes your point.
The over simplification occurs due to the fact that the word "vaccine" in the video will make the viewer think all vaccines are the same.
This just isn't true, live and killed virus vaccines are far more effective in the real world than say vector vaccines. Also the maker of the same type of vaccine can give very different results, eg: Russia vs China.
There are different strategies for different vaccines. Some vaccines try to stop the spread (HPV vaccine), others prevent severe illness and death (small pox vaccine)
The general health of the population is a big factor. Currently the West has 60% of its population with unhealthy BMI(60%). The healthy(40%) are a minority. Health plays a major factor in how the immune system responds to a virus.
And today, politics, the media and money determine what is and isn't a good vaccine and what is and isn't a good strategy. Eg: New Zealand vs the UK.
I guess if you hadn't used vaccines as an example it wouldn't have bugged me.
Oh well, my fault.
He said the people in the example "had been vaccinated against the disease". At no point did he mention which disease and if there is more than one vaccine for it. The fact that you latched onto this show pretty clearly what you are like.
@@Jacquobite hahaha, the irony of your statement also tells us what type of person you are.
My point is still very valid. This is because viruses affect different people differently. Using vaccinations as an example for Simpson's Paradox is an over simplification of vaccines and their effect on people. Eg: Covid, the WHO states that 90 to 99% of people are naturally immune to the virus. Therefore, regardless of whether a vaccine is used or not, 90 to 99% of people won't suffer from it because they are naturally immune, and the death rate is even smaller. You need to prove that it is the vaccine that is preventing the death and suffering and not the natural immune system that is preventing the death and suffering first. This FACT is not being taken into account in the video. Does the example explain Simpson's Paradox? YES, is it a realistic example of vaccination? NO.
See problem here is you're explaining very basic systems, only the people who don't understand them wont want to understand them. You can lead a horse to water but you can't force it to drink.
the contest allowed me 60 seconds. given 5 minutes, i would have gladly elaborated :)
@@MichaelTanMusic I think you misunderstood me, I'm saying that asshole ignorant people won't want to be educated.
@@JonathanXLindqviust ah i understand. this is true! having said that, this concept is very counterintuitive and took me some time to fully understand too.
but, i follow the data and stats, not an ideology :)
I feel like I’m being lied to by the government for the first time
You erned a suscriber
cool! thanks for checking it out :)
when you asked the question ,how in 1 million deaths half are vaccinated,,,,,maybe becz fairly large part of population has got vaccinated,isnt it so ?
that might be the case, but even if it’s the same.. there’s another explanation: the vaccinated people were far older than the unvaccinated
@@MichaelTanMusic From the observation i would tell that people dont get that if they are SICK vaccination should wait :v (old people dont care, and rest are not clever enough to care about it or dont belive in this anyway :v)
When hit songs and samples 18
I don't get it. Let's say for simplicity 100 people died, 55 were vaccinated and 45 were unvaccinated. We need to split these 100 people in 3 groups, so that in each and every group the unvaccinated are more than the vaccinated, and I don't think this is possible! Can anyone find such a combo? Either I or the video missed a critical piece of information!
Thank you for the comment. If I had 5 minutes, I would elaborate more :) The numbers in these graphs do all add up, and your example is completely plausible. It's very counter-intuitive, and even the simplest example ends up with a lot of numbers!
I think the thing you are missing is that, when the unvaccinated deaths are shown to be higher than the vaccinated deaths (in each of the age groups at 0:33), we are looking at a PERCENTAGE within each group. It doesn't necessarily mean that *more* unvaccinated people died; in fact, more OLD VACCINATED people died than any other group.
OK, here's a scenario similar to yours! (It's tricky to reverse-engineer your numbers, but these are similar).
Let's start with the 60+ group; and say that 58 deaths were in vaccinated people, 34 were in unvaccinated.
And let's say this corresponds to 12.5% of those vaccinated, and 25% of those unvaccinated; creating a vaccinated pool of 464 people, and an unvaccinated pool of 136. And puts 600 people in this age group.
In the middle group, say it's 8 and 9 deaths (vax, unvax) corresponding to 2% and 4%. 625 people in the age group.
And the youngest group, say it's 1 and 4 deaths corresponding to 0.5% and 1% of each group. 600 in this group.
So, this adds up to 67 vaccinated deaths and 47 unvaccinated deaths. And within each age group, the vaccinated people die less frequently.
And, add it all up, and the vaccinated deaths are 67/1064 = 6.3%, unvaccinated 47/761 = 6.2%.
Spooky huh!!
@@MichaelTanMusic Ah thanks for the response. This makes sense now, I though the percentages were on the total population, not separately within each group. I believe this is exactly the same thing that happens with gerrymandering and votes right? But from the opposite perspective.
@@AndreasLianos haha yes, if the electorates are deliberately sized differently, a similar effect could be gerrymandered!
In other words; Why we need intersectionality.
what.
hey?
What happened to the "hit songs were sampled"?
No need to see some vaccination stats.
The RUclips copyright protection is making it very difficult to post these videos any more, it was a lot of trial-and-error to know which songs would result in the video being blocked, or forcing me to make sections of it silent. heartbreaking for me :(
@@MichaelTanMusic for the well known songs just mute it
@@BrazilianDaftPunkFan yea but it defeats the purpose when half of the songs are muted haha
@@MichaelTanMusic Just make the copyrighted songs segments shorter and RUclips won't recognize it. Give it a try. I always do that.
@@OrionCassiopeiaAndromeda Yes, but it's getting better. Sometimes it recognises just a couple of seconds!
:/