Done with whole statistics playlist and I can say it was a great help. Now I believe I have the working knowledge of the topics covered in this series of lectures. Thank You!
Unexpected sudden ending of 68 video's. Thank you very much for clarifying all this material for me, your explanations are very clear and easy to watch!
I really like these videos from Khan Academy. I encourage my students to use them to do independent study. There is a definite correlation between using the videos and success in the examinations, but not proof of causality yet :)
I come back to this talk every few months to re-inforce the arguments...it is really necessary to apply these lesson to so many things in life (Global warming, most of Modern cutting edge medicine, etc)...Remember they are measuring correlation and selling it as causality, to get FDA approvals or more research money.
After all these years of hating... I finally start liking statistics :) I am really sad because of the fact that this is the final lesson. Thank You Sal!
@ZergAteu The researchers probably understand the difference and may have even been careful with their wording. The article, however, is clearly trying to have the reader walk away with the thought that eating breakfast makes obesity less likely.
Just finished the playlist. It is remarkable how Sal has the talent to build concepts from ground up making the student really got into what is going on. This is a great example of humility, something usually missing on education (specially higher).
I now finished the whole playlist to get some introductory ideas about statistics. I think I got some now. It really provides a good value for learners, great videos, he explains well to make understand intuitively. He's great.
Wow! This video, apart from explaining the difference between correlation and causality, tells us why it is important to know the difference too. Causality is a correlation. However, not all correlations are causalities. What we need is the objectivity to see a causality as a causality and a correlation as a correlation. Above all, we must have the ability to picture the interactions between a multitude of correlations and causalities while evaluating an event.
In all these examples, yes. But you could also have 2 things that are correlated completely coincidentally. See for example the website/book "Spurious Correlations".
Causality always creates correlation, except in a few pathological cases. We generally say that X causes Y when doing X changes the probability of Y occurring. As a result, X and Y will be correlated. The point of this video was to say that correlation does not imply causation. The converse, however, is generally true. Causation implies correlation.
Correlation is saying that two things can go on simultaniously and cause and effect is saying that one thing happens which affects the next thing in the sequence. Directing this to the video, correlation is saying that sleeping late could impact your ability to eat breakfast because maybe they dont have enough time and vice versa and cause and effect is saying that because they are obese, they could not eat breakfast. That is just an example.
Thanks Prof for these mind boggling videos. Studied stats in my MBA from one of the most revered Faculty from an IIM and he was nowhere half as good as you. I had no idea stats could be this much fun. Thanks a lot. A bit sad to see the abrupt ending of 67 videos.
This is to the point. As I watched I kept thinking about the many articles that "might be wrong" that come out on a daily basis. These are powerful concepts to understand and apply regularly.
I Appreciate for all the statistics videos You are the one of the best teachers I ever meet！Your teaching is clear and kinda funny sometimes make me laugh lol anyway thank you again！！
Ideally, the authors would make the notes that Khan has made and provide some alternatives to causation as Khan did. This way, when the reader reads the article they don't come away thinking, "if I want to lose weight I just need to eat breakfast" but instead, "people are researching ways to help me lose weight, but they still don't know much besides "eat healthy and exercise".
I wish more people thought like this, not taking things at face value, making up their own mind. would certainly help with the media and their inflexions towards news and current affairs.
@ZergAteu If what the study implied was a correlation rather than a causation, they shouldn't have said "eating breakfast may beat teen obesity" because it singles out one correlated factor. Why not name it "getting up early may been teen obesity" or "being more active may beat teen obesity"? Don't get me wrong, I agree with you that the article does not imply causality, but that's only because it's filled with weasel words.
Hi Khan. I'm a big fan of your work and like following you. I hope you don't mind if I make a few suggestions. I think you did a great job at differentiating between causality and correlation. A few things I would recommend mentioning in addition to this are: timing (which came first?), co-linearity (overcoming it to analyze causality), and randomization or multivariate models (how it takes care of the confounding issue you mention). Good luck and great job!
Ok i ve seen an article telling that most rock climbing accidents happen by men rather than women. But as far as I know there are more men that goes rock climbing than women. Can we say there is an correlation or how can we implement it in causality, can you explain i really didn’t get it. The more men rock climbing causes more accidents, am I wrong?
A fluctuation of A and B tend to be observed at the same time, Instead of A causing B; That's a very good example to illustrate Correlation and Causation, in my opinion.
Thank you Professor for this amazing videos. I watched the whole playlist of Statistics and I find it inspiring. Longing to follow with another playlist soon.
Thank u sir, you have expanded my way of thinking or looking at things to a different level, your videos are very helpful and I hope you make more of these useful videos. If you need our help through donations so you can make more videos, than add a link and I will gladly support.
Great. I was also hoping you'd throw in an explanation of what a logical corollary is. Although it's implied throughout. Still it's easy to confuse that with a correlation unless you know why they can't be the same thing.
Great videos. Loved all of it. I feel the regression part could be expanded to talk of linear and multiple regression analysis, and also explaining how to interpret excel outputs of regression. But overall, really good!
So... since life is filled with variables, how does one conclude actual cause and effect other than in a controlled lab or controlled environment where even those are often misinterpreted based on unforeseen conditions?
@ZergAteu I didn't vote it down :-) You might be underestimating the influence the title has over the way the rest of the article is read. If it starts out with "X may beat Y", then the reader will interpret the content as being in line with that summary. It's a tactic frequently used by tabloid journalists.
As I was listening to the first couple of seconds I thought that maybe american titan General Mills is probably the one that funded that study cited in webmd. Anyway, great video, very khana - educational. People on jury duty should probably see this video before making any decision.
Causality and correlation are synonymous? So the following is true? As the number of people who breath air increases, the number of people that die increases. Therefore, breathing air causes death.
Great piece here. The funny thing is, even as a fairly fit guy, in the military, I started struggling with weight management in my thirties like a lot of people do, and one of the first things that really helped me start turning it around was compressing my eating window. Let's also be sure to acknowledge the fact that people dealing with poverty and food insecurity often eat the worst possible macronutrient composition (high fat + high processed carbs + low protein), which can be obesogenic via hormone signaling even if not massively overconsumed.
In this case, the readers do not likely know the difference between causation and correlation so the weasel words result in the reader believing something is true and backed up by science when really it isn't true (not that it's false, just that it's not proven to be true or false yet).
@@tkhasimbashaa This comment was irrelevant too. This topic ACTUALLY critically analyses the importance of ethical interpretation in statistics. n= 2n. bye.
I think giving a more obvious example (e.g. the correlation between the number of known stars and the population of earth) would've been a better way to start, like c0nc0rdance did in his video on the subject.
@ZergAteu @qtutoringhelps I believe a big part of the problem here lies in the target audience. If a paper was published in a scientific journal that used a weasel word here or there the readers, other scientists, would understand the difference between causation and correlation (hopefully). However, the referenced WebMD article is meant to be an easy to digest summary of the study fit for consumption by the masses (non scientists).
[INTERPRETATION ] what makes you so confident in case you encounter a "realistic" causality? is there any robust academic foundation? I couldnt find any clue in your greatest lecture in this clip. Am I an idiot? I would politely ask you to answer or explain bits on this stupid kindergargetn level question please.
carbs are essential, I hate that there is this stigma behind them. Every cell in your body runs of glucose, aka simple/complex sugars... sorry just had to get that out there. great video!
+indigochick Yes, all three macro nutrients are important, all three of them have been demonized at some point in history when it comes to weight management. It is still hard to believe for people that weight management is just as simple as: total kcal intake and total kcal expenditure. But you know, that is not fancy enough for articles etc.
causality and correlation are synonymous. It doesn't matter if you eat cinnamon buns or bran muffins in the morning you will become overweight at some point in your life if you keep this type of diet. This study is typical research news the public is fed via the media. If the public is generally ignorant(or skeptical) of research methods how are they supposed to make informed decisions about diet.
@alique087 We cannot judge the intentions of the article but we can attempt to draw conclusions about the effects the article has on it's target demographic (non-scientists). A non-scientist reading this article is likely to come away with the impression that eating breakfast will help them lose weight. The article does not technically show causation, but it does imply causation to the average reader who is unfamiliar with causation vs correlation.
Correlation sounds boring. Causality and narrative are what our minds crave. Journalists will always imply causality as long as the public is scientifically illiterate.
I understand the message behind the video, but I disagree with Sal in that WebMD was indicating causality. MD used the words "may" and "seem," which are simply indicators of possibility. If MD would have used "will" or "does," then I would have called them on bad write-ups.
u do it by randomized controlled experiments....divide into two groups randomly, and then give them your breakfast, etc..but have to be really randomized.
From what I understood, the text and the research that was talked about were more about comparing the patterns of people who ate breakfast and who didn't and that was that. Rather than saying breakfast causes you to lose weight, which may or may not be true in each case.
Sorry, I think I wrote this in hurry and forgot to correct myself afterwards. Correlation is not necessarily causation but that isn't holy writ either. My point is that if the public doesn't know how research methods are conducted they can't make informed decisions.
I thought this video identifies the difference between correlation and causality, this video should have been named Eating Breakfast Causes Kids to Lose Weight.
Done with whole statistics playlist and I can say it was a great help. Now I believe I have the working knowledge of the topics covered in this series of lectures. Thank You!
Unexpected sudden ending of 68 video's. Thank you very much for clarifying all this material for me, your explanations are very clear and easy to watch!
I really like these videos from Khan Academy. I encourage my students to use them to do independent study. There is a definite correlation between using the videos and success in the examinations, but not proof of causality yet :)
I love this talk. This is one of the best treatments of Causality vs. correlation.
I come back to this talk every few months to re-inforce the arguments...it is really necessary to apply these lesson to so many things in life (Global warming, most of Modern cutting edge medicine, etc)...Remember they are measuring correlation and selling it as causality, to get FDA approvals or more research money.
@@datingprofile what's your profession?
After all these years of hating... I finally start liking statistics :)
I am really sad because of the fact that this is the final lesson.
Thank You Sal!
@ZergAteu The researchers probably understand the difference and may have even been careful with their wording. The article, however, is clearly trying to have the reader walk away with the thought that eating breakfast makes obesity less likely.
Just finished the playlist. It is remarkable how Sal has the talent to build concepts from ground up making the student really got into what is going on. This is a great example of humility, something usually missing on education (specially higher).
I now finished the whole playlist to get some introductory ideas about statistics. I think I got some now. It really provides a good value for learners, great videos, he explains well to make understand intuitively. He's great.
Wow! This video, apart from explaining the difference between correlation and causality, tells us why it is important to know the difference too. Causality is a correlation. However, not all correlations are causalities. What we need is the objectivity to see a causality as a causality and a correlation as a correlation. Above all, we must have the ability to picture the interactions between a multitude of correlations and causalities while evaluating an event.
So basically causality only works in one direction and correlation is 2 things that move together and are related.
My thoughts exactly
In all these examples, yes. But you could also have 2 things that are correlated completely coincidentally. See for example the website/book "Spurious Correlations".
"Breakfast is the most important meal of the day" says the breakfast-related industry
Causality always creates correlation, except in a few pathological cases. We generally say that X causes Y when doing X changes the probability of Y occurring. As a result, X and Y will be correlated. The point of this video was to say that correlation does not imply causation. The converse, however, is generally true. Causation implies correlation.
Correlation is saying that two things can go on simultaniously and cause and effect is saying that one thing happens which affects the next thing in the sequence. Directing this to the video, correlation is saying that sleeping late could impact your ability to eat breakfast because maybe they dont have enough time and vice versa and cause and effect is saying that because they are obese, they could not eat breakfast. That is just an example.
Thanks Prof for these mind boggling videos. Studied stats in my MBA from one of the most revered Faculty from an IIM and he was nowhere half as good as you. I had no idea stats could be this much fun. Thanks a lot. A bit sad to see the abrupt ending of 67 videos.
Same for me
I owe you big time, Sal. Nicely and Perfectly articulated the concepts. Thank you very much.
This is to the point. As I watched I kept thinking about the many articles that "might be wrong" that come out on a daily basis. These are powerful concepts to understand and apply regularly.
I Appreciate for all the statistics videos You are the one of the best teachers I ever meet！Your teaching is clear and kinda funny sometimes make me laugh lol anyway thank you again！！
Ideally, the authors would make the notes that Khan has made and provide some alternatives to causation as Khan did. This way, when the reader reads the article they don't come away thinking, "if I want to lose weight I just need to eat breakfast" but instead, "people are researching ways to help me lose weight, but they still don't know much besides "eat healthy and exercise".
I wish more people thought like this, not taking things at face value, making up their own mind. would certainly help with the media and their inflexions towards news and current affairs.
@ZergAteu
If what the study implied was a correlation rather than a causation, they shouldn't have said "eating breakfast may beat teen obesity" because it singles out one correlated factor. Why not name it "getting up early may been teen obesity" or "being more active may beat teen obesity"? Don't get me wrong, I agree with you that the article does not imply causality, but that's only because it's filled with weasel words.
Hi Khan. I'm a big fan of your work and like following you.
I hope you don't mind if I make a few suggestions. I think you did a great job at differentiating between causality and correlation.
A few things I would recommend mentioning in addition to this are: timing (which came first?), co-linearity (overcoming it to analyze causality), and randomization or multivariate models (how it takes care of the confounding issue you mention). Good luck and great job!
THE END.
Thank you so much Sal for this playlist. :)
Ok i ve seen an article telling that most rock climbing accidents happen by men rather than women. But as far as I know there are more men that goes rock climbing than women. Can we say there is an correlation or how can we implement it in causality, can you explain i really didn’t get it. The more men rock climbing causes more accidents, am I wrong?
A fluctuation of A and B tend to be observed at the same time, Instead of A causing B; That's a very good example to illustrate Correlation and Causation, in my opinion.
Thank you Professor for this amazing videos. I watched the whole playlist of Statistics and I find it inspiring. Longing to follow with another playlist soon.
What is causality but correlation between two variables?......... mind blown.
Thank u sir, you have expanded my way of thinking or looking at things to a different level, your videos are very helpful and I hope you make more of these useful videos. If you need our help through donations so you can make more videos, than add a link and I will gladly support.
This series ending was more abrupt than The Sopranos cut to black. Thanks, Sal!
This video is simple and to the point. Great video!
Completed the entire series. Loved it!
Thank you so much for helping our minds to open up
Thank you sir for this playlist. You are excellent sir.💕💖💖💖
I could not take the ending seriously! How he tackles that example near the end of the video made me laugh.
THANKS, THAT'S A BEAUTIFUL LECTURES.
Great playlist Sir. Thanks. This helped a lot.
Great. I was also hoping you'd throw in an explanation of what a logical corollary is. Although it's implied throughout. Still it's easy to confuse that with a correlation unless you know why they can't be the same thing.
This was an awesome journey Sal. Lots of respect and thanks.
Excellent contents - God bless you
Informative ❤
Great videos. Loved all of it. I feel the regression part could be expanded to talk of linear and multiple regression analysis, and also explaining how to interpret excel outputs of regression. But overall, really good!
that was awesome ,i think you only have to eat when you are hungry eating more doesn't make u skinny.
So... since life is filled with variables, how does one conclude actual cause and effect other than in a controlled lab or controlled environment where even those are often misinterpreted based on unforeseen conditions?
This question is why "social sciences" are seen as weak, or not rigorous -- because it is simply impossible to answer this adequately.
@ZergAteu
I didn't vote it down :-)
You might be underestimating the influence the title has over the way the rest of the article is read. If it starts out with "X may beat Y", then the reader will interpret the content as being in line with that summary. It's a tactic frequently used by tabloid journalists.
Really appreciate the info!
I MADE IT!!! Thanks for the videos, Sal, you helped me so much :)
As I was listening to the first couple of seconds I thought that maybe american titan General Mills is probably the one that funded that study cited in webmd. Anyway, great video, very khana - educational. People on jury duty should probably see this video before making any decision.
This playlist has been a great help sir, Thank you for the hard work.
Very Helpful. 67/67.
Causality and correlation are synonymous?
So the following is true?
As the number of people who breath air increases, the number of people that die increases. Therefore, breathing air causes death.
The more people there are, the more likelihood of friction and conflict, which breeds a ton of other reasons which lead to death.
Great post. This will save me so much tiresome re-explaining of bad health journalism to friends!
Great examples! Well done.
great video! thank you
My brains 🧠 , trying to process these information.
so does causation is more powerful instrument to proove our hypothesis than correlation?
Excellent video lesson, Sal.
Great piece here.
The funny thing is, even as a fairly fit guy, in the military, I started struggling with weight management in my thirties like a lot of people do, and one of the first things that really helped me start turning it around was compressing my eating window. Let's also be sure to acknowledge the fact that people dealing with poverty and food insecurity often eat the worst possible macronutrient composition (high fat + high processed carbs + low protein), which can be obesogenic via hormone signaling even if not massively overconsumed.
In this case, the readers do not likely know the difference between causation and correlation so the weasel words result in the reader believing something is true and backed up by science when really it isn't true (not that it's false, just that it's not proven to be true or false yet).
Do a tutorial on Correlation and Regression please!
Great Video
What are differences between correlation and magnitude? What is magnitude?
Great video Sal!
More about reading research papers please!
Even the same topic with a different example would be good.
Cheers!
best video on the subject
That how you end a series on statistics! With a 'Random' video :)
This is a good teaching
Thanks a lot for the videos, really appreciate it. I was wondering if you've posted any videos on SPSS?
Is this video about math or obesity? I can't tell.
Is obesity about math or breakfast? I can't tell. lol
Clearly about fuirt loops.
irrelevant comment. This topic forms the basics of the Data science analysis and very much related to statistics branch of mathematics
hmm one serious student who doesn't have a sense of humor
@@tkhasimbashaa This comment was irrelevant too. This topic ACTUALLY critically analyses the importance of ethical interpretation in statistics. n= 2n. bye.
Thanks
Thank you.
Amazing how writers in the medical field would write these kinds of articles. Either a lack of education or good amount of corruption
Thanks!
굉장한 설명이었다
For correlation you can’t say it will make things better
I usually have coffee for breakfast. Now I'm wondering if I should at least consider brunch.
How to know if correlation is causal through historical data, no controlled experiments?
I think giving a more obvious example (e.g. the correlation between the number of known stars and the population of earth) would've been a better way to start, like c0nc0rdance did in his video on the subject.
@ZergAteu
@qtutoringhelps
I believe a big part of the problem here lies in the target audience. If a paper was published in a scientific journal that used a weasel word here or there the readers, other scientists, would understand the difference between causation and correlation (hopefully). However, the referenced WebMD article is meant to be an easy to digest summary of the study fit for consumption by the masses (non scientists).
So, how do you know if there is causality?
[INTERPRETATION
] what makes you so confident in case you encounter a "realistic" causality? is there any robust academic foundation? I couldnt find any clue in your
greatest lecture in this clip. Am I an idiot? I would politely ask you to answer or explain bits on this stupid kindergargetn level question please.
carbs are essential, I hate that there is this stigma behind them. Every cell in your body runs of glucose, aka simple/complex sugars... sorry just had to get that out there. great video!
+indigochick Yes, all three macro nutrients are important, all three of them have been demonized at some point in history when it comes to weight management.
It is still hard to believe for people that weight management is just as simple as: total kcal intake and total kcal expenditure. But you know, that is not fancy enough for articles etc.
causality and correlation are synonymous. It doesn't matter if you eat cinnamon buns or bran muffins in the morning you will become overweight at some point in your life if you keep this type of diet. This study is typical research news the public is fed via the media. If the public is generally ignorant(or skeptical) of research methods how are they supposed to make informed decisions about diet.
What is the conclusion of the video?
causation does not imply correlation.
@alique087
We cannot judge the intentions of the article but we can attempt to draw conclusions about the effects the article has on it's target demographic (non-scientists). A non-scientist reading this article is likely to come away with the impression that eating breakfast will help them lose weight. The article does not technically show causation, but it does imply causation to the average reader who is unfamiliar with causation vs correlation.
So smoking correlates with cancer and does not cause it? Smokers tend to engage in other risky activity which promotes cancer?
When can we get to the really cool stuff(like logistic regression)?
Correlation sounds boring. Causality and narrative are what our minds crave. Journalists will always imply causality as long as the public is scientifically illiterate.
I understand the message behind the video, but I disagree with Sal in that WebMD was indicating causality. MD used the words "may" and "seem," which are simply indicators of possibility. If MD would have used "will" or "does," then I would have called them on bad write-ups.
Then how to identify causality?
Yeah thats what I was thinking during this too!
u do it by randomized controlled experiments....divide into two groups randomly, and then give them your breakfast, etc..but have to be really randomized.
You try to induce B by applying A to a test group, rather than simply waiting for A to happen and looking if B is there too.
i dont still get it. where's the correlation in there?
I hope you reply me. I really need the answers for our reporting
From what I understood, the text and the research that was talked about were more about comparing the patterns of people who ate breakfast and who didn't and that was that. Rather than saying breakfast causes you to lose weight, which may or may not be true in each case.
By virtue of its intentional humorous content, this comment compelled me into a state of continuous guffawing.
I also wager that folks who eat breakfast get up earlier in the day than people who do not.
Salman Khan! Man Of Science! All Hail!
beautiful
Lawyered.
I think eating breakfast will not help, if your breakfast is made of big mac, three cheesburgers, fried potatoe and a big coke portion...
This video caused me to want bacon and Fruitloops.
How about argue the difference between common sense and nonsense?
Sorry, I think I wrote this in hurry and forgot to correct myself afterwards. Correlation is not necessarily causation but that isn't holy writ either. My point is that if the public doesn't know how research methods are conducted they can't make informed decisions.
Maybe breakfast eaters are more active because they are actually awake all day instead of sleeping half the day then waking up and not doing shit.
I thought this video identifies the difference between correlation and causality, this video should have been named Eating Breakfast Causes Kids to Lose Weight.
I usually eat relatively high protein breakfasts.
it makes me more carefull to conclude the result, but i become more confuse now