This is hands down the clearest and most intuitive explanation that I've ever encountered, whether on YT or in print. It has always surprised me how folks who normally do a great job at explaining statistics in plain language, when it comes to this particular question, either just evade it or resort to mathematical formulas that I bet look obvious and crystal clear to the inititiated but won't help those who only begin to learn statistics and are simply curious about the details of a notion as fundamental as SD. You, sir, did a great job here and I will be forever grateful to you for this!
I don't get what you're saying around 5:56. If there is less variability with a smaller sample, why do your sample mean drawings have a more spread out distribution with the n=10 than the n=50. Aren't you saying the opposite here?
The reason you're so good at teaching is because you're focused on interaction, understanding and thinking: Instead of writing the answer right away, you ask us, the viewers, to think. Then you explain why we're thinking, and at the end, we understand. Thank you very much.
I had to check when you posted the last video, because I didn't remember when I watched one of your videos the last time. No longer taking stats but it's impossible not to watch. You are simply awesome explaining this stuff.
Wish I had you as my statistics teacher in school. Like so many others here have commented, you provided a very intuitive answer to something I have been wondering abt for a long long time. Thx Mr. Nystrom and please keep more maths videos coming!👍👍
Your awesome man. Keep up the good work. I do have a suggestion Keep letting loose & be yourself. Your passion & love of statistics is infectious. Hold nothing back, the craziervypu are, the more memorable
SUCH A WELL THOUGHT OUT EXPLANATION. I came here after watching: Statquest: where he explains this n-1 thing using derivative of variance formula Dave Your Tutor: who explains using degrees of freedom Knan Academy: where Sal explains this using a simulation. Your video summed it all up so well...
Good God...you are a math angel...that shit has puzzled the docket s outta me for too long......may peace and prosperity be upon your children's children ...thank you my brother..
I am helping my son with this, and I found you...❤❤❤❤❤ it's like watch Steve Irwin talk about crocs or better yet, watching Steve on Blues Clues!!! Thank you for all your energy!!!
I got this video recommended on Google and thought it would be absolutely unjust if I didn't hit the like button! Because this video is just sooo perfect! On point!! 👍👍
You need to have all of your videos aired in every school in the world! I am 50 years old and I took statistics and earned an A from watching you and other videos.. Mainly yours... It just makes sense when you teach it. I applaud you!!!! #teacherofthecentury!!!1
I've watched a few different teachers on TH-cam but I leave still feeling confused, your really good at explaining! Thank you! Keep doing what your doing!
Thank you so much! My stats lecturer glossed over this and I've been worried all week... looks like I'll be following your channel all throughout university!
Mr. Nystrom, YES it did make a TON of sense! I am re-learning statistics on my own. My book (by John Freund) explains what the sample standard deviation is for but doesn't go into all the detail that you do here. (1) The purpose of the sample is to get an unbiased picture of the population, and (2) the bigger the sample, the lesser the correction. BEAUTIFUL! Thank you!
You are such a great teacher! I just asked this exact question on your other video explaining standard deviation whoops! Really appreciate the way you break things down and explain why everything is the way it is in the formulas. Great teaching style and personality:)
Thx! It`s really vivid and clear explaination. I`d like to add on why n-1. Since it can be n+1. quote from wikipedia "In statistics, Bessel's correction is the use of n − 1 instead of n in the formula for the sample variance and sample standard deviation, where n is the number of observations in a sample. This method corrects the bias in the estimation of the population variance. It also partially corrects the bias in the estimation of the population standard deviation. However, the correction often increases the mean squared error in these estimations. This technique is named after Friedrich Bessel. "
Finally I understood the reason for n-1!! I got so irritated everytime I used n-1 without understanding the reason. Thanks for the simple and lively video.
Now I find hope again and I think I'll pass my stats course this year for sure. Thank you. 👏🏻👏🏻👏🏻 Universities should not be accepting any less professors than you, for real:))
Thanks this was just great. Not having a math background at all....and sitting exam next week for social science freaked me out....have to say you done more in these lessons than the lecturer done in 4 weeks! Blessings from Scotland.
Just awesome, you just explained exactly what I needed. Currently, I am facing several problems trying to know which formula to use for different statistical parameters. You just helped me understand one of them, thank you very much.
Thanks man! Great video again. I would love also to have an in depth explanation on the degrees of freedom. An explanation between the normal distribution and the Chebyshev distribution would be great too! Thanks again !
Very good explanation. I think it is more of an observation that an explanation. I would suggest you to include the following at the end "We are using sample mean instead of the actual mean for the sample standard deviation calculation. That is, the term Xi minus the average is not there. That specific Xi minus average is 0. That term is always missing from the sample. It is because we do not have that term, we divided by N -1 instead." It is very difficult to express. You have done a very good job. ☺️
Another way to look at is as follows. Does calculating the standard deviation of a sample with n=1 make sense? Also when one is calculating s, one uses xbar. So if you have n values, and you know (n-1) values and the average of n values, your nth value is not independent anymore.
OUTSTANDING, a new video! Starved and in withdrawal, started calculating the probability of seeing a new video again: p=0.15 it wasn't looking good :-)
I have to admit that you have explained exactly what the leaners needed and diffrenrce between two formuls where the most of people get confusion(welldone)the winds frome Somalia.
Thank you sir, that made lots of sense and finally explains it for me. I'm not even studying stats or maths but just enjoy the topic and your videos are very clear and fun to watch. Keep it up!
Great explanation! This is a hard topic. Somehow mathematically we divide by (n-1) to make our sample standard deviation less deviated from the population standard deviation.
but y subtract 1.. y not 2 or 3? it just kind of seems like a hit and miss way of doing things..is there a more logical reason y subtracting 1 gives a more meaningful answer than subtracting 2 or 3?
its to do with the degrees of freedom..because these samples have been used to estimate one model parameter which is Mean..because of that we reduce the 1 from the sample size which is n..because standard deviation is build upon finding the mean first...So it has already introduced an error to the model..so to reduce that error in a statistical way we make the sample size (n-1). If its in regression where we use Y and X to get the regression error or standard error of the regression model we use (n-2) since we have used mean of Y and Mean of X for estimating the regression error of the model. So I hope this helpe :)
YOU ARE AMAZING!!! When I first found the video I thought there's no way Imma spend 12 minutes to find out why I came across two formulas for the standard deviation. Yet here I am more excited than ever about stats! thanks so much! And btw, could your or anybody reading these comments tell me the difference between the standard error and the standard deviation of the sample? Thanks a lot!
Nice informative video. I request you to make one similar video on the importance of the use of the degree of freedom (n-1) intend of n for ANOVA calculation.
hey, thank you for the video. Still do not get the concept. What you are saying is based on the argument that we always underestimate. But what if samples we took made us overestimate? Is it possible? How do we know if we over or underestimate?
Holy shit this guys enthusiasm is infectious. The world needs more teachers like this guy.
the passion that you have, makes me want to learn
This is brilliant... answering a question I had never had properly explained for years. Cheerful examples are the best way :)
This is hands down the clearest and most intuitive explanation that I've ever encountered, whether on YT or in print. It has always surprised me how folks who normally do a great job at explaining statistics in plain language, when it comes to this particular question, either just evade it or resort to mathematical formulas that I bet look obvious and crystal clear to the inititiated but won't help those who only begin to learn statistics and are simply curious about the details of a notion as fundamental as SD. You, sir, did a great job here and I will be forever grateful to you for this!
Dude, Youre a legend . This was by far the best explanation of this nitty gritty concept Ive seen. You rock man.
This is the best video that was able to explain why n-1 in sample SD makes sense by far
I don't get what you're saying around 5:56. If there is less variability with a smaller sample, why do your sample mean drawings have a more spread out distribution with the n=10 than the n=50. Aren't you saying the opposite here?
I have never seen such a better teacher in my whole life
The reason you're so good at teaching is because you're focused on interaction, understanding and thinking: Instead of writing the answer right away, you ask us, the viewers, to think. Then you explain why we're thinking, and at the end, we understand. Thank you very much.
I had to check when you posted the last video, because I didn't remember when I watched one of your videos the last time. No longer taking stats but it's impossible not to watch. You are simply awesome explaining this stuff.
Joel Batista exact same thing as me
Joel Batista haha same here I still watch this guy even after finishing the course
same
Wish I had you as my statistics teacher in school. Like so many others here have commented, you provided a very intuitive answer to something I have been wondering abt for a long long time. Thx Mr. Nystrom and please keep more maths videos coming!👍👍
Mate. You are my hero. If mathematics was taught mainstream the way you teach it, we would have colonised the galaxy by now.
I'm not even taking the piss.
Man, that got me over a huge mental block! You're the first person to explain that in a way I understood. Thank you so much!!!
The clearest way to explain this! Well done Sir!
This is such a good and logical explanation I have ever heard about n-1 denominator. I love the way you teach.
Thanks for the short class on N-1. You are the first one to make it make sense!
Your awesome man. Keep up the good work. I do have a suggestion
Keep letting loose & be yourself. Your passion & love of statistics is infectious. Hold nothing back, the craziervypu are, the more memorable
SUCH A WELL THOUGHT OUT EXPLANATION.
I came here after watching:
Statquest: where he explains this n-1 thing using derivative of variance formula
Dave Your Tutor: who explains using degrees of freedom
Knan Academy: where Sal explains this using a simulation.
Your video summed it all up so well...
This was awesome. Started an intro course for pob/stats two weeks ago and this was exactly what I needed. Love the energy!
I'd love to see an in-depth video on degrees of freedom. Happy to have you back!
you are the goat man, your explanation was soo frikin clear it will stay in my memory forever
Good God...you are a math angel...that shit has puzzled the docket s outta me for too long......may peace and prosperity be upon your children's children ...thank you my brother..
My professor explained it to me and it went right over - this video made complete sense. I can now explain this to anybody.
I am helping my son with this, and I found you...❤❤❤❤❤ it's like watch Steve Irwin talk about crocs or better yet, watching Steve on Blues Clues!!! Thank you for all your energy!!!
This is the best video I have found that explains the n vs n-1.
I got this video recommended on Google and thought it would be absolutely unjust if I didn't hit the like button! Because this video is just sooo perfect! On point!! 👍👍
You need to have all of your videos aired in every school in the world! I am 50 years old and I took statistics and earned an A from watching you and other videos.. Mainly yours... It just makes sense when you teach it. I applaud you!!!! #teacherofthecentury!!!1
This thing was messing with my brain for 1 week. After one day of crying and scratching my head i found your video Thank You
I'll be 59 this year (2017)... and I'm still learning through your videos. Than you sir.
I just started my stats class, but I don't even bother with the textbook because you explain things so much better!
I've watched a few different teachers on TH-cam but I leave still feeling confused, your really good at explaining! Thank you! Keep doing what your doing!
Thank you so much! My stats lecturer glossed over this and I've been worried all week... looks like I'll be following your channel all throughout university!
Mr. Nystrom, YES it did make a TON of sense! I am re-learning statistics on my own. My book (by John Freund) explains what the sample standard deviation is for but doesn't go into all the detail that you do here. (1) The purpose of the sample is to get an unbiased picture of the population, and (2) the bigger the sample, the lesser the correction. BEAUTIFUL! Thank you!
You are such a great teacher! I just asked this exact question on your other video explaining standard deviation whoops! Really appreciate the way you break things down and explain why everything is the way it is in the formulas. Great teaching style and personality:)
Thx! It`s really vivid and clear explaination. I`d like to add on why n-1. Since it can be n+1. quote from wikipedia "In statistics, Bessel's correction is the use of n − 1 instead of n in the formula for the sample variance and sample standard deviation, where n is the number of observations in a sample. This method corrects the bias in the estimation of the population variance. It also partially corrects the bias in the estimation of the population standard deviation. However, the correction often increases the mean squared error in these estimations. This technique is named after Friedrich Bessel. "
why aren't your my statistics professor, you may be the first one that actually taught me something
One of the best explanations out there. Crystal clear. Good job buddy.
I am straight up addicted to your videos. Your amazing at explaining this stuff!
Perfect , brilliant .This helped to find why n-1 not n-2 with your example of proportion of 5/5 ,10/10 and so on.
Thanks a lot.
Finally I understood the reason for n-1!! I got so irritated everytime I used n-1 without understanding the reason. Thanks for the simple and lively video.
You are absolute genius explaining these things.
Now I find hope again and I think I'll pass my stats course this year for sure. Thank you. 👏🏻👏🏻👏🏻 Universities should not be accepting any less professors than you, for real:))
Dude that saved my grades way back in the day is back!! Lots of love man!
Mr Nystrom , You are a legend!, Your videos helped me to clear my Machine learning exam, Please do continue posting the videos.
You are freaking amazing . For several years I had this question and now i am so so so happy . You are freaking genius dude.
Thanks this was just great. Not having a math background at all....and sitting exam next week for social science freaked me out....have to say you done more in these lessons than the lecturer done in 4 weeks! Blessings from Scotland.
They should introduce a Nobel Prize for Passion.
You'll be the first winner.
Bravo! 👏🏻
After seeing around 5-6 videos, finally I understood here. Thank you.
Just awesome, you just explained exactly what I needed. Currently, I am facing several problems trying to know which formula to use for different statistical parameters. You just helped me understand one of them, thank you very much.
Thanks man! Great video again.
I would love also to have an in depth explanation on the degrees of freedom.
An explanation between the normal distribution and the Chebyshev distribution would be great too!
Thanks again !
Hello and thanks from Michigan! Your passion is infectious.
Damn you're the best! That fraction example to explain how to correct for the bias is awesome.
wooow you are an amazing teacher....with great teaching skills.... I completely understand the whole concept
Yes sir, that made perfect sense! Thank you for your continuous effort to spread good education.
You are amazing!!! I hope our professor would have had half of your teaching skill.
Its easy to understand your explanation. Good work
Thank you so much for making this video, It is so helpful I wish I had educators like you
On spot.... keep it up💥💥💥💥💥💥
this makes me enjoy learning stats more and always an inspiration to watch you teach and spread the knowledge.
Very good explanation. I think it is more of an observation that an explanation. I would suggest you to include the following at the end
"We are using sample mean instead of the actual mean for the sample standard deviation calculation. That is, the term Xi minus the average is not there. That specific Xi minus average is 0. That term is always missing from the sample. It is because we do not have that term, we divided by N -1 instead."
It is very difficult to express. You have done a very good job. ☺️
hey this is the best explaination so far on this thing . salute
This video is amazing!! I have my Algebra 2 quiz tomorrow and this is answering all of my questions!!! I love it!
Brilliant, simply brilliant. Can't be overstated.
Finally, I understood this. My fault for never asking, but I'm equally shock none of my professors explained this.
As in your other vids you explain it so one can really get the idea behind the thing. Awesome teacher! and fun to watch :)
The best 12 minutes of my life so far ❤️❤️ thanks a ton for the amazing explanation
love this presentation....at 10:46 you want to square those deviations from the mean :) .... great teaching!
Finallly i understood after watching so many videos !!!!!!!!! THANKSSS
You'r videos helped me pass my Stat class, i am grateful.
Another way to look at is as follows.
Does calculating the standard deviation of a sample with n=1 make sense?
Also when one is calculating s, one uses xbar. So if you have n values, and you know (n-1) values and the average of n values, your nth value is not independent anymore.
I like the way you explain with pretty cool nature cracking jokes. You made stat not only easy but also funny. Thanks a lot
You make stats understandable and fun. Thank you!
Brilliant explanation for someone who needs deeper understanding!!! And entertaining :)
Thanks for your explanation! Everything becomes more clearer
The best teacher at BHS hands down
He's back! The legend returns!
Thank you!!! You just maybe my savior to getting through Prob and Stats this semester!
Never thought Aston Kutcher is so smart!!! Thank you!
I can’t wait to see all your playlists
MrNystrom your lessons are amazing, explanations are intuitive. Keep up the good work. Also looking forward to your new videos.
OUTSTANDING, a new video! Starved and in withdrawal, started calculating the probability of seeing a new video again: p=0.15 it wasn't looking good :-)
you are one of my top 5 favourite people of all time. i wish you'd make more videos on stats :)
Very helpful explanation and very clearly put. Thanks!
I have to admit that you have explained exactly what the leaners needed and diffrenrce between two formuls where the most of people get confusion(welldone)the winds frome Somalia.
This is a totally excellent video!!!! I love you sooooo much dear mr Nystrom !
Thank you sir, that made lots of sense and finally explains it for me. I'm not even studying stats or maths but just enjoy the topic and your videos are very clear and fun to watch. Keep it up!
Wow! Crystal clear explanation.
Great explanation! This is a hard topic. Somehow mathematically we divide by (n-1) to make our sample standard deviation less deviated from the population standard deviation.
but y subtract 1.. y not 2 or 3? it just kind of seems like a hit and miss way of doing things..is there a more logical reason y subtracting 1 gives a more meaningful answer than subtracting 2 or 3?
when n-2, n-3 and n-4 etc.
its to do with the degrees of freedom..because these samples have been used to estimate one model parameter which is Mean..because of that we reduce the 1 from the sample size which is n..because standard deviation is build upon finding the mean first...So it has already introduced an error to the model..so to reduce that error in a statistical way we make the sample size (n-1). If its in regression where we use Y and X to get the regression error or standard error of the regression model we use (n-2) since we have used mean of Y and Mean of X for estimating the regression error of the model. So I hope this helpe :)
Excellent. I wish I saw this 35 years ago.
4:18 What a reaction. Just loved the video :-)
Thank you, no one ever explained this to me.
great explanation! really allows for an intuitive understanding
YOU ARE AMAZING!!! When I first found the video I thought there's no way Imma spend 12 minutes to find out why I came across two formulas for the standard deviation. Yet here I am more excited than ever about stats! thanks so much! And btw, could your or anybody reading these comments tell me the difference between the standard error and the standard deviation of the sample? Thanks a lot!
Nice informative video. I request you to make one similar video on the importance of the use of the degree of freedom (n-1) intend of n for ANOVA calculation.
Good stuff, buddy. Learning a lot from you. 😃
That was a bloody good description. :)
Great clarity in this presentation for standard deviation.
hey, thank you for the video.
Still do not get the concept. What you are saying is based on the argument that we always underestimate.
But what if samples we took made us overestimate? Is it possible? How do we know if we over or underestimate?
I saw in an text book that for n>40 use n only as the distinction becomes less important the higher the sample becomes.