I'm glad to hear you're watching these for fun -- I think that's an angle that I should play up a little more :) I've been desperately trying to free up some time to get back to video production, and I will find a way, but there's a good chance that I won't be able to make more videos before the summer. I'll see what I can do. Thanks for your interest!
I know it's been 4 years but it doesn't hurt to answer that with the distribution this would likely be impossible, since the t distribution itself is reliant on the sample size
Hello, thank you for making these videos. They are very helpful - I had a question. Why does a larger confidence interval require a larger sample size? I thought increasing the sample size decreases the width of confidence interval - because it decreases the standard error. Am I misunderstanding something? Thank you.
Yes. These calculations are all done under the assumption of a (continuous) normally distributed variable, so there is no meaningless distinction between having a margin of error that is "less than 4" or "no more than 4".
for question like "Suppose a sample of size n is to be selected from a normal distribution where sd is known to be 14.7 What is the smallest value of n to guarantee that the length of the 95% confidence interval for the mean u is less than 3?" i kept on getting 93 and im not sure if it is the correct answer, since apparently the answer is 369 and i got 93 by (1.96(14.7) / 3 ) ^2. Thanks for the help.
Hello @jbstatistics, just wanted to say hi and let you know that there is website www.patreon.com/ where you can offer your knowledge to your subscribers. Same as here on youtube but where you can make a living off of it. 3Blue1Brown and Dr Jordan B Peterson are more known on that site, to name a few. Cheers
Thanks, I appreciate the suggestion. At some point I'll investigate that type of thing further, but I'm a bit of a nutbar -- I'm just doing my bit to improve statistics education.
you are beginning to be my favorite person right now... I have been looking for an easy to understand formula for this for hours
I'm glad to be of help Tycho. Cheers.
This was exactly what I was looking for, thanks for saving my sanity~
I'm glad I could help!
I watch these for fun now. Can we expect another series of videos from you in the near future?
I'm glad to hear you're watching these for fun -- I think that's an angle that I should play up a little more :)
I've been desperately trying to free up some time to get back to video production, and I will find a way, but there's a good chance that I won't be able to make more videos before the summer. I'll see what I can do. Thanks for your interest!
Since the standard deviation is usually not known, can you make a video like this but use the t-distribution instead of the normal distribution?
I know it's been 4 years but it doesn't hurt to answer that with the distribution this would likely be impossible, since the t distribution itself is reliant on the sample size
Sir, why this formula is not being used [z^2(standard deviation) (1-standard deviation) ]\ e^2
If we dont know sigma but we kne the sample st. Deviation , would we be using the t critical value instead of z ?
Can you do this with unknown sigma and population correction factor?
sir, does m also means the width of the confidence interval?
m is half the width of the interval. (The endpoints of the interval are found by adding and subtracting m from x bar.)
Hello, thank you for making these videos. They are very helpful - I had a question. Why does a larger confidence interval require a larger sample size? I thought increasing the sample size decreases the width of confidence interval - because it decreases the standard error.
Am I misunderstanding something? Thank you.
Great Work! , queries solved on comments too! is this an ideal world
I'm glad to be of help!
may i know the reference for this formula??
plz what do excetly mean meu with in 4
help
what does "within four" mean?
"Within 4" means at most 4 units away. For example, 8.2 is within 4 of 10, but 5.5 is not.
Hi, if it is “less than 4” do i sub m=4?
Yes. These calculations are all done under the assumption of a (continuous) normally distributed variable, so there is no meaningless distinction between having a margin of error that is "less than 4" or "no more than 4".
for question like "Suppose a sample of size n is to be selected from a normal distribution where sd is known to be 14.7 What is the smallest value of n to guarantee that the length of the 95% confidence interval for the mean u is less than 3?" i kept on getting 93 and im not sure if it is the correct answer, since apparently the answer is 369 and i got 93 by (1.96(14.7) / 3 ) ^2. Thanks for the help.
The *length* of the interval is less than 3. If the length is 3, what's the margin of error?
It didnt say.. I guess there is no margin of error.. Does the question work without a margin of error?
I'm asking you. If the length of the interval is 3, what is the margin of error? Hint: It's not 3.
Thank you sir =)
You are very welcome Ahmad.
saved me
Hello @jbstatistics, just wanted to say hi and let you know that there is website www.patreon.com/ where you can offer your knowledge to your subscribers. Same as here on youtube but where you can make a living off of it. 3Blue1Brown and Dr Jordan B Peterson are more known on that site, to name a few. Cheers
Thanks, I appreciate the suggestion. At some point I'll investigate that type of thing further, but I'm a bit of a nutbar -- I'm just doing my bit to improve statistics education.