My Math Forum All N samples within one sigma

 January 20th, 2017, 06:19 AM #1 Newbie   Joined: Feb 2012 Posts: 3 Thanks: 0 All N samples within one sigma If I have a random variable which has a normal distribution with mu and sigma, what is the probability that if I draw N samples that they are all less than one sigma away from each other? I am not having any luck coming up with how to calculate this answer. I did write a program to try to determine the answer empirically. 10,000,000 tests per draw set: Draw: 2, All with in sigma: 84.27% Draw: 3, All with in sigma: 66.63% Draw: 4, All with in sigma: 50.96% Draw: 5, All with in sigma: 38.17% Draw: 6, All with in sigma: 28.16% Draw: 7, All with in sigma: 20.57% Draw: 8, All with in sigma: 14.90% Draw: 9, All with in sigma: 10.71% Draw: 10, All with in sigma: 7.69% Would it be something like: $\displaystyle \int_{x-\sigma}^{x+\sigma} \frac {1} {\sigma \sqrt{2 \pi}} e^{\frac {-(x{'}-\mu)^2}{(2 \sigma^2)}} dx{'}$ Any thoughts?
 January 20th, 2017, 02:09 PM #2 Senior Member     Joined: Sep 2015 From: CA Posts: 908 Thanks: 489 I believe you'll want to use order statistics. All of the samples being within a sigma of each other implies $\max(X_i) - \min(X_i) \leq \sigma$ So you're going to have to find the joint distribution of the min and max order statistics and then integrate that over the area that corresponds to the expression above. This isn't impossible but it's fairly involved. Give it a shot and come back w/specific questions if you run into problems.
January 20th, 2017, 06:15 PM   #3
Global Moderator

Joined: May 2007

Posts: 6,203
Thanks: 486

Quote:
 Originally Posted by histrung If I have a random variable which has a normal distribution with mu and sigma, what is the probability that if I draw N samples that they are all less than one sigma away from each other? I am not having any luck coming up with how to calculate this answer. I did write a program to try to determine the answer empirically. 10,000,000 tests per draw set: Draw: 2, All with in sigma: 84.27% Draw: 3, All with in sigma: 66.63% Draw: 4, All with in sigma: 50.96% Draw: 5, All with in sigma: 38.17% Draw: 6, All with in sigma: 28.16% Draw: 7, All with in sigma: 20.57% Draw: 8, All with in sigma: 14.90% Draw: 9, All with in sigma: 10.71% Draw: 10, All with in sigma: 7.69% Would it be something like: $\displaystyle \int_{x-\sigma}^{x+\sigma} \frac {1} {\sigma \sqrt{2 \pi}} e^{\frac {-(x{'}-\mu)^2}{(2 \sigma^2)}} dx{'}$ Any thoughts?
the integration limits are $\displaystyle \mu -\sigma ,\mu +\sigma$

 January 21st, 2017, 07:57 AM #4 Newbie   Joined: Feb 2012 Posts: 3 Thanks: 0 @mathman: I think that would be the probability that the random variable drawn is with in 1 $\displaystyle \sigma$. I need to know if multiple sample are drawn (N), the probability they are all within 1 $\displaystyle \sigma$ of each other. @romsek: Googling, reading and digesting...
January 21st, 2017, 04:55 PM   #5
Global Moderator

Joined: May 2007

Posts: 6,203
Thanks: 486

Quote:
 Originally Posted by histrung @mathman: I think that would be the probability that the random variable drawn is with in 1 $\displaystyle \sigma$. I need to know if multiple sample are drawn (N), the probability they are all within 1 $\displaystyle \sigma$ of each other. @romsek: Googling, reading and digesting...

 Tags samples, sigma

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post EvanJ Probability and Statistics 1 September 4th, 2016 05:50 AM Jessicasamide Algebra 1 November 24th, 2012 06:47 PM jhon13 Advanced Statistics 1 May 23rd, 2012 11:40 PM luketapis Calculus 2 March 27th, 2012 11:30 AM hunnybee Algebra 1 June 23rd, 2011 05:26 AM

 Contact - Home - Forums - Top