Does Sampling Variance Converge to Theoretical Variance in Simulations?

  • Context: Graduate 
  • Thread starter Thread starter jimmy1
  • Start date Start date
  • Tags Tags
    Variance
Click For Summary

Discussion Overview

The discussion revolves around the relationship between sampling variance and theoretical variance in the context of simulations involving a random variable X. Participants explore whether the average of variances from multiple simulations can converge to the theoretical variance, despite individual variances potentially differing significantly.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • One participant questions if the average variance of a random variable X from multiple simulations will converge to the theoretical variance, despite individual variances potentially differing widely.
  • Another participant distinguishes between theoretical variance, which is fixed, and experimental variance, which can vary across experiments, stating that the theoretical mean of experimental variance aligns with the theoretical variance.
  • A later reply reiterates the initial question about the convergence of average variances and introduces the consideration of sample size and the implications of using n-1 versus n in variance calculations.
  • There is a suggestion to find the expectation of the sample variance to demonstrate that it may not equal the theoretical variance.

Areas of Agreement / Disagreement

Participants express differing views on the relationship between sampling variance and theoretical variance, indicating that the discussion remains unresolved with multiple competing perspectives.

Contextual Notes

There are limitations related to the assumptions made about sample sizes and the definitions of variance being used, which may affect the conclusions drawn from the discussion.

jimmy1
Messages
60
Reaction score
0
if a random variable X has variance sigma, does it mean that the average variance of X will be sigma, or will X always have variance sigma.
So if I did 10 simulations of the random variable X, is it plausiable that the variances differ widely, but the average of the variances converge to the theoretical variance, or do each of the variance have to be the same.
 
Physics news on Phys.org
It looks like you are mixing up two different usages of the term variance. There is a theoretical variance which is fixed.

However the experimental variance can vary from experiment to experiment. The theoretical mean of the experimental variance is the theoretical variance.
 
Thanks, that cleared things up!
 
jimmy1 said:
So if I did 10 simulations of the random variable X, is it plausiable that the variances differ widely, but the average of the variances converge to the theoretical variance, or do each of the variance have to be the same.

If you do not consider n-1 to be approximately equal to n, the avarage variance will not be equal to the theoretical variance nor do each of the sample variances will be same. Find 'expectation' of the sample (of size n) variance and check that it is not equal to theoretical variance.
 

Similar threads

  • · Replies 42 ·
2
Replies
42
Views
6K
  • · Replies 1 ·
Replies
1
Views
436
  • · Replies 6 ·
Replies
6
Views
3K
Replies
5
Views
6K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 28 ·
Replies
28
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 7 ·
Replies
7
Views
6K
  • · Replies 9 ·
Replies
9
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K