Is there a central limit theorem for Median?

Click For Summary
SUMMARY

The discussion addresses the existence of a central limit theorem for the sample median, establishing that under specific conditions, the sample median converges to a normal distribution. The key assumptions include a model defined by F(x - θ) with F(0) = 1/2 and a positive density f(0). Under these conditions, the sample median satisfies the asymptotic distribution √n(θ̂ - θ) → N(0, σ²), where the asymptotic variance is σ² = 1/(4f²(0)).

PREREQUISITES
  • Understanding of central limit theorem for means
  • Familiarity with statistical distributions and convergence
  • Knowledge of sample median and its properties
  • Basic concepts of probability density functions
NEXT STEPS
  • Research the conditions for the central limit theorem for medians
  • Study the implications of unique population medians in statistical analysis
  • Explore asymptotic distributions in statistics
  • Learn about the role of density functions in determining variance
USEFUL FOR

Statisticians, data analysts, and students studying statistical inference who are interested in the properties of sample medians and their convergence behavior.

grossgermany
Messages
53
Reaction score
0

Homework Statement


Hi, We know the famous central limit theorem for means.
I wonder if there is a central limit theorem for Median?
If so under what regularity condition, does the median converge to a normal distribution with mean and variance equal to what?

Homework Equations





The Attempt at a Solution

 
Physics news on Phys.org
One of the simplest results is this:
* Assume the model is F(x - \theta)
* Assume both F(0) = 1/2 and that the density f(0) > 0

(these assumptions mean the population median is unique)

Then the sample median satisifies

<br /> \sqrt n \left(\hat \theta - \theta\right) \rightarrow n(0, \sigma^2)<br />

in distribution, where the asymptotic variance is given by

<br /> \sigma^2 = \frac 1 {4f^2(0)}<br />
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 23 ·
Replies
23
Views
4K
Replies
3
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 31 ·
2
Replies
31
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
11K