2
$\begingroup$

How do you measure the central tendency of a cauchy distribution?

I'm aware that the mean is not a good measure of the central location for cauchy. Can I use median?

I've been searching with the key word: central "tendency" of "cauchy" distribution stackexchange for a long time, and found no satisfactory answer that states how I can measure central tendency of Cauchy distribution. Everybody says that "there exists techniques that deals with...bla...bla..." that never gives a solid example of what there are.

Can someone give me an actual example?

$\endgroup$
3
  • 1
    $\begingroup$ Wikipedia on Cauchy Dist'n under Estimation: "One simple method is to take the median value of the sample as an estimator of $x_{0}$ [center]... Other, more precise and robust methods have been developed. For example, the truncated mean of the middle 24% of the sample order statistics produces an estimate for $x_{0}$ that is more efficient than using either the sample median or the full sample mean. $\endgroup$
    – BruceET
    Commented Sep 23, 2019 at 6:42
  • $\begingroup$ Also, because the density function is positive at the median, the distribution of the sample median is asymptotically normal for increasing sample size. $\endgroup$
    – BruceET
    Commented Sep 23, 2019 at 7:34
  • $\begingroup$ My first search of this site for Cauchy tendency yielded several good answers. Variants of this search produce even more. $\endgroup$
    – whuber
    Commented Sep 23, 2019 at 15:50

1 Answer 1

1
$\begingroup$

This may not be the best example of the claim in my Comment, but here are estimates of the center $10$ of a Cauchy with scale parameter 1.

Medians and trimmed means were found (in R) for a million samples of size $n = 100.$ The standard deviation of the trimmed means is a little smaller.

set.seed(922)
h = replicate(10^6, median(rcauchy(100,10)))
mean(h); sd(h)
[1] 9.999961
[1] 0.1585856
a.t = replicate(10^6, mean(rcauchy(100,10), tr=.38))
mean(a.t); sd(a.t)
[1] 10.00004
[1] 0.1535004

enter image description here

$\endgroup$
3
  • $\begingroup$ Sampling doesn't seem to have anything to do with a question that is solely about a distribution. $\endgroup$
    – whuber
    Commented Sep 23, 2019 at 14:52
  • $\begingroup$ Pondered that, but "Everybody says that 'there exists techniques that deals with...bla...bla...' that never gives a solid example of what there are." made me think OP was asking about estimation. // For the dist'n, isn't the center just the median without further discussion? $\endgroup$
    – BruceET
    Commented Sep 23, 2019 at 15:46
  • $\begingroup$ I, too, found the question to be unclear. $\endgroup$
    – whuber
    Commented Sep 23, 2019 at 15:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.