Mean/ Median

Mean/ Median

I try to find out for samples N (0, 1): n
______prob. [(mean / median) <0]

This is equivalent, of course, to mean and median with different signals.
Because mean and median have Expectations = 0 and owing to symmetry I expected prob. = ½ independently from sample sizes n.
Results, I admit, are much richer and somewhat puzzling (against common sense?) (Program SIMIL).
_____Here are

____n= 5_____0.182
______ 9_____0.191
______11____ 0.196
______15____ 0.198
______25____ 0.201
______35____ 0.203 |
______45____ 0.205 |
______55____ 0.204 | mean = 0.204
______65____ 0.203 |

Data was obtained (for each line) simulating 400´000 n sized samples.

Luis Amaral Afonso (Te Moderator Destroyer)
.

Relevant Pages

  • Re: Mean/ Median
    ... This is equivalent, of course, to mean and median with different signals. ... Results, I admit, are much richer and somewhat puzzling. ... Luis Amaral Afonso (Te Moderator Destroyer) ...
    (sci.stat.math)
  • Re: Mean/ Median
    ... This is equivalent, of course, to mean and median with different signals. ... Results, I admit, are much richer and somewhat puzzling. ... Luis Amaral Afonso (Te Moderator Destroyer) ...
    (sci.stat.math)
  • Re: WHAT Jack Tomsky-John Smith conclude from this?
    ... Luis Amaral Afonso (The Moderator Destroyer) ... For example, when n=5 and the true median is less than -2.012 or greater than +2.012, the probability that Afonso's confidence interval covers the true median is zero. ...
    (sci.stat.math)
Loading