Re: how to find z-values for a "stick-out" statistic?

On Mar 15, 9:01 pm, John Nagelson <johnnagel...@xxxxxxxxxxx> wrote:
How to calculate the z-values for results of the following
experiment?

Run 100 trials for which the output can be A, B, C, or D.
A, B, C, and D are hypothesised to be equally probable,
so the expected scores for each outcome is 25.

Define the "stick-out" as the amount by which the HIGHEST SCORE
(H) exceeds the NEXT HIGHEST SCORE (N).

Obviously the expected stick-out is not zero, but I'm not sure
what it is.

I want to define a statistic for stick-out in two different ways.
First, define S = H-N. This can vary from 0 to 100.
Second, define S' = (H-N)/N. This can vary from 0 to infinity.

Is there a way to use chi-squared to get z-values for observed
values of S and S'? Or does something else need to be done?

I'd still be grateful for any help with this, even (especially) if
there's a simple answer!

Many thanks!

John

.

Relevant Pages

  • how to find z-values for a "stick-out" statistic?
    ... How to calculate the z-values for results of the following experiment? ... so the expected scores for each outcome is 25. ... Define the "stick-out" as the amount by which the HIGHEST SCORE ... Obviously the expected stick-out is not zero, but I'm not sure what it ...
    (sci.math)
  • Re: how to find z-values for a "stick-out" statistic?
    ... Define the "stick-out" as the amount by which the HIGHEST SCORE ... First, define S = H-N. ... As a first approximation, you can put 10,000 values of each in an array, ...
    (sci.math)