Re: Normal distribution of ratio of means

> Hello all:
>
> Assuming we have (the estimates of) two means, each
> calculated from the
> same sample. Is the ratio of these two means normally
> distributed (so
> that one can determine the 99%-confidence intervall
> in the following
> way: ratio +/- (2.58 multiplied by the standard error
> of the ratio))?
>
> Thanks,
>
> Fritz


There's an exact result for the confidence interval of the ratio of two means from a bivariate normal, even in the presence of correlation. It's called Fieller's Theorem.

Jack
.

Relevant Pages

  • Re: About dyeing...
    ... I did that Jack, thanks a lot! ... These two pictures are of Black Walnut ... Black Aniline and the other with a 4-1 ratio of red Aniline dye. ...
    (rec.crafts.woodturning)
  • Re: Distribution of ratio of two normal distributions?
    ... The exact distribution of the ratio of two normal ... variables for arbitrary mean, variance, and correlation is a ... well-solved problem - albeit one with an ugly answer. ...
    (sci.stat.math)
  • Re: It would be nice if Google would..
    ... and that space is a correlation associated with two bodies (A cross- ... and that the ratio of a cross-correlation and an auto-correlation ... And your ratio would be the cotangent, ...
    (sci.physics)
  • Re: Historic N Sea temperatures
    ... warmer than that "February" graphic providing some correlation with ... Jack ... Prev by Date: ...
    (uk.sci.weather)
  • Statistics for comp.lang.lisp
    ... - Original Content Rating is the ratio of the original content volume ... to the total body volume. ... - Please send all comments to Christopher Browne ... Prev by Date: ...
    (comp.lang.lisp)