On 09-Jul-01 Beasley , Brent wrote:
> I'm having trouble locating an equation to calculate 95% confidence
> intervals for a likelihood ratio. I've tried several websites and
> texts.
> Any help is appreciated.
As an approximation, increasingly good for increasing amount of data,
2*log( Max Likelihood ) - 2*log( Likelihood at true parameters )
has a chi-squared distribution with d.f. equal to the number
of parameters (with similar results if "true parameters" have
been estimated within constraints, e.g. "independence"
constraint for 2x2 table).
In many cases the "approximation" is exact.
You can go on to variants of the above.
Without knowing more about your specific question, I can't be
more definite or helpful. If you want to follow it up, drop
me a line.
Best wishes,
Ted.
--------------------------------------------------------------------
E-Mail: (Ted Harding) <[log in to unmask]>
Fax-to-email: +44 (0)870 167 1972
Date: 10-Jul-01 Time: 03:09:22
------------------------------ XFMail ------------------------------
|