I would like to compute a 99% confidence interval for the odds ratio based on the likelihood ratio statistic. Here is what I have so far; the output provides a confidence interval for the odds ratio but is this based on the LR statistic?
data vaccine2;
input age reaction count;
cards;
65 5 0
65 1 3752
42 5 4
42 1 11301
;
run;
proc freq;
weight count;
tables age*reaction/ cmh chisq BINOMIAL(CL=LIKELIHOODRATIO) alpha=0.01;
run;