ABSTRACT
The odds ratio in the population will be denoted by ψand it can be estimated from the observed frequencies in the table as the ratio of the estimated odds of category 1 to category 2 of variable 1 in category 1 of variable 2 to the corresponding estimated odds in category 2 of variable 2, i.e.,
bc
Estimator of the variance of log(ψˆ ): var(log ) = / / / /ˆ ψˆ 1 1 1 1a b c d+ + + Approximate 95% condence interval for log(ψ):
log var )( log ˆ ) . ˆ ( ˆψ ψ± ×1 96
If the limits of the condence interval for log(ψ) obtained in this way are ψ ψL U, then the corresponding condence for ψ is exp(ψL ) and exp(ψU).
