I want to calculate the z-score for a distribution for which I know the odds-ratio, and the p-value. I also know the upper and lower 95% confidence limits for the odds-ratio. An example with two-data points is shown here.

OR       OR_95    OR_95U   p-value
0.997804 0.970573 1.025798 0.876215
1.039562 1.010116 1.069866 0.00815
  • 2
    $\begingroup$ "Z score" of what? (By "distribution" surely you mean data, so what you have is an estimated odds ratio (and its p-value).) And what is the second row of statistics you show? A second example? $\endgroup$ – whuber Sep 6 '13 at 18:43

Odds ratios standard errors, CIs, etc are calculated on the log of the odds.

Here's some R code. First we'll take the log of the odds:

OR <- 0.997804
ORL95 <- 0.970573

logor <- log(OR)
logorl95 <- log(ORL95)

Then calculate the difference between the odds ration and the lower (or upper, it doesn't matter, they're the same) confidence limit.

diff <- logor - logorl95

The confidence limit is 1.96 * se, so:

seor <- diff / 1.96

z is estimate / se.

z <- logor / seor

Let's calculate p, just to make sure we didn't make a foolish mistake along the way:

(1 - pnorm(abs(z))) * 2

Here are the important bits of the output:

> z 
[1] -0.155723

> (1 - pnorm(abs(z))) * 2
[1] 0.8762514
  • 1
    $\begingroup$ In case you want more detail, this is covered in Agresti's An Introduction to Categorical Data Analysis. $\endgroup$ – Sycorax says Reinstate Monica Sep 6 '13 at 19:06
  • $\begingroup$ Thanks! Does it matter if the odds-ratios are greater than or less than 1? For the second data point I know the Z is 2.185, however when I try to follow the directions you stated above, the answers do not seem to match. Am I misunderstanding something? $\endgroup$ – sanyal Sep 6 '13 at 19:41
  • $\begingroup$ I'm not sure if you're misunderstanding, but that Z is not correct. If Z is 2.185 then p is not 0.008. A z of 2.185 gives p of 0.029. (Zs of tour 2, 2.5 and 3.3 give ps of tour 0.05, 0.01, 0.001.) $\endgroup$ – Jeremy Miles Sep 7 '13 at 0:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.