Hello Issam,

1. The obtained OR of 0.181 is significantly different from 1 (equality).

2. Whatever batch was represented by the numerator of the OR has significantly lower odds of being in the target category than does the batch represented by the denominator of the OR (95% CI: 0.042-0.777).

It's up to you to explain how such a difference is meaningful, as well as to what implications that might have for the target audience.

Good luck with your work.

Hello Professor David Morse ,

There are some considerations should be explained in related to sample size and the confidence limits (i.e., wide or narrow).

The question; is there a ratio bet. upper and lower limit to be considered wide/ narrow?

Thank you ...

The ratio that you ask about in your reply always exists but doesn't have to be a rational number.thus wide and ma

Wide or narrow have nothing to do with anything in your question. David Booth

Considering that an OR less than 1 can range from 0 to 1, I'd say that a confidence interval from 0.04 to 0.8 is pretty wide. But I don't know of any established criteria for this assessment.

As context for readers, the following table has an OR and CI close to the example in the question.

Gender Pass Fail

Male 6 14

Female 14 6

The square root of the length of a piece of string is considered wide, at least without you providing any context.