Thread: Significant Odds Ratio and Less Significant p-value

Hi-

Why would I get a more significant odds ratio (i.e. 1.4) and a p-value of 0.031 for one variable, and then for another variable, I have an odds ratio of 1.1, and a p-value of 0.013? I would have thought that the more significant the odds ratio, the more significant the p-value will be.

Thanks!

Just to clarify, what I don't understand is why the variable with the highest OR has a less significant p-value, than the variable with the lowest OR.

Any help would be appreciated!

Originally Posted by lem33

Just to clarify, what I don't understand is why the variable with the highest OR has a less significant p-value, than the variable with the lowest OR.

But it doesn't in your example. 1.4 > 1.1, and 0.031 > 0.013. Doesn't it?

Yes, but for p-values, the lower the number the more significant it is. For odds ratios, the higher the number the more significant it is. Thus, for the p-values, 0.013 is more significant than 0.031, and for the odds ratios, 1.4 is more significant than 1.1.

Originally Posted by lem33

Yes, but for p-values, the lower the number the more significant it is. For odds ratios, the higher the number the more significant it is. Thus, for the p-values, 0.013 is more significant than 0.031, and for the odds ratios, 1.4 is more significant than 1.1.

Oh yes I see what you mean. I admit to not having used these terms for years. But don't we have to be careful what we are talking about? As I understand it (having now looked it up!) the "p-value" needs to refer to the probability of the result arising due to pure chance, i.e. if there is no correlation (the null hypothesis). So a low number suggests it probably is not due to chance, i.e. there probably is a correlation. With an odds ratio you don't refer to the null hypothesis…I think.

Are you calculating the p-value and odds ratio yourself or are these numbers you have been given?