Odds ratio of 1.00 while the IV is statistically significant

[Dc731]

I apologize in advance for the stupid question. I am new to statistics.

I got an odds ratio of 1.00 while the independent variable is statistically significant. How do I explain this?

 

[obh]

Can you please give more details?

 

[hlsmith]

Yes, please provide your software output this will clear up many questions. Those two values should align, so something is missing here.

 

[noetsi]

With enough power, millions of cases, you can get statistical significance for nearly anything. That says nothing about the effect size per se other than it is likely the true effect size in the population.

That said, a odds ratio of 1 is essentially a linear regression slope of zero so the results don't make a lot of sense. Your results are telling you there is no relationship between your X and your ability to predict Y.

 

[Dason]

Most likely the value wasn't exactly 1 but the software rounded for the display.

 

[noetsi]

Most likely the value wasn't exactly 1 but the software rounded for the display.

That is probably true, but I would bet 1 is still in the CI.

 

[Dason]

We really haven't been told much of anything so any bet seems premature at this point.

 

[ondansetron]

the lower bound on the CI can still exclude 1 with the point estimate and maybe even CI rounded to 1. This is the same as a p-value of .0000 because the program rounds rather than the value actually being 0.