0
$\begingroup$

So I've been working on a logistic regression model, and all has been going well for the most part. However, I just started analyzing the results and trends of the model, and one particular odds ratio is giving me trouble. The odds ratios for the model (with four independent variables) are 0.9999999628643, 1.00000072117459, 0.0393990796298753, and 23974346.5342214.

Now this is obviously a huge problem. Most of them are fine, but then there's that one odds ratio in the twenty millions. While a larger odds ratio makes sense, as that's the variable that's probably going to have the largest impact on the output, I didn't anticipate having something that high. Is this unusable? Did I do something wrong? Thanks for any help you can offer!

Edit

The CI for the odds ratios are [0.999998601016136, 1.00000132471432], [0.999997388410105, 1.00000405395019], [1.02226897236851E-06, 1518.4726501918], and [0.810502926670688, 709151408131148].

Improve this question
$\endgroup$
10
  • $\begingroup$ The point estimates of ORs are not good. Always provide at least 95% Confidence Intervals - they also have drawbacks, but they are obviously better than point estimates. $\endgroup$
    – German Demidov
    Commented Nov 10, 2019 at 15:00
  • $\begingroup$ Thanks, I fixed my question to include the CIs. $\endgroup$
    – James Connor
    Commented Nov 10, 2019 at 15:21
  • 1
    $\begingroup$ I think you already see that all of your CIs are including 1, no matter how huge the actual point estimate is. "Is this unusable?" - kind of, you may say that "there is no statistically significant difference in ORs" (you should think about the correct formulation of the conclusion, I provided a rough idea) $\endgroup$
    – German Demidov
    Commented Nov 10, 2019 at 15:23
  • $\begingroup$ Huh. So basically, those CIs mean that none of the four inputs have a statistically significant effect on the output? $\endgroup$
    – James Connor
    Commented Nov 10, 2019 at 16:04
  • $\begingroup$ It's quite likely that the independent variables are collinear. Have you checked? $\endgroup$
    – Peter Flom
    Commented Nov 10, 2019 at 16:09

0

Reset to default

Browse other questions tagged or ask your own question.