What if a confidence interval ends at 1.0?

[SEMAN OSMAN]

Dear all users,

I am working the analysis of multilevel logistic regression for association between maternal exposure to Any form of IPV and use of Health Facility Delivery and found the following results: (AOR = 0.60, (95% CI:0.36 - 1.00), p= 0.051).
My question is that the confidence interval includes 1.00 then can I say the result is not statistically significant despite the P-value is significant at 95% CI; because it includes the null?

[Felix Bittmann]

Well apparently the CI and the p-value are telling you the same since p is also larger than 5%, so the p-value in your example is not significant.

[Carlo Lazzaro]

Seman:
set aside any consideration about the usefulness of the p-value (

https://www.amstat.org/asa/files/pdfs/p-valuestatement.pdf ;

https://www.tandfonline.com/doi/full...5.2016.1154108) your p-value exceeds 0.05 and this is consistent with the 95% CI including 1.
Therefore, you can say that result is not statistically significant at 0.05 arbitrary threshold.

[SEMAN OSMAN]

Dear users thank you so much for your professional responses.

[John Mullahy]

To avoid confusion you might note somewhere that the upper limit on the .95 CI is 1.002 (as some readers might wonder whether 1.00 is a number less than 1 rounded up or a number greater than 1 rounded down).