I am reading a statistics book and there is example that odds ratio8 and CI 1-15 . Here the odds ration is not 1 but CI includes 1 . So, do we consider it statistically significant or not?
No. But almost, because the lower confidence limit is almost above 1.
Thank you but I still can't get it . What matters most: value of CI or odds ratio?
Briefly, the odds ratio is the "point estimate" (i.e., your "best guess") for an association.
The CI of your point estimate (the odds ratio) indicates the level of uncertainty around the point estimate. CI are used because the data you are analysing include only a small part of the population. With CI we can infer that the true population effect is between the upper and lower CI.
In your scenario it seems that the true population effect is between 1 (i.e., no effect) and 15 (i.e., a quite large effect). This is why the odds ratio, and therefore the association, is considered "not statistically significant" even if is quite big (8 in your case).
- Badges
- Science topic
- Similar topics
- Books