Which overlaps with the CI on other COVID variants. This is a small sample study, and the media and governments are probably too quick to take the centre of the estimate. Sampling variation could explain the outcome as much as the South African variant.
I would assume the very wide CI mentioned above means that, since it includes 0, we can’t say there is a statistically significant effectiveness (it is not significantly different from 0% effectiveness).
> we can’t say there is a statistically significant effectiveness
Agreed, although to clarify and be more precise, it also doesn't prove it's not effective. It effectively says that the sample was too small to tell anything.
> (it is not significantly different from 0% effectiveness).
This statement isn't true - it means that they are 95% sure the range of possible values is between -50 and +60, but it does not mean that the average of those two values is correct. 0% effectiveness is possible, but was not shown in this study.
You’re right, semantics are everything. I left implicit the “in this study” part. As always, failing to reject the null doesn’t mean the null is true :)
This is not really correct. CI means we are 95% sure that the true result falls somewhere within the range of -50 to 60. What this really means is that the study lacks statistical power to say anything interesting at all about vaccine effectiveness.
Right. I phrased it a bit differently below. By “we can’t say” I meant that the study/researchers aren’t able to claim that the vaccine has non-zero effectiveness, i.e., they haven’t rejected the null.
