“So the number you see is a worthless number”
This is regarding the MMR autism study.
Why do you say so?
It has a very large p-value (specifically the low-risk group).
has a VERY wide confidence interval of 0.76-1.54 which means that the true value can be anywhere in that range.
This means if this study were to be repeated 100 times, the true estimate would fall between 0.76 and 1.54.
So the one fractional difference you found between that one group, is extremely insignificant.
The point was, regardless of which way you slice the data, the numbers are the same other than the ones that are at higher risk, which is older siblings (genetics).