For example, Times reporter Donald McNeil reports that
In the data itself, the real margin of success was razor-thin: 23 Thais out of 16,395.Got that? The 1/3rd protect meme that was propagated last week came from the fact that 23 fewer Thais out of more than 16,000 participants who would have been expected to get HIV, did not.That is, three years after getting the vaccine or a placebo, 74 in the placebo arm of the trial became infected while only 51 in the vaccine arm did.
Bloggers with a taste for biostatistics — and one rival AIDS vaccine specialist who declined to be quoted — said it would take only a handful more infected Thais in the vaccine column to shift the results from “statistically significant” to meaningless. Even one more would have weakened the data enough to make headlines saying “One Quarter Protected” more likely, given the way journalists round off numbers.
No comments:
Post a Comment