On: Turns out lowly thymus may be saving your life

[contravariant]

I take it a relative risk of 1 is a null result?

If so a confidence interval of 1.02 to 2.2 after arbitrarily excluding some data is not exactly ground breaking.

[nerdponx → contravariant]

It's suggestive of p-hacking. "Look, the bottom of the interval is just above 1! Finally I can publish this."