3 ms·
... the 95% upper and lower confidence limits of the non-detection varied between -2 and 5 parts per billion I can't imagine the degree of confusion which led
by cperciva 14y ago
... the 95% upper and lower confidence limits of the non-detection varied between -2 and 5 parts per billion
I can't imagine the degree of confusion which led someone to say that it's possible for Mars' atmosphere to contain a negative concentration of methane.
- zitterbewegung 14y agoStatistically your confidence interval can go negative but practically a negative concentration doesn't exist.
- cperciva 14y agoStatistically your confidence interval can go negative No, it really can't. If it looks like it is, you're using a symmetric distribution function (probably a normal distribution) in a context where its assumptions do not hold.
- zerostar07 14y agoI think they said variation, not confidence interval