posted by ![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
emperor at 11:12am on 03/11/2003
emperor at 11:12am on 03/11/2003Not a bad weekend in the end. The TS turned up, and the various meetings were good and productive (even if not strictly quorate). I managed to feed spag bol followed by rice pudding to 11 people without killing anyone, which was quite an achievement. It even tasted plausible!
Didn't make it to the party on Saturday, though. Neither did my housemates, due to lurg or alcohol (delete as appropriate :) I have a feeling I may be about to catch the lurg-of-doom, too :-/
Plan for this week: ballroom tonight, CTS tommorow (shock!), evening in on Wednesday, RnR/Round on thursday, ?ringing on Friday, parents visiting/parties Saturday.
Finally got questionnaire approved by supervisor today after a bit more back-and-forth. Hoping to get a new chair soon (this one is too low, and so bad for RSI). No sign of Cerian today.
Mostly back to reading for now - the Rhino stuff is too depressing. Any statisticians in the audience?
Didn't make it to the party on Saturday, though. Neither did my housemates, due to lurg or alcohol (delete as appropriate :) I have a feeling I may be about to catch the lurg-of-doom, too :-/
Plan for this week: ballroom tonight, CTS tommorow (shock!), evening in on Wednesday, RnR/Round on thursday, ?ringing on Friday, parents visiting/parties Saturday.
Finally got questionnaire approved by supervisor today after a bit more back-and-forth. Hoping to get a new chair soon (this one is too low, and so bad for RSI). No sign of Cerian today.
Mostly back to reading for now - the Rhino stuff is too depressing. Any statisticians in the audience?
okay
(no subject)
But, fireworks followed by mulled wine and toffee apples..?
(no subject)
(no subject)
(no subject)
(no subject)
First, the simplified form: I have a series of observations. How do I assess whether their mean is significant?
The actualy problem: I have 3 rhinos, which have each been anaesthetised a fair few times. In each case, the time to onset of anaesthesia has been recorded in a proportion of cases. How do I tell whether the values I have tell me anything useful? (i.e. mean time of onset might be 5 minutes, but is this significant?)
(no subject)
If I remember correctly (and I never did these at A Level, for no reason I'm aware of) you normalise to a variance of 1 (obtaining global variance estimates from sample variance is a case of multiplying by n/(n-1), IIRC), and a mean of 0, and then pull out a lookup table, which will give the boundary values beyond which that mean is significant at the (say) 5% level.
I'm assuming here that you're adopting the standard approach to hypothesis testing, i.e. you have a control sample, and thus a 'typical' mean, and you are trying to verify the assertion that a particular sample is (say() greater than the mean, and that there is less than a 5% chance this could have come about randomly.
Hmmm... That's almost certainly useless to you, but I think the critical point is to look up a t-test in a stats textbook of whatever level you have the background for.
IIRC, you're trying to verify the assertion that the onset time increases with repeated applications. Were I attempting to do this without doing some serious reading beforehand, I'd compare likelihoods of a fixed mean, and one which decreased with time. Still using a t-test, but 'normalise to mean 0' takes on a slightly different meaning.
It's really quite hard to explain maths by email, so I can't imagine this makes much sense. Still, hope it helps a little.
(no subject)