The mango news makes me feel a bit better.
Why couldn't Giles have shackles like any self-respecting bachelor?
Xander ,'Beneath You'
Natter 39 and Holding
Off-topic discussion. Wanna talk about corsets, duct tape, or physics? This is the place. Detailed discussion of any current-season TV must be whitefonted.
§ ita § - Oct 12, 2005 7:19:11 am PDT #5489 of 10002
Well not canonically, no, but this is transformative fiction.
Katie M - Oct 12, 2005 7:21:25 am PDT #5490 of 10002
I was charmed (albeit somewhat perplexed) by the fannish sensibility of many of the music choices -- it's like the director was trying to vid Canada. --loligo on the Olympic Opening Ceremonies
Hey, is there anyone around who knows statistics? I've got a yes/no survey here with a response rate of about 56%, and I want to know if there's any way to get from that to a margin of error on the percentage of yeses. It seems like there must be, but frankly it's been six years since I had stats and most of it is apparently gone from my brain.
Jesse - Oct 12, 2005 7:27:48 am PDT #5491 of 10002
Sometimes I trip on how happy we could be.
You need the total sample size, Katie, I'm pretty sure. Like here: [link]
Jesse - Oct 12, 2005 7:40:03 am PDT #5492 of 10002
Sometimes I trip on how happy we could be.
In other news, I am now convinced that the Vitamin Water is magic, as I seem to be pretty much all better today. So that's something.
§ ita § - Oct 12, 2005 7:41:46 am PDT #5493 of 10002
Well not canonically, no, but this is transformative fiction.
I've never thought of it as magic, but I have thought of it as crack.
Rick - Oct 12, 2005 7:42:36 am PDT #5494 of 10002
As Jesse says, just by knowing the sample size you can get what pollsters call margin of error, which is a broad concept that applies to all of the numbers in their poll. It's fairly useless from a statistical standpoint, but good enough to tell whether you should take the numbers seriously at all.
If you want a specific confidence interval for your .56 you compute the standard error. You take the square root of:
(.56*(1-.56)/N where N is the sample size.
Then once you have the standard error you multiply it by the factor that will give you a given confidence interval. For instance, + or - 1.96 times the error gives you a 95% confidence interval for the percentage.
Jesse - Oct 12, 2005 7:44:23 am PDT #5495 of 10002
Sometimes I trip on how happy we could be.
I've never thought of it as magic, but I have thought of it as crack.
Possibly that too, and I'll only feel good as long as I keep drinking it.
It is truly pathetic how fast I forgot everything I "learned" in stats class last year.
sarameg - Oct 12, 2005 7:44:54 am PDT #5496 of 10002
Whenever I see vitamin water or smart water or any of the other kinds of minorly doctored waters, I think of miracle-gro. I can't help it. Maybe it was one too many years of grade school "which is the best fertilizer" experiments.
Katie M - Oct 12, 2005 7:45:40 am PDT #5497 of 10002
I was charmed (albeit somewhat perplexed) by the fannish sensibility of many of the music choices -- it's like the director was trying to vid Canada. --loligo on the Olympic Opening Ceremonies
Here's my concern (because, yeah, I started with margin of error and then decided I was on the wrong track). Can I call "the folks who actually answered" (in other words, the 56%) my sample size? Or is my sample size the entire population, all of whom were contacted?
shrift - Oct 12, 2005 7:46:12 am PDT #5498 of 10002
"You can't put a price on the joy of not giving a shit." -Zenkitty
There is an unopened can of chicken noodle soup in the work freezer. It baffles me to the point of obsessive speculation.
Do they know canned soup isn't perishable? Are they trying to see how long it takes to freeze? Is it some kind of bet? Does it taste better frozen? Did somebody forget their coffee this morning? What's the frequency, Kenneth?