Random Stuff XLV

[6_bleen_7]

The results of the 2008 Silly Straw Poll of Most Favored Feline Water Sources are now in. Cats from all corners of the household were asked to rate the available venues for liquid refreshment; what follows is a summary of the results.

In a stunning upset, Droplets Licked from Dad’s Shins with Sandpaper Tongues as He Emerges from the Shower was

Comments

[chillyrodent]

Just for the heck of it, I polled our flock. The prosaic favorite is the water fountain made for cats, but we have one very strong vote for stale/dirty water in a glass. A puddle would also be fine, as long as it was truly stagnant, and not just fresh water stopping to rest.

Second place went to toilet water.

I'll bet you'd have astronomers guffawing at your poster, because they'd know just what they were seeing. A collision! Friends ... drive drunk - chortle! Me? I was there with Kathy.

[6_bleen_7]

I'm delighted that you think that the PSA may, at least in theory, be funny to someone.

That your intended water source is the favorite astonishes me to no end.

Also, note that even though the sample size in our poll was 2, the results are perfectly accurate because we sampled the entire population. When surveys report a sampling error, they assume that they've taken a representative sample from a much larger population of interest. In fact, the larger proportion of the population sampled, the more accurate the estimate of whatever is being polled.

[chillyrodent]

I know. It's like they don't suspect we want them to drink from it.

Nice epidemiology.

[6_bleen_7]

On the contrary, I think they know we do want them to drink from it, and so they'll avoid it like it was plutonium just to be difficult (i.e., to remind us, in their usual subtle manner, just who's boss).

Thanks-I was thinking even more broadly about the statistical problem of sampling from a finite population. If n is your sample size and N is the population size from which you draw your sample, you need to reduce the standard error of your estimate by a factor of sqrt(1 - (n - 1)/(N - 1)).