Tuesday, August 18, 2009

And There You Are

I'm going to take a cue from the Velvet Blog and repost some favorites. Repost from 2007...

Every once in a while I like to pose this question to my boys: "One third of accidents take place within a mile of home. Why?"

And my big one says "Because you're running late and you're in a hurry." And my little one says "Because you have to pee."

Good answers. But the one I was looking for is: "Every journey begins and ends at home."

And I meant to say that the fact is just an example of probabilities. That's just the statistics of being in a certain place a lot. And it turns out the insurance company article I got the fact from doesn't mention that conclusion (which I find strange for a company with actual actuaries on staff). They say it's because home is where you let down your guard.

Mashup I am not going to make: "Morning Train" and "Ruby Tuesday"

Touche Cliche: You can't go home again if you're too scared to tell them you wrecked the car.


Sue said...

I love your explanation, and will have to remember that for Statistics next term. It reminds me of an example from the textbook about a reporter who was sent out by his boss during a snow storm to collect information on traffic accidents, who was really impressed with what seemed like a large number of traffic accidents, until he thought to ask them how that compared to an average work day (the answer was dramatically less!).

Qaro said...

Thanks! I also need to find a good explanation for the bell curve and standard deviation.

ChefNick said...

Two words . . . "Freak" "Accident"

I pride myself on looking out for those.

So far, so good.

Sue said...

I think I have a fairly good explanation that I give my students of the standard deviation, too lengthy to post here, I'll try to remember to e-mail with attachments when I'm at school one day. But what it boils down to is making an analogy to finding the "average" deviance or distance all the measurements are from the mean -- but showing them what happens if we try the normal way to find an average (adding all the distances up and dividing by the number of measures), i.e., all the deviances/distances add up to zero and you can't divide into zero, etc., etc.

I use examples from nature to explain a normal curve, like heights, size of leaves, etc. We also some times -- when the class is large enough -- have everyone draw 8 twenty-five person samples from 99 person population, and figure the sample mean of a known parameter (like how many people are Democrats), and then everybody reports their findings and we create a histogram, and everyone is amazed to see this more or less normal curve appear. This works even with sample sizes of 25 (and therefore a margin of error of about 20%!).