The AP Statistics kiddos are heading into the land of sampling distributions when we go back on Monday. The introduction to this unit is the German Tank problem, which is intended to hammer the concept of unbiased estimation into their heads.
Last year, they were given the general story and told to come up with an unbiased statistical estimate of the number of tanks in the German fleet…platoon…army…whatever. You know, explore and discover “statistical estimators” on their own.
Last year sucked. My fault: I didn’t give them any ideas.
In my usual tradition of overcompensating, I have probably made it too easy for them. But they have a place to start this time.
The only thing that isn’t in the worksheet is the instruction to have the teacher seed the calculators with the mystery number of tanks the Germans have. (I cheat: I have the kiddos from the class right before seed them.) I use 501: a lot more than the Germans had (so they can’t look up the answers on the web), but you can seed the calculators with whatever number you would like. Store them as N to be consistent with the worksheet.
Proof that I sucked at this last year: I printed and cut out 501 little tiny numbers for them to draw out of a bag. The bag is missing. I’m not doing it again.
Yes, they can hack the calculators and just get the answer. I tell ’em at the end of the project. If they don’t want to play along…
If you’ve never heard of the German Tank problem, Wikipedia does a decent job.