Reminder: Assignment: pp. 581-: #1, 3, 4, 8, 12. These will be
attached to your upcoming exam, coming up 11/3, and constitute 10% of
the exam.
Your exam will cover the material since the previous exam:
infinity,
dimension,
geometry, and
statistics.
You might want to check out this information I worked up about the sampling methods using coins that we have looked at (and that you'll look at today).
We've an election before us: who will win?
Today we're going to introduce and use another technique for answering embarrassing questions, the two-coin-toss method
It's similar to the one-toss method -- why is it any better?
This method provides some anonymity, some cover, which we hope will result in honest answers.
This method involves no data loss!
Curiously enough lying turns junk into
gold. Before, we lost 50% of the data because we gave
an uninformative answer 50% of the time. Now we give
either the honest truth or an honest lie!;)
Randomization still introduces noise into the
results.
Let's first use the new, improved method to determine the rate of embarrassing pattern maleness in this class.
Toss two times;
tell the truth if you get a tail;
lie -- tell the opposite of the truth -- if you get two heads.
Now let's use the new, improved method to see who's going to win the upcoming presidential election, provided this class is a good representative sample of the population of voters:
First off, is this class a "good representative sample of the population of voters"?
Let's assume that there are only two candidates running, McCain and Obama.
To further provide anonymity, let's have you write the answer -- the result you should give, depending on your pair of coin tosses -- on a scrap of paper. Write M or O, depending on what the coin and your own intent tell you to do.
Let's predict who's going to win, and by how much.
Website maintained by Andy Long.
Comments appreciated.