From: Damon Jansen (dkjansen@yahoo.com)
Date: Sat Dec 09 2000 - 17:13:36 PST
Message-ID: <20001210011336.8671.qmail@web108.yahoomail.com> Date: Sat, 9 Dec 2000 17:13:36 -0800 (PST) From: Damon Jansen <dkjansen@yahoo.com> Subject: The Election, Precision and Significant Figures
The issue of the precision of the vote count is
something I have seen addressed in a few newspaper
(the San Francisco Chronicle) and radio (NPR) stories
in the past couple weeks. I have heard the election
referred to as a "statistical tie" and I read a
citation that one would have to have fewer than one
error in 7,000 votes to be sure you got the winner
correct. So, I decided to hold a ballot-counting
session in class.
We looked up on the ABC news site and got numbers
which were current at some point. (Though they do not
agree with the numbers I've been seeing cited the past
two days (today is Saturday 9 December), they'll do.)
Bush: 2 910 457
Gore: 2 910 029
Difference: 438
Total votes: 5 820 486
If you take the total divided by the difference
(5,820,486 / 438)= 13289, you realize that if the
counters made only one mistake every 13,000 votes, you
could still not be "sure" of the outcome. I think the
story I read considered (which said 1 in 7000)only the
votes for one candidate. (This is a pretty
non-statistical analysis--it would be interesting to
apply some better stats to look at confidence levels.)
To illustrate how tough this would be to do, I cut up
pieces of paper so that we had 500 ballots and then
got 3 people to do the voting, 2 people to number the
ballots (though this step is not necessary), one
person as the runner to bring ballots between groups,
and everyone else was a counter. The counters lined
their desks next to each other so they could just pass
the ballots desk-to-desk. The voters simply put a "G"
or a "B" on each ballot.
I did this in five classes throughout the day. I
figured that there would be variations between
students, but I was surprised at just how poorly they
did. In a given class, there was at least a +/- 10
vote spread between most students, and there were
always some outliers who were +/- 30. A very useful
question came up--students asked me what the "real"
answer was, and I could tell them that I had not
counted them, and even if I had, I would just be one
more data point.
Paul D always says "counting is difficult," and I
think students believed it at the end of the day.
I was really surprised at how much fun students had
voting and counting all the ballots. I thought their
interest would wane after 50 or 100, but they kept at
it. I highly suggest this activity, especially if
you're looking for something to do the last day of
school before the holiday!
--Damon Jansen
__________________________________________________
Do You Yahoo!?
Yahoo! Shopping - Thousands of Stores. Millions of Products.
http://shopping.yahoo.com/
This archive was generated by hypermail 2b29 : Tue Oct 16 2001 - 12:22:05 PDT