t-tests, calculator skills, and baseball


My AP stats kids are really not wanting to learn the calculator skills. It is killing me! They want to do it all by hand and then type in the math in the normal screen instead of using “stats” and “ttest” instead. KILLING ME!

So, to force the issue, and give them something interesting to work on I turned to the intertubes and my Google Reader for inspiration. As I was reading this morning, I came across this article.  Hmm, that sounds worth while. I am not a huge baseball fan, but I do like the stats!

Hmm, now the article used the data from every player, and I don’t want my learners needing to use a computer. They need a data set that is manageable on the calculators.  So I poke around on one of provided links and find this link that has team data

That looks promising. But I better look at it myself before I give it to them, right? I take the team data and graph it using JMP and find out it is no-where near normal. That is good. It means that the learners will run into problems doing the t-test if they actually LOOK at the data as I keep asking them to. If they don’t graph the data, they will get a happy answer, and it will be wrong. Nice.

I did the t-tests, and no matter how you slice it, none of them are significant. The means are all around .26 (hmm, regression to the mean comes up here as well.)

Then I think I will give them Barry Bond’s highest batting average (2002, .443 found here) and ask the groups to decide what data they need (and then give it to them) to determine if Barry’s batting average is significant.

So, the instructions, purposefully vague, will be:

Is the National League or American League’s batting averages significantly different in 2000? How about 2010?

Are the averages significantly different between 2000 and 2010?

What information do you need to decide if Barry Bond’s batting average is significantly higher than the leagues, and then is it significant?

The learners, working in groups, will need to write an appropriate Ho, Ha, do all conditions checks, and then, if appropriate, do the appropriate tests.

The conditions checks will fail on the nearly normal condition, but I will ask them to continue anyway and explain the problem in their conclusion.

Any feedback on this project?


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.