Tuesday, September 2, 2008

Becoming a Great Teacher: Part 4

(Readers new to this series should start at the beginning)

The most important quality of a great teacher is to give dynamic lectures. One way to give dynamic lectures is to show short, entertaining tv or movie clips. Kim Anderson, who teaches sales in our department, is king in this area. I believe almost every lecture of Kim's contains a video of some sort. The funnier the video, the better. Of course, it should relate to the lecture objectives.

While not claiming to be one of these great teachers, every now and then I do give a damn good lecture. Here is one I plan to give today.

In my regression class I like to make sure students understand that regressions measure correlation, not causation. This is achieved by the following.

  • First, I show them a clip from The Simpsons, episode Much Apu About Nothing (Seventh Season). In this clip Homer commits the fallacy of confusing correlation with causation, and ends up purchasing a rock that keeps tigers away. While I cannot post the video due to copyright concerns, an interesting article containing the relevant dialogue can be found here.
  • After discussing the fact that the rock didn't really keep away tigers, I remind them of the regression we used for explaining why some people hit softballs further than others: Predicted Hitting Distance = 144.109 – 66.777(female) + 3.570(years experience) (see my previous post on collecting these data).
  • The data indicates that obtaining one more year of experience playing on a baseball or softball team will increase one's hitting distance by about 3.57 feet. But will it? The data say that class members with more experience playing on a team hit further, but what really caused them to hit further?
  • It certainly is the case that the more you play on a team the further you should be able to hit, but it is not the case that increasing experience by one year will increase the average hitting distance by 3.57 feet.
  • The reason is that people self-select into ball teams. At this point I identify a guy in the class who is obviously strong and contains more experience than me (I, for those who do not know, am very skinny at only 155 lbs). I ask: who looks like they can hit further? Everyone says the other guy. I then ask: who is more likely to want to play ball and show off their athleticism? Everyone says the other guy. Thus, I illustrate that the correlation between experience and hitting distance may not be totally causation. In part, both experience and distance are caused by natural athleticism.
  • Here comes another fun part. I ask the students to use the data here (look under Data Available for Download in the left menu) to estimate the relationship between being religious and being happy. Indeed, the data suggest that the more religious one is, the more happy one tends to be. But does religion cause happiness, or do happy people tend to be more religious, perhaps because they have more to be thankful for?
  • To illustrate this latter point, I ask if any student is not religious. At least a few are certain to raise their hand. I then inform them about this church called the Church of the Flying Spaghetti Monster. It sounds silly, but it is arguably as real as any other church. I then ask the person, and the class, if they expect to be made happier by converting to this Church of the Flying Spaghetti Monster. A good laugh and discussion usually follows. Students love finding out about this church, especially the fact that its heaven has a beer volcano.
  • Also, at the Church's website is the prophet Bobby Henderson's claim that global warming is caused by a reduction of pirates. To prove this, he graphs the number of pirates against CO2 emisssions - sure enough, the number of pirates is negatively correlated with C02 emissions. This last example really drives home the point of not confusing correlation causation!