Ulaanbaatar, MongoliaThursday, April 29, 2010
We are just starting a statistics unit in 10th grade and a few days ago a good article in the Economist caught my eye. The once-in-decade discussion about whether or not to use statistical sampling to "adjust" the American census results has apparently just come around again.
In conjunction with the article, I used a bowl of M&M's to simulate statistical sampling methods. "Imagine this bowl represents all the people living in the United States. Instead of counting them all, we reach in and grab a sample (with my 1/4 cup measure) and examine the sample closely...and then we do it again...and again...and finally draw conclusions about the overall population of the United States..."
Although I am not the overly patriotic type, I did observe that these particular M&M's were not quite the same quality I am used to back in the States. Yes, they tasted fine, but the candies were not the perfect disks one finds at home. And, say, what happened to the blue M&M's?
But they were perfectly adequate for the experiment. And my obliging students refrained from eating their data until they were allowed to do so...
