Data to use for in-class sampling exercises?

Mark Street writes:

I teach a high school (grade 11) statistics class outside the USA and I am always looking for hands-on demonstrations. In fact, last week (the start of our school year here), I did the in-class exercise about “guessing ages of ten pictures” (p. 11-13) from your book “Teaching Statistics – A Bag of Tricks”.

I am interested in using the “candy weighing” demonstration (p. 120-121) to talk about random sampling.

I agree with your advice (p.48, Sec 5.1) that it’s better to have students do sampling from actual populations. I also agree with your suggestion that actual doing personal interviews is not an effective use of time, except for larger projects.

To that end, can you suggest some sources of actual population data that I could use in class? As I am outside the USA (in Thailand), we do not have phonebooks here. Certainly, this must soon be a problem for people in the USA with phonebooks going the way of the 8-track tape. I even looked online for digital images of phonebook pages, but there were surprisingly few to choose from.

My reply: Random words from the newspaper, perhaps. They could estimate how many are nouns, verbs, etc?

5 thoughts on “Data to use for in-class sampling exercises?

  1. I’m not sure if Thai students will be interested in this, but instead of phone books you could look at old US Census records. 72 years after each Census they become publicly available, so for example you can look at scans of the original enumeration forms from 1940:
    http://1940census.archives.gov/

    You’ll see not just names but age, sex, education, occupation, and other more interesting covariates than you can get from a phone book.

  2. We had a similar exercise, sampling words at random from a book, in one of my grad school classes. Having lived in Thailand and being able to speak some Thai, I chose a Thai to English dictionary and estimated the total number of words in the dictionary that I knew from the pages that I sampled at random.

    Perhaps the students might find that an interesting twist?

  3. When I was a TA teaching a class on sampling methods, a group of students did a project in which they sampled wikipedia pages using the “random article” link. I thought it was quite clever of them. If your students have internet access, something along these lines might work well.

  4. Give each student a penny. Everyone flips the penny, and if it lands on heads, they raise their hand. Count up the number of heads and record it on the board. Repeat this 10 or so times so that people will start to get the idea of sampling variation.

Comments are closed.