Ways to Choose a Sample, and Which Ones Are Fair

0:00Do a school's 300 students want a longer lunch? This whole group is the population.

0:07Asking everyone is slow, so we ask a sample, a smaller part: 30 students.

0:14A sample is fair if every student has the same chance of being chosen. If not, it is biased.

0:21One: a simple random sample. Put all 300 names in a hat and draw 30.

0:28Each student has a 30 in 300 chance, which is 1 in 10. Fair.

0:34Two: a systematic sample. Pick a random start from 1 to 10, then take every 10th name.

0:42Start at 4: take 4, 14, 24, up to 294. That is 30 names, 1 in 10 each. Fair.

0:53Three: a stratified sample. Split the school into its three grades: 120, 90 and 90 students.

1:02Draw one tenth of each at random: 12, 9 and 9, making 30. Fair.

1:09Four: a convenience sample. Ask the 30 students who happen to be nearby.

1:15They are often friends in one grade, and the other 270 have no chance. Biased.

1:23Five: a voluntary sample. Post an online poll, and whoever chooses to answer is the sample.

1:30Students with strong views answer more, and some never see the poll. Biased.

1:37A common mistake is thinking a poll speaks for everyone. 27 yes out of 30 replies is 90 percent of the replies, not of the school.

1:48Fair samples give everyone the same chance: simple random, systematic and stratified. Convenience and voluntary are biased.