I’ve noticed most U.S. polls query around 1,000 people and sometimes even less. Somehow that qualifies for headlines like “Americans say…” or “Most Americans…” How is it acceptable that 0.0002% of the population is accepted as representative?
In: Other
Much the same way a chef can figure out what’s wrong with the soup by tasting a spoon-full, rather than drinking the whole pot. Or, put another way, there’s a joke my professor used to tell: “If you don’t believe in random sampling, next time you need a blood test, tell the doctor to take it all.” Opinion polling can do the same thing: reliably determine a group’s opinion on an issue (the soup) from a small sample of the population (the spoon-full).
Imagine that you’re in charge of planning a party for your neighborhood, and you’re trying to decide whether to buy hamburgers or hot dogs. You expect 100 people will show up, so you knock on ten random doors, and ask. The first ten answers come back: three people want hot dogs, seven people want hamburgers.
Mathematically, we can draw conclusions about all 100 people from those ten. It’s *possible* you got really lucky/unlucky and found the only seven people who like hamburgers in the whole neighborhood, but it’s extremely unlikely. In fact, there are statistical equations that can tell us exactly how unlikely it is.
Polling companies use those equations to figure out how many people they should survey in order to get a good estimate. We know that 95% of the time, asking 1000 randomly selected people from the entire country whether they prefer hot dogs or hamburgers, the result will be within +/- 4% of the answer for the whole country.
Latest Answers