Interested in embarking on a citizen science project? One way to learn about the world around you is to take a sample. In fact, this spring the radio and podcast program, Science Friday, encouraged listeners to take samples, which sparked a variety of ideas from listeners.

So how do you go about getting a sample? As Charles Wheelan writes in Naked Statistics, it’s like soup! In all seriousness, best practice is to take a representative sample.  Wheelan explains that:

[t]he key idea is that a properly drawn sample will look like the population from which it is drawn. In terms of intuition, you can envision sampling a pot of soup with a single spoonful. If you’ve stirred the soup adequately, a single spoonful can tell you how the whole pot tastes.

This soup analogy is informative. If a sample is not representative (or the soup is not well-stirred), we cannot make generalizations. Wheelan explains:

[s]ize matters, and bigger is better…it should be intuitive that a larger sample will help smooth away any freak variation. (A bowl of soup will be an even better test than a spoonful.) One crucial caveat is that a bigger sample will not make up for errors in its composition, or “bias.” A bad sample is a bad sample.

From a sample, we may learn something about a population, but we must take care not to overgeneralize. For more on samples and other statistical concepts, Naked Statistics is a useful primer and one of the books that our team has enjoyed.

Source: Wheelan, Charles. Naked Statistics: Stripping the Dread from the Data. New York: W.W. Norton, 2014.

Image: “Pot Steaming Hot Cooking Kitchen Stove Cooker” by Republica, on Pixabay. CC0 Public Domain.