Researchers in Management and Strategy worry a lot about bias – statistical bias. In case you’re not such an academic researcher, let me briefly explain.
Suppose you want to find out how many members of a rugby club have their nipples pierced (to pick a random example). The problem is, the club has 200 members and you don’t want to ask them all to take their shirts off. Therefore, you select a sample of 20 of them guys and ask them to bare their chests. After some friendly bantering they agree, and then it appears that no fewer than 15 of them have their nipples pierced, so you conclude that the majority of players in the club likely have undergone the slightly painful (or so I am told) aesthetic enhancement.
The problem is, there is a chance that you’re wrong. There is a chance that due to sheer coincidence you happened to select 15 pierced pairs of nipples where among the full set of 200 members they are very much the minority. For example, if in reality out of the 200 rugby blokes only 30 have their nipples pierced, due to sheer chance you could happen to pick 15 of them in your sample of 20, and your conclusion that “the majority of players in this club has them” is wrong.
Read the whole story: Forbes