Do Not Marvel at Your Existence: Self-Selection Bias
Traveling from Philadelphia up to New York, I (Rolf Dobelli) got stuck in a traffic jam. “Why is it always me?” I groaned. Glancing to the opposite side of the road, I saw carefree southbound drivers racing past with enviable speed. As I spent the next hour crawling forward at a snail’s pace, and started to grow restless from braking and accelerating, I asked myself whether I really was especially unlucky. Do I always pick the worst lines at the bank, post office, and grocery store? Or do I just think I do?
Suppose that, on this highway, a traffic jam develops 10 percent of the time. The probability that I will get stuck in a jam on a particular day is not greater than the probability that one will occur. However, the likelihood that I will get stuck at a certain point in my journey is greater than 10 percent. The reason: Because I can only crawl forward when in a traffic jam, I spend a disproportionate amount of time in this state. In addition, if the traffic is zooming along, the prospect never crosses my mind. But the moment it arises and I am stuck, I notice it.
The same applies to the lines at bank counters or traffic lights: Let’s say the route between point A and point B has ten traffic lights. On average, one out of the ten will always be red, and the others green. However, you may spend more than 10 percent of your total travel time waiting at a red light. If this doesn’t seem right, imagine that you are traveling at near the speed of light. In this case, you would spend 99.99 percent (not 10 percent) of your total journey time waiting and cursing in front of red traffic lights.
Whenever we complain about bad luck, we must be wary of the so-called self-selection bias. My male friends often gripe about there being too few women in their companies, and my female friends groan that theirs have too few men. This has nothing to do with bad luck: The grumblers form part of the sample. The probability is high that a man will work in a mostly male industry. Ditto for women. On a grander scale: If you live in a country with a large proportion of men or women (such as China or Russia, respectively), you are likely to form part of the bigger group and accordingly feel hard done by. In elections, it is most probable that you will choose the largest party. In voting, it is most likely that your vote corresponds with the winning majority.
The self-selection bias is pervasive. Marketers sometimes stumble into the trap in this way: To analyze how much customers value their newsletter, they send out a questionnaire. Unfortunately, this reaches only one group: current subscribers who are clearly satisfied, have time to respond, and have not canceled their subscriptions. The others make up no part of the sample. Result: The poll is worthless.
Not too long ago, a rather maudlin friend remarked that it bordered on the miraculous that he—yes, he!—ever existed. A classic victim of the self-selection bias. Only someone who is alive can make such an observation. Nonentities generally don’t consider their nonexistence for too long. And yet precisely the same delusion forms the basis of at least a dozen philosophers’ books, as they marvel year in, year out at the development of language. I’m quite sympathetic to their amazement, but it is simply not justified. If language did not exist, philosophers could not revere it at all—in fact, there would be no philosophers. The miracle of language is tangible only int the environment in which it exists.
Particularly amusing is this recent telephone survey: A company wanted to find out, on average, how many phones (landline and cell) each household owned. When the results were tallied, the firm was amazed that not a single household claimed to have no phone. What a masterpiece.
* Source: The Art of Thinking Clearly by Rolf Dobelli