In the late 1980s, competition between the two largest brands of cola was fierce. Pepsi Cola ran a campaign they called “the Pepsi Challenge.” Their workers set up booths in public places and invited passerbys to taste two colas—Pepsi and Coke. The labels would be hidden, so samplers would not know which cola was which. They would take a sip of each cola and report which they preferred. According to Pepsi, a large majority preferred their product to that of their rival, Coca Cola.
I took the Pepsi Challenge once at the DuPage County Fair. After tasting both samples, I truthfully told the worker that they tasted the same, that I had no preference. However, I also stood by and watched them work. Each time someone agreed to take the challenge, the worker would pour about an ounce of cola from bottle A into a cup, then would pour the same amount from bottle B into another cup. The worker would then hand the cup from bottle B to the sampler, and after he or she had tasted that cola, he or she would receive the sample from bottle A.
No wonder more people thought they preferred Pepsi! Invariably, bottle A was Coca Cola and bottle B was Pepsi Cola. The Pepsi sample would always be fresh from the bottle still bubbling, while the Coke sample would have had a little time to go flat. The contest was rigged; yet, most people who took the challenge probably left the booth convinced that Pepsi tasted better than Coke because of that single experience.
People are easily fooled. Studies and surveys and polls offer copious amounts of information, but often they are skewed by procedures as subtle as the Pepsi Challenge. Average people are not often able to observe the way information is gathered the way I was able to observe the Pepsi Challenge. Professional surveys contain many kinds of data that help a person to determine the reliability of the results, but that data is not distributed as widely as the results themselves. Without that data (and knowledge of how to interpret it), a person has little chance to know which surveys and studies are accurate and which are rigged. Healthy skepticism is always recommended. J.