# Nimble Numbers (Do They Tell the Truth?)

Numbers will scream at you all your life. “9 out of 10 dentists recommend Sparkle tooth paste.” “70% of people prefer dogs over cats.” “66.6% of all girls prefer the color purple over the color pink.” Websites and textbooks and advertisers often have numbers for everything. But are those numbers always true?

Those numbers are most commonly found by asking people to answer a question with one or two answers (a poll) or a series of questions with a larger choice of answers (a survey). By asking the right questions, polls and surveys can get answers that don’t quite tell the truth.

Let’s look at advertisers who use dentists to sell a dental product. And imagine that you’ve just invented a brand new tooth paste that tastes like candy floss. You take samples to ten dentists and ask them if they would recommend it to their clients.

The dentists try the tooth paste and, while other tooth pastes they know would be better at dental care, they see nothing wrong with Candy Floss Paste. They know that some kids don’t like brushing their teeth and think that maybe they would be more likely to brush if the paste tasted like candy floss. Nine give you the recommendation.

The tenth thinks, “Well, this won’t hurt anyone, but other brands work better and still taste good.” He doesn’t give you a recommendation.

Still, you can brag that 9 out of 10 dentists recommended your paste. But what if you had given the dentists your paste as well as one of the most effective and yummy-tasting brands and asked which they would recommend? Think about it—then think about whether or not 9 out of 10 is really the truth.

Another way numbers can lie is found in who you ask. Imagine you stand outside a dog show and ask all the people coming in to watch the show: “Do you prefer dogs or cats?” Of course, most of the people will say they prefer dogs. After all, they are going to a dog show. If you want to get a poll that says more people prefer cats, take your poll outside of a cat show.

You might think this is an obvious example, but consider that many polls and surveys take place in your favorite shopping mall. Teenagers in a mall would be more likely to say yes to the question: “Do you plan on buying a new cell phone in the next year?” If you asked that same question in front of a senior citizen’s home, what do you think the results would be?

Now think about how you would get 66.6% of girls to say they preferred purple over pink. Here’s one way: Go to a schoolyard or mall and look for girls wearing purple. Ask them if they prefer purple over pink. If they are wearing purple, chances are very good that they will say that they prefer purple over pink. One, though, might prefer pink but didn’t have an clean pink shirt to wear that day.

Here’s another way. Get a celebrity or a person who resembles a popular singer. Dress that person in an expensive, cool purple shirt or dress. Studies have shown that people will give poll responses that they hope will get the approval of the attractive person asking the question. That answer might not be the truth, but it might help make purple more popular.

Next time you see a commercial or read a news story with a percentage in it, think about that number. More importantly, think about how that number was created. That number might not be as true as it sounds.