Controversy & Debate

Can People Lie with Statistics? Josh: discussing job prospects after graduation) Well, you know, college students today just aren’t as smart as they were fifty years ago. Sam: Come on, that’s not true at all. Josh: (smugly) Sorry, pal. I happen to have the data to prove it.

e have all been in arguments when someone has presented us with “data” as if that were “proof.” But are numbers the same as “truth”? It is worth remembering the words of the

nineteenth-century English politician Benjamin Disraeli, who once remarked, “There are three kinds of lies: lies, damned lies, and statistics!” In a world that bombards us with numbers—often de-scribed as “scientific data” or “official figures”—it is important to realize that “statistical evidence” is not necessarily the same as truth. For one thing, any researcher can make mistakes. More important, because data do not speak for themselves, someone has to decide what they mean. Sometimes people (even sociologists) “dress up” their data almost the way politicians deliver campaign speeches—with an eye more to winning you over than to getting at the truth. The best way to avoid being fooled is to understand how people can mislead with statistics.

1. People select their data. Many times, the data presented are not wrong, but they do not tell the whole story. Let’s say someone who thinks that television is ruining our way of life presents statistics indicating that we watch more TV today than people did a generation ago. It also turns out that during the same period, SAT scores have fallen. Both sets of data may be correct, but the suggestion that there is a cause-and- effect link here—that television viewing is lowering test scores—is not proved. A person more favorable to television might counter with the additional “fact” that the U.S. population spends much more money buying books today than it did a generation ago, suggesting that television creates new intellectual interests. It is possible to find statistics that seem to support just about any argument.

2. People interpret their data. People can also “package” their data with a ready-made interpretation, as if the numbers can mean only one thing. The pie chart shows the results of one study of U.S. children living in poverty

W

(National Center for Children in Poverty, cited in Population Today, 1995). The researchers reported that 43 percent of these children lived in a household with no working parent, 39 percent lived in a household with one or two parents employed part time, and 18 percent lived in a household with one or two parents working full time. The researchers labeled this figure “Majority of Children in Poverty Live with Parents Who Work.” Do you think this interpretation is accurate or misleading? Why or why not?

3. People use graphs to spin the truth. Graphs, which often show an upward or downward trend over time, are a good way to present data. But using graphs also gives people the opportunity to spin data in various ways. The trend depends in part on the time frame used. During the past ten years, for instance, the U.S. crime rate has fallen. But if we were to look at the past fifty years, we would see an opposite trend: The crime rate rose sharply. The scale used to draw a graph is also important because it lets a researcher “inflate” or “deflate” a trend. Both graphs shown here present identical data for SAT critical reading scores between 1967 and 2012. But the left-hand graph stretches the scale to show a downward trend; the right-hand graph compresses the scale, making the trend seem steady. So understanding what statistics mean—or don’t mean—depends on being a careful reader!

What Do You Think?

1. Why do you think people are so quick to accept “statistics” as true?

2. From a scientific point of view, is spinning the truth acceptable? Is this practice OK from a critical approach, in which someone is trying to advance social change?

3. Find a news story on some social issue that you think presents biased data or conclusions. What are the biases?