25.02.2015 Views

Asking the Right Questions, A Guide to Critical Thinking, 8th Ed

Asking the Right Questions, A Guide to Critical Thinking, 8th Ed

Asking the Right Questions, A Guide to Critical Thinking, 8th Ed

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Are <strong>the</strong> Statistics Deceptive? 161<br />

hair spray? As for <strong>the</strong> second statement, wouldn't you want <strong>to</strong> know how<br />

many of those deaths would have been prevented if an SUV were not<br />

involved, how many o<strong>the</strong>r mo<strong>to</strong>r vehicle fatalities not involving an SUV<br />

<strong>the</strong>re were, <strong>the</strong> number of SUVs on <strong>the</strong> road compared <strong>to</strong> how many<br />

deaths <strong>the</strong>y were involved in, and how many miles SUVs travel compared<br />

<strong>to</strong> how many deaths occur in SUVs? With reference <strong>to</strong> <strong>the</strong> third statement,<br />

how does <strong>the</strong> budget of one particular movie relate <strong>to</strong> <strong>the</strong> budget of o<strong>the</strong>r<br />

movies, and is this one case highly unusual, or is it typical of <strong>the</strong> movie<br />

industry?<br />

When you encounter statistics, be sure <strong>to</strong> ask, "What relevant information<br />

is missing?"<br />

Risk Statistics and Omitted Information<br />

"Daily use of Nepen<strong>the</strong> brand aspirin will lower <strong>the</strong> chance of a second heart attack<br />

by 55 percent."<br />

"Routine physicals have been linked <strong>to</strong> finding early cures and lowering people's<br />

likelihood of early death by 13 percent."<br />

A common use of statistics in arguments—especially arguments about health<br />

risks—is <strong>the</strong> reporting of risk reduction as a result of some intervention. Such<br />

reports can be deceptive. The same amount of risk reduction can be reported<br />

in relative or absolute terms, and <strong>the</strong>se differences can greatly affect our<br />

perceptions of <strong>the</strong> actual amount of risk reduction.<br />

Imagine a 65-year-old woman who just had a stroke and is discussing<br />

treatment options with her doc<strong>to</strong>r. The doc<strong>to</strong>r quotes statistics about three<br />

treatment options:<br />

(1) Treatment X will reduce <strong>the</strong> likelihood of a future stroke by 33 percent,<br />

(2) Treatment Ywill reduce <strong>the</strong> risk by three percent, and<br />

(3) With treatment Z, 94 percent of women are free of a second stroke for<br />

10 years, compared <strong>to</strong> 91 percent of those who go untreated.<br />

Which treatment should she choose? Our guess is that she will choose <strong>the</strong><br />

first. But all of <strong>the</strong>se options refer <strong>to</strong> <strong>the</strong> same size treatment effect. They just<br />

express <strong>the</strong> risk in different ways. The first (<strong>the</strong> 33 percent) is <strong>the</strong> "relative<br />

risk reduction." If a treatment reduces <strong>the</strong> risk of heart attack from 9 in<br />

100 <strong>to</strong> 6 in 100, <strong>the</strong> risk is reduced by one-third, or 33 percent. But <strong>the</strong>

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!