types of supporting Material 8.3 171
USING STATISTICS AS SUPPORT Statistics can be expressed as either counts
or percentages. Verizon CEO Ivan Seidenberg used a count and a percentage in
the same sentence in his speech to a communications conference:
Using smart grids and mobile technologies to manage electric power
could create 280,000 new jobs and cut carbon emissions by more than
20 percent by 2020.^19USING STATISTICS EFFECTIVELY The following guidelines can help you to
analyze and use statistics effectively and correctly:
- Use reliable sources. It has been said that figures don’t lie, but liars figure!
Indeed, statistics can be produced to support almost any conclusion desired.
Your goal is to cite authoritative and unbiased sources. - Use authoritative sources. No source is an authority on everything; therefore,
no source can be credible on all subjects. The most authoritative source is
the primary source—the original collector and interpreter of the data. If you
find an interesting statistic in a newspaper or magazine article, look closely
to see whether a source is cited. If it is, try to find that source and the origi-
nal report of the statistic. Do not just assume that the secondhand account,
or secondary source, has reported the statistic accurately and fairly. As often
as possible, go to the primary source. - Use unbiased sources. As well as being reputable and authoritative, sources
should be as unbiased as possible. We usually extend to government research
and various independent sources of statistics the courtesy of thinking them
unbiased. Because they are, for the most part, supposed to be unaffiliated
with any special interest, their statistics are presumed to be less biased than
those coming from such organizations as the American Tobacco Institute, the
AFL-CIO, or Microsoft. All three organizations have some special interest at
stake, and the data they gather are more likely to reflect their biases.
As you evaluate your sources, try to find out how the statistics were
gathered. For example, if a statistic relies on a sample, how was the sample
taken? A Thursday afternoon telephone poll of 20 registered voters in Brooklyn
is not an adequate sample of New York City voters. The sample is too small
and too geographically limited. In addition, it excludes anyone without a
telephone or anyone who is unlikely to be at home when the survey was
conducted. Sample sizes and survey methods do vary widely, but most
legitimate polls involve samples of 500 to 2,000 people, selected at random
from a larger population. Of course, finding out about the statistical meth-
odology may be more difficult than discovering the source of the statistic,
but if you can find it, the information will help you to analyze the value of
the statistic. - Interpret statistics accurately. People are often swayed by statistics that sound
good but have, in fact, been wrongly calculated or misinterpreted. For ex-
ample, a speaker might say that the number of children killed by guns in