How to Lie with Statistics

Statistical thinking will one day be as necessary for efficient
citizenship as the ability to read and write. -- H. G. Wells

Back in the mid eighties I was watching a Sunday news
discussion where the pros and cons of Capital punishment (The Death
Penalty) was the focus of debate. The pro-death penalty guest
showed that in the decade since the reinstatement of the death
penalty by the United States Supreme Court ( in 1976) Capital
crime was down across the country. This broad statement is true.

However, closer examination of the data showed that [in any
given year] States that adopted capital punishment laws were
as likely to experience a rise or decline in those crimes as none
capital punishment states [during the same years]. This fact
was credibly demonstrated by the opposing guests, who produced
detailed graphs to make their case. 
This example shows how statistic are used to promotes
agendas through lies and half truths. In other words take from
the data what supports your cause and throw the rest away.
And so it is with much of the Statistics we're fed by the Media.

This book is a good primer for a basic understanding of
statistics and the dangers in they use. It is said a picture is
worth a million words, if by implication that suggest instant
gratification, then in this book every chapter is preceded by
an illustration that captures the written text.
Darrell Huff prefaces this very informative piece with this
introduction:

 ... Henry G. Felsen, a humorist ... pointed out quite a
while ago, proper treatment will cure a cold in seven days,
but left to itself a cold will hang on for a week. So it is with
much that you read and hear. Averages and relationships
and trends and graphs are not always what they seem
...The secret language of statistics, so appealing in a
fact-minded culture, is employed to sensationalize, inflate
confuse, and oversimplify. Statistical methods and statis-
tical terms are necessary in reporting the mass data of
social and economic trends, business conditions ...
But without writers who use the words with honesty
and understanding and readers who know what they mean
, the results can only be Semantic nonsense...statistic
are making many an important fact " look like what she
ain't." A well-wrapped statistic is better than Hitler's "big
lie"; it misleads, yet it cannot be pinned on you.

Someone once went to a good deal of trouble to find
out if cigarette smokers make lower college grads than
non-smokers. It turned out that they did...smoking makes
dull minds...the fallacy is an ancient one which however
has a powerful tendency to crop up in statistical 
material , where it is disguised by a welter of impressive
figures.
 
It is the one that says if B follows A, then A has caused
B....It seems a great deal more probable however, that
neither of these things has produce the other, but both are
a product of some third factor. Can it be that the social
sort of fellow who takes his books less serious is likely
to smoke more? The point is when there are many
reasonable explanations you are hardly entitled to pick
one that suites your taste....

... Another kind of co-variation is one in which the relation-
ship is real but it is not possible to be sure which of the
the variables is the cause and which the effect. In some of
these instances cause and effect may change from time
to time or indeed both may cause an effect at the 
same time.... Perhaps the trickiest of them all is the very
common instant in which none of the variables has any
effect at all on the other, yet there's a real correlation. A
great deal of dirty work has been done with this one. The
poor grade among cigarette smokers is in this 
category, as are all to many medical statistics that are all
quoted without the qualification that although the relation-
ship has been shown to be real, the cause-and-effect
nature of it is only a matter of speculation.

These are just a few example from the book -- garden variety
statistic-- of the type we are fed by the media almost daily: 

I was listening to a top rated news station in New York in mid
March, 2001 when the reporter stated : "Statistics show that crime
is down and the prison population is up," he further stated, "no
reason was given but an official attributed this disparity to longer
prison sentences," handed out in the courts. Obviously there's
something missing about the data here. This one's for you folks.

This book has many examples showing numbers an graphs that
are a bit dated in terms of population, but the underlying principles
are quite current. Knowledge of basics statistics is a quality of life
issue I think.

Norton paperback reissue 1993 
"How to Lie with Statistics"