7.4 Reasoning with statistics 273
Commentary
Well, the facts (as presented in the graphs)
may speak for themselves, but do they speak
for Mr Keyes of Indiana? Chart 3 indicates that
reported crime in the USA rose sharply and
increasingly through the 1960s; and, with
occasional temporary falls, throughout the
1970s and 1980s too. It peaked around 1991 at
close to 15 million reported crimes. Over the
next two decades it fell back to just over
10 million, a decrease of 4.5 million, or 30%.
Meanwhile the numbers in prison (Chart 4),
which had been under 200 per 100,000 of the
population prior to the 1970s, rocketed over
the next three decades, with one small
reduction in the late 1990s and another
around 2009. The increase between 1970 and
2008 was over 360%.
So, for two decades – 1970 to 1990 – crime
rates and imprisonment both rose. But whilst
the imprisonment rate rose continuously, the
crime rate fell back three times before reaching
its peak. If the increasing imprisonment rate
was ‘working’ it looks as though it was
working for a time, then failing again. Then,
after 1991, with imprisonment still on the
same steep rise, crime began a more or less
steady descent. But we do not know what
happened after that, or what will happen in
the future. You must decide whether there is a
sufficiently strong pattern or trend in Chart 3
to make a reliable prediction, or to support
Keyes’ hypothesis.
Remember that the task you were set was to
criticise the data as evidence. This does not
mean that there is anything wrong with the
data itself. The graphs are based on official
statistics, and therefore come from reputable
and reliable sources. You are not asked to
assess their credibility. It is the interpretation
that we are concerned with. The question is
whether the statistics:
A positively support [3]
B fail to support [3]
C contradict or disprove [3].
The answer is almost certainly B; and here are
some reasons why. Firstly, even if the data is
interpreted as a strong correlation between
the rise in imprisonment rates and the fall in
crime rates (which is questionable on the
basis of Chart 3), there is nothing to indicate
which is the cause and which the effect. In
Chapter 2.10, and several times in Unit 4, the
fallacy of assuming cause on the basis of
correlation was discussed, and you shouldIn answering the question you will need to
critically assess the statistics represented
by the graphs as evidence for [3]. You are
not asked to assess the accuracy of the
data: assume it is correct. But do ask
yourself how clear and transparent the
presentation is. Does it hide or distort any
of the relevant information?CHART 316,000,000
Reported crimes in the USA: 1960−201014,000,000
12,000,000
10,000,000
8,000,000
6,000,000
4,000,000
2,000,000
0
1960 1970 1980 1990 2000 2010
CHART 4195019550100200300400500600Prisoners per 100,000 population700800USA imprisonment rate: 1950–200819601965197019751980198519901995200020052010