I am amazed at how many people draw incorrect conclusions
by confusing absolute and relative values. My grand-mother,
may she rest in peace, was so convinced she had saved money
by buying on sale something she did not need in the first place
(a fallacy of the type “the more you buy, the more you save”).
In this age of environmental concern, I notice the same type
of flawed logic about fuel consumption. Because a cold engine
is less fuel-efficient, I see much advice against short car trips,
during which the engine does not reach nominal temperature.
A longer trip on the freeway is admittedly more fuel-efficient
than a short one in the city, but it still consumes more fuel.
In other words, the planet would be better off if we tried hard
to suppress our longer car trips rather than our shorter ones.
Just a week ago, the Belgian newspaper Le Soir was running
a five-page report on their survey of religious faith in Belgium.
The report included 14 tables counting 9 to 40 numbers each,
but only one graph displaying… four numbers! This imbalance
is in itself disappointing—graphs typically excel at displaying
many numbers in a small space—but the incorrect conclusions
are even more bothersome. For example, the survey indicates
that the fraction of Catholics in Belgium went from 65 to 60%
between 2005 and 2010 but also that the Catholics pray more
than in the past: 50% now pray every day, against 46% in 2005.
The text goes on to point out the same kind of “strengthening”
of Catholic faith: reading the Bible, believing in miracles, etc.
If you do the math, however, you will conclude more correctly
that nothing has changed: for example, 50% of 60% for 2010
is exactly as many people who pray every day as 46% of 65%
for 2005 (assuming a constant population over the five years).
That is, the 5% who no longer regard themselves as Catholic
were most likely not among those praying every day in 2005.
A better visual representation, possibly combined with a table,
would help lift this kind of naive interpretation of the numbers.