Book review
Freakonomics
Stephen Levitt, May 2005
Author: Stephen Levitt
Publisher: William Morrow
Review by Donald Clark
It’s not often that an economist tops the best
seller lists, but Steven Levitt is no ordinary economist. He rose
to fame when he spotted that the famed drop in New York crime, much
vaunted by Mayor Giuliani, was in fact mirrored in many other large
US cities, without the famous zero-tolerance methods. The data showed
that the 40% drop in crime rate was closely correlated with those
states that legalised abortion in the early 70s. Quite simply, the
criminals weren’t being born. This caused mayhem but he appears
to have been right. It was a political myth.
Regarded by some as the smartest person in America or the ‘Indiana
Jones of economics’, his specialism is economic detective
work. He wrestles with problems and surprise culprits usually emerge
from his deep investigation into the economic data. He’s in
the business of scotching myths and defying ‘conventional
wisdom’, a phrase invented by the economist John Galbraith
to describe common beliefs that are usually wrong.
The book is of particular interest to those of us in the education
and training as many of the case studies look at the business of
teaching and learning.
Educational myths
Looking at parenting he noted that parents are terrible at assessing
risks. Swimming pools, for example, are 100 times riskier than guns
in the US, with a far higher death rate (1 in 11,000 as opposed
to 1 in 1 million plus). This led him to examine school data to
see what influence parents had on their child’s educational
development.
He rightly quotes Stephen Pinker and Judith Harris, who have written
widely on the subject, rocking the psychological and educational
establishment with their reassessment of the role of parents and
teachers in child development. More importantly he had access to
great data from Chicago where students were allocated to schools
by lottery. This was an almost perfectly designed social experiment
as it made analysis of the data cleaner. To cut to the chase –
school choice barely mattered at all in terms of educational achievement,
and when it did it was in those non-academic kids who were given
vocational training.
He then turned his attention to data from a huge Early Childhood
Longitudinal Study (20,000 students), applied regression analysis,
and uncovered some fascinating conclusions. The things that DO NOT
enhance your child’s educational attainment are:
- Attending Head Start or Sure Start programmes
- Reading to them every day
- Taking them regularly to museums
- Mother working between birth and kindergarden
- Stopping them watching a lot television
In short, don’t worry if your kid has not attended a nursery,
that you failed to read to then before bedtime, that you had no
time to take them to museums, that you had kids when you were older
or that they watch a lot of television. It makes no difference.
Having tried these hypotheses out on a number of parents recently,
be prepared for some furious reactions.
The things that do seem to matter are:
- Mother giving birth after age of 30
- Educational attainment of parents
- Having lots of books in the house
- Adoption (negatively)
- Low birth weight (negatively)
From another set of data concerning the behaviour of parents in
picking up their children from nursery, Levitt looked at incentives.
Fed up of parents turning up late for their children the nursery
in question turned to $3 fines for late parents. Lateness got worse
as parents could turn up late, pay the fine and feel no guilt. Even
after the fines were dropped the latecomers remained high. Such
is the odd world of incentives.
In another study, looking at student test scores from 1996 onwards,
he found unequivocal evidence of widescale teacher cheating in the
exams. As he says, teacher cheating is rarely looked for, and just
about never punished. Because the cheaters were unsophisticated
and lazy in their techniques it was relatively easy to write a computer
programme that spotted unusual answering patterns. He found definitive
cheating in over 200 classrooms per year. In another study, some
35% of respondents said they had witnessed their fellow teachers
cheating in some fashion. In Chicago, a far thinking administrator
implemented a scheme to retest 120 classrooms against a control
group. The results were compelling. Some of the teachers were sacked,
but more importantly all had been warned. Cheating fell by 30% in
the following year.
Other myths
The study on online dating is also fascinating, if only on the
level of gossip. In an analysis of 30,000 online daters, 50% on
west coast, 50% on east coast they found that:
- Getting a date is hard – 57% of men receive no emails,
23% of women
- Photos matter 75% less emails if you don’t provide one
- Men looking for a long term relation ship do great (the others
don’t)
- Women looking for an occasional lover do great (the others don’t)
- Tall men get more responses
- Overweight men seem not to be at a great disadvantage
- Blonde women get more responses
- For women, being overweight is deadly
To get some idea of the weird scope of his studies, other chapter
titles include:
- How is the Ku Klux Clan like a group of Real-Estate Agents?
- Why do drug dealers still live with their moms?
- Roshanda by any other name would smell as sweet
This is a refreshing return to books that use data rather than
clever hypotheses to make a point. Books like Blink, Slow, Blur,
Emergence, The Tipping Point and The Wisdom of Crowds pick an abstract
noun or concept and spin an all-encompassing theory around the word
or concept. This has become a sort of publishing genre.
Freakonomics looks as though it’s part of this trend –
but it’s not. Levitt works on the opposite principle. His
is the reverse, not idea driven but analysis driven. He looks at
data and tries to find causes from within the data. There is no
‘theme’ and as the philosopher Nozick said, when interviewing
Levitt, ‘He’s 26 years old. Why does he need to have
a unifying theme?… He’ll take a question and he’ll
just answer it, and it’ll be fine’.
The world of education and training could do with more of this
hard science. Far too much of our theory is faddish and non-empirical.
Some myth-busting would be very useful.
|
