Reviews
| DOI | http://doi.org/10.1111/1468-2230.t01-1-00344 |
| Published date | 01 July 2001 |
| Date | 01 July 2001 |
REVIEWS
Joseph L. Gastwirth (ed),Statistical Science in the Courtroom, New York:
Springer, 2000, xxii + 443 pp, hb, £41.00.
This collection covers a wide range of topics: DNA evidence, sampling,
epidemiology, deterrent effects of the death penalty, economic analysis of warranty
contracts and the estimation of quantities of smuggled drugs for sentencing purposes.
Beyond their interest in statistics and law, the contributors share no particular
perspective: even among those writing on DNA, some adopt classical, and others
Bayesian approaches. The papers differ enormously in their technical difficulty:
some are accessible to those with no prior knowledge of statistics, others require
some familiarity with calculus or basic statistical concepts, still others will, I
imagine, prove taxing even for some statisticians.
Despite the volume’s heterogeneity, there is plenty of interest in it. To pick out just
a few topics: Wendy Wagner presents an insightful analysis of judicial review of the
decisions of regulatory agencies, and of how the blunt criticisms of the courts can
lead these agencies to adopt rather slipshod approaches to their tasks. The Shonubi
case is something of a cause ce
´le
`bre in law and statistics circles: Shonubi was caught
entering the United States with 103 heroin filled balloons in his digestive tract. From
an analysis of the contents of just four of these balloons, the court needed to estimate
the total weight of heroin smuggled on this and seven other trips. Two papers in this
collection shed new light on the problem, suggesting that this sample was too small
by itself to give a straightforward estimate of the weight of drugs carried on even the
last trip. And the ‘trip effect’ – the assumption that a cautious Shonubi would have
carried smaller quantities of heroin on his earlier trips – is also questioned. Then there
are several papers on DNA. I am fairly familiar with statistical discussions of DNA
evidence, but David Balding’s paper here impressed me with its fluid and persuasive
analysis of the key issues, all presented at a level that only the most mathophobic
lawyers will find impenetrable.
I interspersed my reading of the – sometimes rather dry – papers in this collection
with Alain Desrosie`res’ The Politics of Large Numbers (1998). One of the many
themes of this subtle social history of statistical reasoning is that during the
nineteenth century the assumptions underlying statistical models were more
transparent because they were called into question in the various ‘contact zones’
between statistics and other disciplines. But now that numbers rule the world the
contingencies underlying them are seldom scrutinised. Parts of the legal system,
however, do today seem to collide with statistics in such a way as to force some
rethinking of suppositions. In Shonubi, the appeal court twice found itself dissatisfied
with attempts to use the quantities of drugs found on other smugglers as the basis for
extrapolating the total amount smuggled by Shonubi, demanding, instead, ‘specific
evidence’ of his grand total. Statisticians are probably right to suspect that here there
is something of the ‘statistics can’t prove anything’ mentality at work. But an
awareness of how acts of classification – ‘creation of spaces of equivalence’, in
Desrosie`res’ terms – underpin statistical reasoning might alert one to the possibility
that something subtle is going on here. In Shonubi, the appeal court and the
statisticians are perhaps working with different ways of classifying people and their
behaviour. The court, used to dealing with individuals, is naturally reluctant to use
ßThe Modern Law Review Limited 2001 (MLR 64:4, July). Published by Blackwell Publishers,
108 Cowley Road, Oxford OX4 1JF and 350 Main Street, Malden, MA 02148, USA. 643
To continue reading
Request your trial