The Fallacy of DNA Databases


Today’s Headlines from Libertarian Sites

Nation of Weather Weenies?
by Patrick J. Michaels (CATO

With the Yield Curve?
by Frank Shostak (Mises Institute)


DNA evidence has proven extremely
useful in solving crimes. Many people convicted of crimes before the advent
of DNA evidence have been exonerated and released. But while DNA is a
very useful tool, it has its limitations and unfortunately governments
in Europe are misusing DNA — and of course law enforcement officials
in the United States want to join them.

The potential tragedy that
misuse of DNA evidence can cause was illustrated recently when police
in Great Britain admitted they misidentified a suspect based on DNA. Although
the suspect’s DNA matched DNA found at the scene of the crime, the gentleman
had an airtight alibi and after additional tests were performed the man’s
DNA was found to be slightly different. According to “experts” quoted
in numerous stories, this was a random problem that has only a 1 in 37
million chance of happening. Unfortunately that claim is nonsense — because
of the way Great Britain utilizes DNA evidence, there is about a 1 in
52 chance of such mismatches happening routinely.

DNA evidence is compelling
when comparing DNA found at the scene of a crime with that of a suspect
identified through traditional police methods. If a woman is killed and
her husband is a primary suspect, a DNA test might be useful in attempting
to exclude him as a possible suspect. The odds of any two randomly selected
individuals having a false positive is extremely low.

The problem with DNA starts
when governments start amassing huge databases. In much of Europe, DNA
samples are collected and kept on file for anyone arrested (much like
fingerprints). In Great Britain, for example, almost 700,000 DNA samples
are now included in a law enforcement database. Many law enforcement officials
in the United States want to do the same thing — take DNA from everyone
arrested and dump it into a huge database.

Then, anytime DNA is found
at a crime scene it can simply be compared to the samples in the database
and the perpetrator discovered more quickly. Except, as the British recently
found out, the result is likely to be junk. While there may be a 1 in
37 million chance of any two people in Great Britain giving the same result
on its DNA test, when you compare a DNA strand found at a crime scene
to almost the DNA of 700,000 samples, the odds of a false positive rise
to 1 in 52. That means for every 52 matches generated by such a search,
it should be expected that at least 1 is a false positive.

What happens when law enforcement
starts entering literally the DNA of literally millions of people into
its databases? Given the large number of arrests for drug and other related
offenses in the United States it is not inconceivable that within 10 years
such a DNA database would contain material taken from 20 million suspects.
At that point, the odds fall to 1 in 2 that any given match is a false

But actually the odds of a
false positive are much higher for a couple reasons. First, not all U.S.
law enforcement agencies use the same methods of quantifying DNA. When
DNA samples are analyzed laboratories usually look at up to 10 different
markers to compare samples. Using 10 markers yields the 1 in 37 million
odds of any two samples being identical. But many state agencies currently
collecting DNA evidence use only 6 or 8 markers (the more markers, the
higher the cost of the test). So existing DNA evidence is even more prone
to false positives.

Second, the 1 in 37 million
figure assumes that there is an equal distribution of DNA characteristics
throughout the population. This may be true when comparing samples of
two random individuals, but certainly will not be true of any DNA database
created in the United States. Some DNA characteristics vary by race, obviously,
and since a much larger percentage of minorities would be included in
a U.S. DNA database than non-minorities (compared to their distribution
in the normal population), the result would be that the distribution of
DNA characteristics in the resulting database would be far from even.
The only way around this would be to include the DNA of every person in
the general population which, unfortunately, would be self-defeating since
it would raise the odds of a false positive rise dramatically; assuming
a U.S. population of 260 million, running a random DNA sample against
such a database would on average yield more than 7 false positives for
every genuine match.

Now in Great Britain this may
not be much of a problem. British authorities have already done away with
their version of the Fifth Amendment — judges and juries can now consider
a suspect’s unwillingness to testify as evidence of his or her guilt.
There is also talk by the ruling Labor Party of doing away with trial
by jury for a whole host of crimes such as burglary (jury trials are said
to be too expensive). But the United States was founded on the view that
protecting the rights of the individual against the state is of paramount
importance. The People insisted that the right against self-incrimation
and the right to a trial by jury were preserved in the Bill of Rights
specifically to avoid the potential shenanigans that now dominate British

The recent false accusation
caused by a British DNA database is one more reasons to be skeptical of
poorly thought out European imports to the American legal system.

Leave a Reply