Koehler, J. J. (1997). One in Millions, Billions and Trillions: Lessons from People v.Collins (1968) for People v. Simpson (1995). Journal of Legal Education. 47, 214-223.

ABSTRACT

In both Collins and Simpson, very tiny statistical frequencies were used by prosecutors to link defendants with a crime. And in both cases, the matching characteristics that lurked behind these frequencies (eyewitness characteristics in Collins, DNA characteristics in Simpson) did have probative value.

But in both cases, the probative value of this evidence, as suggested by the frequency statistics, was severely constrained by the possibility of error. This is not to say that error was likely in either case. But because the possibility of error was much more likely than one in 12,000,000 in Collins , and one in 57 billion in Simpson, a serious FRE 403 concern arose. That is, jurors might not only have failed to understand how to combine the possibility of error with the theoretical frequency statistics to identify the probative value of the DNA evidence, but they may have believed mistakenly that the possibility of error was either irrelevant or was already assimilated into the frequency statistics.

If jurors are confused by DNA statistics, then there is a real risk that they may mistakenly believe that frequencies like 1 in a million or 1 in a billion estimate the chance that defendants are innocent of the crimes they are accused of committing. This translation error, which has been called the "inverse fallacy," is well-documented in the scientific literature on probabilistic reasoning.

In light of the problems statistical evidence and arguments can create in the courtroom, mechanisms for handling these problems should be discussed. Some have argued that statistics should be excluded from court because they are easily manipulated and difficult to understand. I disagree but believe that the odds are greater than one in a million that this issue, and others raised here, will spark lively debate in law school classrooms for years to come.