State v. Bloom

¶1(dissenting).

¶2I dissent. I agree, of course, as I did five years ago when we decided in State v. Schwartz, 447 N.W.2d 422 (Minn.1989), that when a proper foundation has been laid—that is, when there has been a showing that the tests were performed in accordance with appropriate laboratory standards and controls — DNA test results are admissible into evidence. At that time we hedged DNA testing about with the same procedural safeguards applicable to other kinds of scientific evidence: to insure a fair trial, the test data and methodology must be available for independent review by or on behalf of the opposing party, and the admissibility of statistical probability evidence should be limited pursuant to Rule 403 of the Minnesota Rules of Evidence, which permits the exclusion of relevant evidence if its probative value is substantially outweighed by the danger of unfair prejudice. See State v. Kim, 398 N.W.2d 544, 548 (Minn.1987).

¶3Today the majority has bulldozed that hedge into oblivion, ostensibly in the interest of “scientific proof,” with only a backward glance at the reason for excluding statistical probability evidence. Statistical probabilities are certainly useful tools for indicating the direction and utility of further scientific research and for supporting certain conclusions with respect to the validity and use of the results of the research. Statistics have, however, been known to lie — either inadvertently or purposefully. The debate surrounding the National Research Council’s recommendation for the development of a conservative “interim ceiling method” and the Council’s acknowledgment that the methods of calculating the probability of a random match, current at the time its 1992 report was issued, rest on inadequate data and unjustified assumptions about population substructure, both bear witness to the fallibility of statisti*173cal evidence. More germane to the present case, perhaps, is the wide divergence in the statistical probabilities proffered by the state’s two experts, both professing to use the “interim ceiling method:” The BCA’s forensic technician is prepared to testify to a 1 in 93,700 chance that a randomly selected person’s DNA would match at the five loci where he matched the defendant’s record to the specimen found at the crime scene. The imported expert, however, is prepared to testify to a 1 in 634,687 chance of a random match.

¶4The accuracy of a statistical probability may be of little moment to the scientific community because although discovery of the falsity of an assumption underlying a statistical probability may destroy the value of the statistic, that discovery may confirm the validity of the basic research. If, however, an erroneous statistical probability plays any significant role in the conviction of an innocent person, the error has not only destroyed the life of the innocent person but has in some sense dehumanized the community. The point is illustrated by the proof of mathematical probability received as evidence in the infamous Dreyfus case in 1899.1

¶5In the light of the potential for invalidity of the assumptions about the frequency of a particular pattern of DNA at a given locus, about population substructure, and about statistical independence, I am not persuaded that there is any more justification for admitting statistical probabilities with respect to DNA than there was for admitting probabilities in State v. Carlson, 267 N.W.2d 170 (Minn.1978), or in any of its progeny. Cf., Friedman v. Commissioner of Public Safety, 473 N.W.2d 828, 838, 843-45 (Minn.1991) (Coyne, J., dissenting) (explaining my approach to overruling established precedent; dissent joined by Keith, C.J.) If the underlying assumption is wrong, the statistics are worthless. Conversely, the price of error is too high.

¶6More to the point, perhaps, than the potential for error in these statistics is their potential for misuse. The state’s imported expert witness is prepared, as I have already noted, to testify that there is a 1 in 634, 687 chance of a random match at the five loci in question. That figure is, in a sense, fictitious; the witness concedes that he used the “interim ceiling approach” with its built-in adjustments. Statistical probabilities are, of course, only probabilities — estimates, not certainties. As the witness undoubtedly recognizes, however, the mere recitation of the number 634,687 conveys an aura of mathematical precision, lending the probability a credibility it may not deserve and a weight to which it has no logical claim.

¶7A quarter-of-a-century ago the California Supreme Court observed that “[mjathemat-ics, a veritable sorcerer in our computerized society, while assisting the trier of fact in the search for truth, must not [be allowed to] cast a spell over him.” People v. Collins, 68 Cal.2d 319, 66 Cal.Rptr. 497, 438 P.2d 33, 33 (1968) (en banc). Surely, the almost universal acceptance of accuracy and precision in mathematical calculations has not diminished in the intervening years.

¶8More inimical to our traditional system of legal proof in criminal trials than undeserved credibility is the danger that the fact finder will understand the statistic represents something which it does not — namely, that the probability refers to the odds that someone else than the defendant is the source of the specimen found at the scene of the crime. That mistake has so often found its way into the rhetoric of public figures and in erroneous reports in the public press that it may be impossible to erase it from the mind of the fact finder. Even the state’s expert fell into that trap on cross-examination.

¶9Moreover, it seems to me that the preoccupation with statistical evidence is very likely to divert attention away from those aspects of DNA evidence which are deserving of closer examination. Are the test results reliable? What is the laboratory’s “false positive match rate”? Does the laboratory which performed the tests observe rigorous, scienti*174fic standards and quality control? Did the technicians follow laboratory protocol in the tests in question? Does the laboratory have an objective and quantitative procedure for identifying patterns, and a clearly defined procedure for declaring a match, identifying potential artifacts2 and designing internal controls to determine the presence of the artifact in a particular test? In short, there are inherent limits to any particular technique, test results can be skewed by a medium which does not meet quality controls or by defective equipment, and there is always the possibility of human error. If the match reported in the present case is erroneous at any one of the five loci, there is no match at all and the statistical probabilities of a random match have no relevance whatever.

¶10Despite my conviction that the admission of statistical probabilities is more likely to be misleading than helpful and that it introduces unfairness into a criminal trial, I do believe that DNA typing has now demonstrated sufficient reliability and particularity to justify some modification of the rule set out in the Kimcase, supra. It seems to me that it is more accurate to inform the jury that matches like those present in this case are “rare” or even “extremely rare” than to say only that the test is “consistent with” defendant being the source of the specimen found at the scene of the crime. In keeping with expert testimony in other contexts, I also consider admissible an expression, given an appropriate foundation for the opinion, that to a reasonable degree of scientific certainty the defendant is the source of the specimen. That information places the jury in a suitable position to assess the DNA evidence and protect the right of the defendant to a fair trial while at the same time avoiding the pitfalls attendant upon the illusion of mathematical precision.

¶11This court has heretofore adhered to the principle that conviction of the defendant in a criminal trial depends on the jury’s determination that, based on all of the evidence before it, the state has proved the defendant’s guilt beyond a reasonable doubt and that a criminal trial should not be resolved by a battle of experts. That a civil case should turn on the opinion of competing experts is considered by some as cause for alarm. That a criminal trial should be determined on the basis of probability estimates ought to give all of us pause.

¶12That mine is the lone dissenting voice in this matter does not shake my conviction that the day will come when this court regrets today’s decision.