Government of the Virgin Islands v. Byers

941 F. Supp. 513, 35 V.I. 240, 45 Fed. R. Serv. 1247, 1996 WL 599116, 1996 U.S. Dist. LEXIS 15582

• Court: District Court, Virgin Islands
• Decided: October 11, 1996
• Docket: Crim. 1992-52
• Status: Published

Opinion

MOORE, Chief Judge

OPINION

I. Factual & Procedural Background

II. Standard of Admissibility

A. Qualification of Government's Expert

B.The Reliability of the DNA Profiling Technique

1. DNA

2. The DNA Profiling Process

a. Laboratory

b. Matching

c. Population Genetics and Statistics

3. The 1992 NRC Report

a. The Ceiling Principle

b. Estimating Error Rates

4. The Daubert Factors

a. Whether the Proffered Technique Can be (and Has Been) Tested

b. Whether the Theory or Technique Has Been Subject to Review

c. The Existence and Maintenance of Standards Controlling the Technique's Operation

d. The Known or Potential Rate of Error

e. Degree of Acceptance By a Relevant Scientific Community

5. The Downing Factors

a. The Novelty of DNA profiling and its Relationship to More Established Modes of Scientific Analysis

b. The Qualification and Professional Stature of Expert Witness

c. The Non-Judicial Uses to Which the Method Has Been Put

C. Relevance of FBI's DNA Profiling Evidence

D. Rule 703 Considerations

E. Overwhelming, Confusing or Misleading Jury

III. Conclusion

MEMORANDUM

The prosecution has moved in limine to admit deoxyribonucleic acid ["DNA"] profiling evidence. For the reasons articulated below, the motion will be granted.

I. Factual & Procedural Background

The defendants have been charged in a 37-count information with conducting a joint four-month 1 crime spree in St. Thomas that included rape, robbery, kidnapping, and burglary. Between January 18 and April 28,1990, two masked men broke into four homes, robbed the occupants of their valuables, and raped the female occupant of each home. According to the victims, both men were black. The ordeal of the first rape victim was especially brutal. After she was raped at home, her two assailants forced her to drive them to a remote location where she was raped a second time by the taller of the two men. Twelve days later she was kidnapped again by the taller of the two men, and raped a third time. Because her assailant removed his mask during this last incident, the victim was able to positively identify him as Reginald Sylvester Byers, Jr. Byers has since pled guilty to three counts of aggravated rape in exchange for his testimony against his codefendant, Caswell Fredericks.

Fredericks was never identified by any victim as an assailant in these crimes. But after speaking with Byers' family and friends, police learned that Fredericks and Byers were acquaintances. The police obtained blood samples from Fredericks which, according to the Federal Bureau of Investigation ["FBI"], match the DNA profile of the semen recovered from three of the four rape victims. 2 The FBI estimates that the probability of randomly selecting another person with a DNA profile matching the semen samples recovered from the victims would be as follows:

Victim M.S.: 1 in 1,000 for Caucasians,

1 in 550 for Blacks, and 1 in 300 for Hispanics (using two probes);

Victim B.B.: 1 in 300,000 for Caucasians,

1 in 100,000 for Blacks, and 1 in 35,000 for Hispanics (using three probes);

Victim M.M.: 1 in 300,000 for Caucasians,

1 in 100,000 for Blacks, and 1 in 35,000 for Hispanics (using three probes).

The defendant's opposition to the introduction of this DNA evidence is based, in large part, on the FBI's failure to implement certain recommendations contained in a 1992 report by the National Research Council ["1992 NRC Report"]. 3 The defendant also contends that the FBI's profiling process did not account for population substructure within the black race. According to the defendant, comparing DNA samples from Afro Caribbeans against a population database composed almost entirely of African Americans in the mainland United States would tend to skew the statistical estimates prejudicially.

Our task, then, is to decide whether the DNA profiling results against Fredericks should be admitted at trial, given the defendant's objections about their reliability and the possibility that these results may unfairly prejudice the jury. This task is made somewhat easier by a recent, comprehensive opinion of this Court 4 which examined the process of DNA profiling in considerable detail using the standards enunciated by Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 125 L. Ed. 2d 469, 113 S. Ct. 2786 (1993). Government of Virgin Islands v. Penn, 838 F. Supp. 1054 (D.V.I. 1993) (McGlynn, J.). Because the method of DNA typing remains essentially unchanged since the Penn decision and because the FBI also conducted the DNA profiling in that case, much of Penn's description on the FBI's method of DNA profiling is applicable here. To some extent, therefore, we have adopted Penn's explanations of the technical aspects of DNA profiling. Of course, whether these procedures are reliable and relevant under the applicable standard is a decision we must make on our own after an independent evaluation of the evidence presented at the gatekeeper hearing.

II. Standard of Admissibility

We begin our analysis by reviewing the applicable standards for admissibility In federal courts, the admissibility of novel scientific evidence is governed principally by Rule 702 of the Federal Rules of Evidence, which provides that

if scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education may testify thereto in the form of an opinion or otherwise.

This Rule requires the trial judge to act as a "gatekeeper" by admitting only scientific testimony that is relevant and reliable. Daubert, 509 U.S. at 589. Daubert superseded the more restrictive Frye test 5 which the Court found to be at odds with the "liberal thrust" of the Federal Rules. Id. at 588. Frye's general acceptance test is no longer the sine qua non of admissibility but merely one of many factors to be considered.

Rule 702 has three major requirements: (1) the proffered witness must be an expert; (2) the witness must testify to scientific, technical, or other specialized knowledge; and (3) the witness' testimony must assist the trier of fact. United States v. Velasquez, 33 V.I. 265, 64 F.3d 844, 849 (3d Cir.), reh'g denied (1995) (citing In re Paoli R.R. Yard PCB Litigation, 35 F.3d 717, 741-42 (3d Cir. 1994), cert. denied, 115 S. Ct. 1253 (1995) ("Paoli II")). The Court of Appeals for the Third Circuit has construed the first requirement liberally. "'We have held that a broad range of knowledge, skills, and training qualify an expert as such,' and have 'eschewed imposing overly rigorous requirements of expertise.'" Velasquez, 64 F.3d at 849 (quoting Paoli II, 35 F.3d at 741).

The second requirement of Rule 702 goes to the reliability and trustworthiness of the expert's testimony. Velasquez, 64 F.3d at 849. The testimony must be "'based on the methods and procedures of science' rather than on 'subjective belief or unsupported speculation.'" Paoli II at 741-42. The inquiry into reliability is focused not on results but on the method of arriving at those results. Daubert assumes that "if the principles, methodology and reasoning are scientifically valid then . . . the inferences, assertions and conclusions derived therefrom are scientifically valid as well." United States v. Bonds, 12 F.3d 540, 556 (6th Cir. 1993), reh'g denied (1994).

Daubert identified five nonexclusive factors to be weighed when determining scientific reliability: (1) whether a theory or technique can be (and has been) tested; (2) whether the methodology is subject to peer review and publication; (3) whether, and the frequency with which, the methodology leads to erroneous results; (4) whether the standards of controlling the technique's operation can be maintained; and (5) whether the theory or technique has been generally accepted in the particular scientific field. Daubert, 509 U.S. at 593-594. The Third Circuit Court of Appeals has appended three more factors to be considered: (6) whether, and how strong, a relationship the technique has to methods which have been established to be reliable; (7) whether the expert witness is qualified to testify based on the methodology; and (8) whether, and how often, the method has been put to nonjudicial use. Paoli II, 35 F.3d at 742 n.8. No single factor or combination of factors is dispositive.

The third requirement of Rule 702 is that an expert's testimony be relevant to some disputed factual issue in the case. Daubert, 509 U.S. at 589; Paoli II, 35 F.3d at 743; United States v. Downing, 753 F.2d 1224, 1237 (3d Cir. 1985). "Thus, even if an expert's proposed testimony constitutes scientific knowledge, his or her testimony will be excluded if it is not scientific knowledge for purposes of the case." Paoli II, 35 F.3d at 743.

Although Rule 702 is the "primary locus" of the judge's gatekeeper function, Daubert explained that other evidentiary rules, most notably Rules 703, 706 6 and 403, may inform a judge's decision whether to admit expert testimony. Daubert, 509 U.S. at 595. Rule 703 7 has not enjoyed widespread use in DNA cases, perhaps because — in the Third Circuit at least — the standard for reliability under Rule 703 is the same as that for Rule 702. Paoli II, 35 F.3d at 748.

Rule 403, 8 by contrast, is considerably more independent of Rule 702. Paoli II, 35 F.3d at 746. A finding of reliability under Rule 702 may nevertheless result in exclusion under Rule 403. Under Rule 403 the

trial court must. . . balance its assessment of the reliability of a novel scientific technique against the danger that the evidence, even though reliable might nonetheless confuse or mislead the finder of fact, and decide whether the evidence should be admitted.

Downing, 743 F.2d at 1240. Whether the shortcomings of the technique can be effectively communicated to the jury; whether the data on which the expert relied can be inspected by the jury; and whether there are other forms of incriminating evidence in the case are some of the considerations that enter into this balancing. Downing, 753 F.2d at 1239,1243. As a rule, exclusion under Rule 403 should be rare, Paoli II, 35 F.3d at 747 n.16, but because expert testimony can be misleading more often than other types of evidence, Rule 403 gives judges more power over experts than over lay witnesses. Daubert, 509 U.S. at 595.

In this case, then, the government must cross five major hurdles before the jury may hear its incriminating DNA profiling evidence against Fredericks:

A. its proffered witness must be qualified as an expert;

B. the DNA evidence itself must be deemed reliable (i.e., based . on the methods and procedures of science) after filtering for the factors enumerated in Daubert and Paoli IP,

C. the DNA evidence, although reliable, must be relevant to some disputed factual issue;

D. the data underlying the expert's opinion must be reliable under Rule 703; and

E. the evidence must not tend unduly to prejudice, overwhelm, confuse, or mislead the jury.

From a distance this legal standard may appear as an almost insurmountable barrier to admission. But given the sometimes unfortunate impulse toward liberal use of expert testimony encouraged by the federal rules and the fact that no federal court has ever excluded DNA evidence under Daubert, 9 the barrier becomes far less formidable upon closer inspection. The remainder of our analysis considers each of these five hurdles.

The government proffered Dr. Bruce Budowle, chief of the FBI Forensic Research Unit, as an expert in the fields of forensic biology, molecular biology, human genetics, statistics and population genetics. Transcript of October 16, 1995 Hearing ["Tr."] at 19. Dr. Budowle's credentials are impressive. He has been qualified as an expert in more than 40 cases in at least 16 states, the District of Columbia, and the Virgin Islands. He has been published on more than 100 occasions on DNA-related topics and has addressed numerous conferences and symposia. He is a member of three professional forensic organizations and serves on the editorial board of several scholarly journals. Tr. at 13-19. He has received several awards in his field, including the 1981 American Academy of Forensic Sciences Recognition Award. Under Rule 702's "liberal policy of admissibility," 10 we had little difficulty accepting Dr. Budowle as an expert in the fields of forensic biology, molecular biology, human genetics, statistics, and population genetics.

B. The Reliability of the DNA Profiling Technique

Doubts about the reliability of the FBI's method of DNA profiling form the core of defendant's arguments for exclusion. In Penn, Judge McGlynn addressed the process of collecting, processing, and interpreting DNA evidence in depth, and we need not repeat that comprehensive discussion here. We also take judicial notice of cases from other jurisdictions that have found the FBI's DNA profiling technique reliable under Daubert, 11 Against this backdrop, we focus on the impact of the 1992 NRC Report, if any, on the reliability of the FBI technique. A brief overview of DNA profiling and the controversy surrounding its forensic applications will suffice to limn the issues raised by the defendant.

1. DNA

DNA is a long, spiraling molecule that contains the genetic blueprint for the physical characteristics of all living organism. In humans, DNA is found in chromosomes in the nucleus of most cells, including semen, saliva, and white blood cells. Chromosomes are threadlike structures that carry genetic information from each parent. There are 46 chromosomes in each cell, 23 from each parent.

The DNA molecule is composed of two intertwined strands of nucleotide bases with the base on each strand weakly bonded to a base on the other to form a double helix, resembling a spiral staircase. The bonded bases forming the rungs of this staircase are called "base pairs". There are four such nucleotide bases assigned the letters "C", "T", "A", and "G". The C on one strand always pairs with the G on its complementary strand, and the A always pairs with the T. 12 The strand of base pairs in the DNA molecule in each cell is about three billion letters long. Tr. at 21.

The sequences of base pairs in the DNA in each individual is remarkably similar — ninety-nine to 99.99% is identical from person to person. 13 Yet it is an accepted fact of human genetics that, except for identical twins, no two individuals have precisely the same ordering of base pairs. 14 It is thus the 1% to 0.01% variation in the sequencing of these base pairs from person to person that allows for DNA profiling. The challenge of forensic DNA profiling, obviously, is to isolate for analysis those rare portions of DNA that vary from person to person, called alleles. Most alleles contain genetic coding for variable human characteristics, such as eye or hair color. Each person also has some DNA molecules, called "junk DNA", which do not code for any known physical trait. Tr. at 23.

Certain regions of junk DNA contain high concentrations of alleles, and are therefore ideal for measuring differences between people. See Tr. at 34. These regions are known as "variable number of tandem repeats" (VNTRs) and are made up of core sequences of base pairs. Both the composition of this core sequence and the number of times that it is repeated changes from chromosome to chromosome. Everyone has the same core VNTRs sequence at a particular location ["locus"] on a given chromosome, but the number of repetitions of this core sequence at that locus varies widely from person to person. The more repetitions of the core sequence, the greater the length of the VNTRs and vice versa. The essence of DNA profiling is its ability to differentiate people by selecting several standard loci on particular chromosomes and measuring the varying lengths of VNTRs at those loci. 15

2. The DNA Profiling Process

As stated in Penn, the DNA profiling process can be divided into at least three stages: the laboratory procedure, matching, and applying principles of population genetics and statistics. Penn, 838 E Supp. at 1059. 16

The first step in the laboratory procedure is to extract DNA from the blood, semen, or hair samples. Next, DNA molecules are cut into shorter fragments using molecular scissors designed to search for and separate out fragments containing VNTRs at the standard chromosomal loci. The result is a mixture of DNA fragments of many different lengths. 17

The fragments are then segregated by size with a process known as electrophoresis, in which the DNA fragments are placed in an agarose gel and a small electrical charge applied for several hours, with the positive charge at the bottom. Since DNA is negatively charged, the fragments migrate toward the positive pole of the electrical charge. The distance migrated varies according to the fragment's size. While still in the gel, the fragments are subjected to a chemical agent and denatured or separated into single-stranded DNA fragments.

The single strands of DNA fragments are then transferred to a nylon membrane using a process known as "Southern blotting." The nylon membrane is placed over the agarose gel and "soaks up" the DNA fragments by capillary action, much like a paper towel soaks up liquids. The single strands are immobilized on the nylon membrane to match their placement in the agarose gel by baking the nylon membrane.

The VNTRs fragments at the standard chromosomal loci are then "illuminated" by inserting from four to six 18 single-stranded, radioactive probes, each with a core base letter sequence identified with each respective standard location. 19 These single-stranded probes, which are inserted sequentially, will bind or "mate" with each of the complementary core base letter sequence repeated in the VNTRs fragments on the nylon membrane. The result is radioactive hybrid VNTRs fragments that will show up on x-ray film. The laboratory is thus able to make a permanent "photograph" of the nylon membrane on an autoradiograph ["autorad"] by exposing to x-ray film the membrane for several hours. The radioactive hybrid VNTRs appear as bands on the autorad, which gives a picture of the size of each of the target VNTRs by showing the distance that these VNTRs traveled in the agarose gel. See Penn, 838 R Supp. at 1060-62.

The next step is the matching stage, which includes both a subjective and objective component. Laboratory technicians make a visual inspection of the autorad to determine if there is a match between the known sample (K) from the defendant and the questioned or crime scene sample (Q) from the victim. The FBI also uses a control to assure the accuracy of its results. A match occurs when the VNTRs in both samples migrate the same distance through the agarose gel during electrophoresis. If a visual inspection reveals an approximate match, the laboratory uses a computer imaging program to increase the accuracy of the VNTRs measurements. Penn, 838 F. Supp. at 1062. The FBI computer model compensates for variations in measurement by using a match window of +— 2.5%. Id. at 1062-63. The computer measures the hybrid VNTRs several times, and if the individual measurements of a hybrid vary by more than this match window of 2.5%, the FBI does not declare a match. A match is declared only when the computer program confirms the visual inspection. The FBI estimates that one third of all suspects are cleared because their DNA profiles do not match the sample from the crime scene.

It is the final phase of the DNA profiling process — applying principles of population genetics and statistics — that has generated the most controversy. 20 This phase measures the frequency with which VNTRs fragments of a given length occur in the general population, including those at the selected standard chromosomal loci. The laboratory compares the suspect's DNA profile to the DNA profiles of randomly selected people in the FBI's databases. Using DNA from anonymous donors at blood banks and paternity testing labs, the FBI has constructed a database for each of the major racial groups. These databases are quite small. The Caucasian database, for instance, consists of DNA profiles of only 750 Caucasians from Florida, Texas, and California.

To measure the frequency with which VNTRs fragments of a given length occurs at standard chromosomal loci in a racial group, the FBI uses a system known as the "fixed bin process". The FBI created a DNA profile for each person in its population database and sorted the VNTRs fragments into bins. VNTRs of the same length were placed in the same bin; some bins contain more VNTRs fragments than others. The FBI built a number of features into its fixed bin process to adjust for measurement errors and to add a degree of conservatism to its DNA frequency estimates. One such feature is large bins. Generally, the larger the bin, the more VNTRs fragments it contains, and the greater the frequency of VNTRs fragments of that length. Tr. at 52-53. Another feature designed to minimize the possibility of error is minimum bin frequencies. Any bin with less than a 5 count (5% of the VNTRs fragments) was combined with another bin so that the sum of the two bins is greater than 5. Id. at 55.

The FBI arrives at a DNA frequency estimate for a given case by comparing the number of times or frequency a VNTRs of the same length of the suspect (taken from bands on an autorad) occurs in the several different population databases the FBI has established, not just for the racial group to which the suspect belongs. Based on the assumption that VNTRs at different loci are not somehow connected and that they behave independently, 21 the laboratory uses the Product Rule 22 to arrive at a probability of a coincidental match. Tr. at 57. Penn, 838 F. Supp. at 1064-65. This assumption is critical to the FBTs DNA frequency estimates, for if VNTRs are not independent, the Product Rule would yield inaccurately low frequency estimates. This means that the suspect's VNTRs would be found to appear less frequently in the population than it really does. The net effect of this would be to mislead the jury into concluding that it was more likely the defendant was the perpetrator than is scientifically or statistically reliable. It is this basic assumption of independence that has been challenged in recent years.

3. The 1992 NRC Report

In April 1992, after two years of study, a panel of scientists, judges, and lawyers published a major report on the use of DNA technology in forensics. Although it found the current laboratory procedures for detecting DNA variations "fundamentally sound," 23 the report made recommendations in several areas, including statistics and quality assurance. Not surprisingly, the National Resource Council's recommendations in the area of statistics proved very controversial. Most controversial of all was the ceiling principle.

The impetus for the ceiling principle came from a 1991 article by two prominent geneticists who found the statistical methods at DNA forensic laboratories "unjustified and generally unreliable." 24 These geneticists argued that the presence of population subgroups with varying DNA patterns and the tendency to mate within those subgroups contradicted the fundamental assumption that the bands on an autorad were independent of each other. 25 Therefore, multiplying VNTRs frequencies using the Product Rule would understate the frequency with which VNTRs of a given length occur in that subgroup. Thus, the ceiling principle was designed to account for the possibility that the bands on the autorad may not be completely independent.

The ceiling principle requires laboratories to draw random samples of 100 persons from each of 15-20 population subgroups, e.g., Italians, Puerto Ricans, Haitians, Koreans. When computing a DNA profile, the frequency assigned to each band of the autorad would be the highest observed frequency in any of the population groups or 5%, whichever is larger. 26 This approach tended to yield overly conservative estimates of the likelihood of a random match. 27 The FBI believes the ceiling principle is "unabashedly conservative" and has refused to adopt it. Tr. at 115.

The convergence of several events since the publication of the 1992 NRC Report has placed the two geneticists7 theory on the genetic effects of population subgrouping in doubt and has challenged the need for the ceiling principle. In 1993 the FBI completed a worldwide population study 28 which concluded that, while population subgroups exist, their effects on DNA profiling is not forensically significant. Tr. at 80-81. Outside experts who have tested the FBTs population databases have confirmed that the bands on the autorad behave independently. Tr. at 135-36. In addition, several studies published shortly before the 1992 NRC Report argue that the effects of population subgrouping may have been overstated. 29 Another criticism of the ceiling principle was that by emphasizing racial, ethnic, and geographic subgroups, it tended to ignore the reality that the population suspects belong to are usually mixed. 30 Population subgrouping is only relevant if the suspect is known to belong to a particular subgroup. Perhaps the most damning response was the suggestion that the ceiling principle is not science at all but merely an arbitrary policy decision designed to placate the defense bar. Given this level of critical response, at least one member of the commission that published the 1992 NRC Report has expressed reservations about the continued validity of the ceiling principle. 31

More importantly, in a report published earlier this year the NRC abandoned the ceiling principle altogether and recommended that "in general, the calculation of a profile frequency should be made with the product rule." National Research Council, The Evaluation of Forensic DNA Evidence, ES-4 (May 2, 1996) (prepublication copy) ["1996 NRC Report"]. If the race of the suspect is known, the database for the person's race should be used, but if the suspect's race is unknown "calculations for all the racial groups to which possible suspects belong should be made." Id. According to the 1996 NRC Report, subgroups are relevant only if the suspect is known to belong to that subgroup. If a suspect is known to belong to a particular subgroup, then VNTRs frequencies for that specific subgroup should be used in preparing the probability estimate. Id. at ES-5. But if the VNTRs frequencies for that subgroup are not available, data for the full population should be used with certain variations in the formula for estimating probability. Id.

In general, the 1996 NRC Report offers a strong endorsement of the DNA profiling process. "The technology for DNA profiling and the methods for estimating frequencies and related statistics have progressed to the point where the reliability and validity of properly collected and analyzed DNA data should not be in doubt." Id. at p. 2-14. Thus, the body that devised the ceiling principle in the first place has now signaled a retreat from its use.

These developments bode ill for the defendant's arguments on the effect of population substructure on VNTRs frequency distributions. These same arguments were considered and rejected in Penn, which found that the differences in VNTRs frequencies between blacks on the mainland and blacks in St. Thomas 32 were not forensically significant. Penn, 833 F. Supp. at 1070-71. "Any concern that the St. Thomas black population's bin frequencies are drastically different from those of the United States' black population is unwarranted." Id. at 1071. We make several additional points on this issue in light of Dr. Budowle's testimony at the gatekeeper hearing in this case and the findings of the 1996 NRC Report.

First, Dr. Budowle testified that he compared the VNTRs frequencies of Afro Caribbeans (from the Bahamas, Flaiti, and the French Antilles) with those of African Americans from the mainland United States and found only slight differences. Tr. at 80. These differences were not forensically significant, meaning that they did not affect the basic assumption that the bands on the autorad behave independently. Dr. Budowle testified that he also considered the effect of racial mixing on VNTRs frequencies by comparing African Americans from the South (where there was less racial mixing) with those from the North. Tr. at 75. Again, his results showed that racial mixing did not have a forensically significant effect on VNTRs frequencies. Tr. at 81. Moreover, the randomness of the individuals making up the FBI's Black database inherently accounts for racial mixing in the black population.

Second, as the 1996 NRC Report concluded, population subgrouping is important only if we know that the suspect is a member of a particular subgroup. All that was known about the suspect in this case was his race. The victims did not indicate whether he was a transplanted North American, a native St. Thomian, or an immigrant from one of the other Caribbean islands. As recommended by the 1996 NRC Report, the FBI's database for Blacks was used in comparing the defendant's DNA profile since the suspect's race is known in this case. Because investigators did not know the subgroup to which the suspect belonged, there was no need to compare the defendant's DNA profile with any subgroup. The FBI procedure of giving DNA frequency estimations for several different racial groups was more than adequate under the circumstances.

b. Estimating Error Rates

Another NRC recommendation that the FBI has declined to follow concerns the estimation of laboratory error rates.

Laboratory error rates should be measured with appro priate proficiency tests and should play a role in the interpretation of results of forensic DNA typing. . . . Proficiency tests provide a measure of the false-positive and false-negative rates of a laboratory. Even in the best of laboratories, such rates are not zero.

A laboratory's overall rate of incorrect conclusions due to error should be reported with, but separately from, the probability of coincidental matches in the population.

Both should be weighed in evaluating evidence.

1992 NRC Report at 94. Dr. Budowle testified that this recommendation is impractical for at least two reasons. First, the 1992 Report did not furnish a method for calculating these rates. Tr. at 241. 33 Second, the very process of observing and correcting errors makes it difficult to arrive at a historical error rate. Once the laboratory corrects for an error, it will no longer register as an error in future tests. Id. Recognizing these practical difficulties, the drafters of the 1996 NRC Report concluded that it was no longer feasible to estimate the likelihood of laboratory error. "Auditing and proficiency testing cannot be expected to give a meaningful estimate of the probability that a particular laboratory has made ... an error in a specific case. An unrealistically large number of proficiency tests would be needed to estimate accurately even a historical error rate." 1996 NRC Report at ES-4.

The FBI's laboratory procedures differ from the 1992 NRC Report recommendations in several other significant respects. 34 But after reviewing the 1992 NRC Report and evaluating the testimony of Dr. Bruce Budowle, we are convinced that none of these divergences — even on the issues of ceiling principle and error rate estimations — affect the basic reliability of the FBI's DNA profiling technique. The mere fact that there are two competing methods of arriving at a result does not necessarily mean that one of them is unreliable. The debate concerning the ceiling principle illustrates this point quite well. Both the ceiling principle and the product rule may be based on the methods of procedures of science; the fact that one method yields more conservative results does not automatically invalidate the less conservative method. Since Daubert focuses on principles and methods, not on conclusions, both may be admissible under Rule 702. The defendant's criticisms of the FBI's failure to adopt the NRC recommendations go mostly to the weight of the DNA profiling evidence, not to its admissibility.

Although we have made an independent assessment of the various Daubert factors, Penn's detailed analysis of these factors, as they relate to the FBI profiling technique, is highly persuasive, and we adopt them here. We write separately, however, to highlight instances where Dr. Budowle's testimony influenced our consideration of these factors.

a. Whether the Proffered Technique Can be (and Has Been) Tested

Dr. Budowle testified that the FBI profiling process has been tested repeatedly both by internal and external sources. The process of measuring the lengths of VNTRs is itself relatively simple and has been tested dozens of times every day by use in both forensic and nonforensic applications. Tr. at 46. Such widespread use in a variety of applications suggests that the process is sufficiently tested and found reliable.

b. Whether the Theory or Technique Has Been Subject to Review

Dr. Budowle's numerous publications on the FBI's DNA profiling process invites review from peers in the field of human genetics. Indeed, the act of "getting published" itself entails a certain element of review, since the editorial board of the journal must preliminarily determine whether the material to be published passes muster. Aside from Dr. Budowle's publications, however, the FBI has made its database and its findings available to scientists for independent review. Tr. at 135-36. An indication of the level of peer review that the FBI DNA profiling technique has received can be seen from the defendant's use of the published criticisms of other experts to cross-examine Dr. Budowle. See, e.g., Tr. at 153-54, 194-96.

c. The Existence and Maintenance of Standards Controlling the Technique's Operation

The FBI's various quality assurance standards and the built in conservatism in the DNA profiling process is adequately chronicled in Penn. Penn, 838 F. Supp. at 1066-73. Two other points bear mentioning. Dr. Budowle testified that the FBI lab is in the process of accreditation by the American Crime Directors Laboratory Accreditation Board. Tr. at 243. Accreditation of DNA profiling labs was one of the recommendation made by the NRC in 1992. 35 Second, the FBI lab voluntarily adheres to certain uniform technical standards and guidelines. Id. 36

d. The Known or Potential Rate of Error

Like the Sixth Circuit in Bonds, we too are concerned about the lack of any meaningful reference on the frequency of errors in the FBI's DNA profiling process. Bonds, 12 F.3d at 560. Samples can be contaminated by even microscopic quantities of another person's DNA. Knowledge of the frequency with which errors are made helps the trier of fact weigh the value of the evidence. Otherwise, the jury may be left with the impression that the technique is infallible. To some degree, the defendant may be able to elicit information about observed errors and the accuracy of the process during cross-examination. But observed error and error rates are two different concepts. Ultimately, this is but one factor of the reliability inquiry, and even though it weighs somewhat heavily against reliability, it is not dispositive of the issue.

e. Degree of Acceptance By a Relevant Scientific Community

Penn and Bonds' analyses of the degree of acceptance of DNA profiling evidence is persuasive. See Penn, 838 F. Supp. at 1073; Bonds, 12 F.3d at 562-65. Further discussion on this factor is unwarranted.

a. The Novelty of DNA profiling and its Relationship to More Established Modes of Scientific Analysis

According to Downing, this factor measures "the likelihood that the scientific basis of the new technique has been exposed to critical scientific scrutiny." Downing, 753 F.2d at 1239. Although scientists first discovered the DNA molecule in 1871, forensic DNA testing was not introduced until 1986.According to Downing, this factor measures "the likelihood that the scientific basis of the new technique has been exposed to critical scientific scrutiny." Downing, 753 F.2d at 1239. Although scientists first discovered the DNA molecule in 1871, forensic DNA testing was not introduced until 1986. 37 Indeed, VNTRs were not discovered until 1985. 38 Traditional dermatological fingerprinting, by contrast, has been used in courts since the turn of the century. 39 Although DNA profiling and traditional fingerprinting are often compared, the two processes are quite different. Unlike traditional fingerprinting, where two people cannot have the same fingerprint, DNA profiling does not purport to be able to definitively identify the defendant. Even if there is a perfect match between the defendant's DNA and the DNA taken from the crime scene, there is still the possibility that the sample at the crime scene came from a different person whose patterns at the targeted loci are indistinguishable from the defendant's. State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 486 (N.H. 1992).

DNA profiling has more in common with techniques that test for genetic markers in blood such as the human leukocyte antigen test ["HLA"] and the electrophoretic testing of dried blood stains. HLA testing is used in paternity cases to calculate the probability that the putative father is the actual father. See, e.g., In re Paternity of M.J.B., 144 Wis. 2d 638, 425 N.W.2d 404, 404-08 (Wis. 1988); Turek v. Hardy, 312 Pa. Super. 158, 458 A.2d 562, 563-64 (Pa. Super. Ct. 1983). Unlike traditional blood tests, HLA testing involves tissue typing of white blood cells and is used to include the defendant as a possible father. Electrophoretic testing is used in criminal trials to determine whether dried blood stains at the crime scene belong to the defendant. See, e.g., People v. Reilly, 196 Cal. App. 3d 1127, 242 Cal. Rptr. 496, 501 (Cal. App. Ct. 1988). Both of these procedures use statistical data to estimate the likelihood of inclusion; both have been admitted in courts of law. 40 Thus, the concept of extracting genetic information from samples and calculating statistical probabilities based on this data is not entirely new.

In truth, DNA profiling is actually an amalgam of several established scientific fields including chemistry, molecular biology, statistics, and genetics. Given our finding (under the second Daubert factor) that DNA profiling has been subjected to peer review and given the fact that this technique is similar to other established techniques and relies heavily on the methods and techniques of other proven sciences, this first Downing factor militates in favor of a finding of reliability.

b. The Qualification and Professional Stature of the Expert Witness

We need not tarry long on this factor. Dr. Bruce Budowle is among the foremost experts in the field of DNA profiling as evidenced by his many publications and lectures in this field and by his position as head of one of the leading forensic DNA labs in the world.

c. The Non-Judicial Uses to Which the Method Has Been Put

DNA profiling has many non-judicial applications. It has been used by pathologists to facilitate the identification of bodies. It has also been used to determine parentage and heirship to screen for genetic diseases, and to determine the pedigree of animals. These varied uses of DNA profiling weigh in favor of reliability.

Thus, seven of the eight factors identified by Daubert and Downing weigh in favor of reliability. We therefore find that the FBTs DNA profiling procedure is reliable under Rule 702.

The DNA evidence that the government wishes to introduce is clearly relevant to whether the defendant committed the crimes charged in the superseding information. Obviously, a match between the defendant's blood sample and the semen sample retrieved from the victims can assist the trier of fact in determining Fredericks' guilt or innocence on the rape counts. This evidence also bears on several of the other crimes charged in the information as well, since it would tend to place Fredericks at the scene of these crimes.

Our gatekeeper role also requires us to decide whether the facts underlying the FBI's DNA profiling evidence are of a type reasonably relied on by experts in the field. Paoli II, 35 F.3d at 748. We find that the data underlying the DNA evidence in this case, such as the use of anonymous samples from blood bank and paternity labs in composing population databases, the size of the match window (+ 2.5%), and the size and minimum frequencies of the FBI's fixed bins, meet the requirements of Rule 703.

E. Danger of Overwhelming, Confusing or Misleading Jury

Even though we have found that the DNA profiling evidence in this case passes muster under Rules 702 and 703, we nevertheless may exclude this evidence if it will unfairly prejudice the jury. Paoli II, 35 F.3d at 746; Downing, 753 F.2d at 1240. There is always the danger that any evidence touched by the mantle of science will hold inordinate sway in the minds of jurors. But this modern predilection towards believing all things scientific is not always enough to exclude novel scientific evidence. Paoli II, 35 F.3d at 746. Something more is required, such as when it has "assumed a posture of mythic infallibility" among lay persons. Downing, 735 at 1239. In our view, DNA profiling evidence comes perilously close to achieving that stature.

Much of this has to do with the nature of DNA itself. In the popular imagination, DNA retains the power to unlock life's genetic secrets. It can show us why we look the way we do. It can tell us whether we will fall prey to certain diseases; what our children will look like; and even whether we are kin to corpses long at rest in the grave. It can unravel mysteries from the past; 41 it can warn us of the possibility of trouble in the future 42 It can tell us more about ourselves and our heredity that we sometimes care to know. 43 Little can be hidden from this genetic witness. For the lay person, therefore, it is easy to conceive that this same science that can reveal the genetic secrets of the living, the dead, and the yet unborn is potent enough to solve the most perplexing crime. There is something very primal about DNA and genetic science that lends itself to a posture of "mythic infallibility." Of course, the inability in trial cases to provide the jury with meaningful error rates serves only to strengthen this posture.

We are also cognizant of the likelihood that DNA evidence may confuse or overwhelm the jury. The differences between DNA profiling and traditional modes of forensic analysis substantially increase this likelihood. Fingerprint, foot print, handwriting, and bite mark evidence, for example, are easily comprehensible to the jury because they are taken from everyday life. The least experienced juror has observed a fingerprint, footprint, or bite mark, on many different occasions long before he or she enters the courtroom. And the fact that handwriting tends to vary from person to person is evident from grade school. DNA profiling evidence, on the other hand, is totally alien to most jurors. The specter of bands on an autorad is no more familiar to the average juror than the innermost workings of the stealth bomber.

Also, unlike voice, handwriting, footprint, fingerprint or bite mark evidence, neither the DNA evidence nor the faceless laboratory technicians who performed the critical laboratory tests and procedures are present in the courtroom for independent credibility evaluation by the jury. DNA evidence is invisible to the naked eye, so bringing the used agarose gel to court would be pointless. In the words oí one appellate court, DNA evidence, "unlike that presented with fingerprint, footprint or bite mark evidence, is highly technical, incapable of observation, and requires the jury to either accept or reject the scientist's conclusions that it can be done." Andrews v. State, 533 So. 2d 841, 850 (Fla. Dist. Ct. App. 1988), rev. den., 542 So. 2d 1332 (Fla. 1989). The juror's almost complete dependence on the testimony of the expert and her inability to independently evaluate the accuracy of the DNA profiling results thus weighs against admissibility. 44

Another danger of admitting DNA profiling evidence under Rule 403 is that DNA evidence is easily misinterpreted. The statistical probability of a random match is too easily converted, by lawyers, judges and jurors alike, into the probability that it could not have been anyone other than the defendant who committed the crime. 45 This is especially true when the probability estimate is very low. Vanishingly small probabilities of a random match may tend to establish guilt in the minds of jurors and are particularly suspect.

Ultimately, however, whether DNA profiling evidence is unfairly prejudicial depends on the facts of the case, to which we now turn. For the semen recovered from one of the rape victims, M.S., the probability of a random match is actually rather high: 1/1000 for whites; 1/550 for blacks; and only 1/300 for Hispanics. The DNA frequency estimates for the semen recovered from the two other rape victims (1/300,000 for whites; 1/100,000 for blacks; and 1/35,000 for Hispanics) is still comparatively high. 46 The availability and probative value of the other evidence in the case also impact the Court's consideration of the prejudicial effect of DNA evidence. The Court is aware that the government will also attempt to introduce footprint evidence and eyewitness evidence from Fredericks' alleged accomplice. 47 It is very likely that the jury will give equal consideration to the other evidence in the case and not give inordinate weight to the DNA profiling evidence against Fredericks. Given the relatively high probability of a random match and the presence of other incriminating evidence against Fredericks, we find that the ever-present danger that DNA profiling evidence will unfairly prejudice the jury does not warrant exclusion here.

After evaluating the FBI's DNA profiling evidence under Rules 702, 703 and 403 and weighing the factors applicable to each Rule, we hold that the DNA profiling evidence presented in this case is admissible at trial. Though we have reservations about the overwhelming, confusing and misleading effect DNA profiling evidence is likely to have on jurors, the circumstances of this case presently do not justify exclusion on that basis. Accordingly, the government's motion in limine to admit this evidence will be granted. A separate order follows.

ENTERED this 11th day of October 1996.

ORDER

Upon consideration of the premises, the Court being duly advised, and for the reasons set forth in the Memorandum of even date, it is hereby

ORDERED that the government's motion in limine to admit DNA profiling evidence is GRANTED.

DATED this 11th day of October, 1996.

1 Reginald Sylvester Byers, Jr. is alleged to have embarked on a crime spree of his own after his ventures with Caswell Fredericks ceased. Byers' spree did not end until June 1991.

2 It would appear that the FBI did not compare Fredericks' blood sample with the semen recovered from the first victim.

3 National Research Council Committee on DNA Technology in Forensic Science, DNA Technology in Forensic Science (1992) (1992 NRC Report).

4 A court may take judicial notice of expert testimony that has been offered in earlier cases to support or dispute the merits of a particular scientific procedure. United States v. Downing, 753 F.2d 1224, 1239 (3d Cir. 1985).

5 Frye v. United States, 54 App. D.C. 46, 293 E 1013, 1014 (1923).

6 Rule 706 is not relevant here.

7 Rule 703 reads:

The facts or data in the particular case upon which an expert bases an opinion or inference may be those perceived by or made known to the expert at or before the hearing. If of a type reasonably relied upon by experts in the particular field in forming opinions or inferences upon the subject, the facts or data need not be admissible in evidence.

8 Rule 403 provides:

Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence.

9 See, e.g., United States v. Johnson, 56 F.3d 947 (8th Cir. 1995) (finding DNA evidence in car jacking admissible); United States v. Davis, 40 F.3d 1069 (10th Cir. 1994), cert. denied, 115 S. Ct. 1387 (1995) (affirming admission of DNA evidence in attempted armed robbery case); United States v. Chischilly, 30 F.3d 1144 (9th Cir. 1994), cert. denied, 130 L. Ed. 2d 890, 115 S. Ct. 946 (1995) (admitting DNA evidence against Navajo Lidian defendant in sexual abuse and murder case); United States v. Bonds, 12 F.3d 540 (6th Cir. 1993), reh'g denied, (1994) (admitting DNA evidence in conspiracy and firearms case); United States v. Martinez, 3 F.3d 1191 (8th Cir. 1993), cert. denied, 510 U.S. 1062, 126 L. Ed. 2d 697, 114 S. Ct. 734 (1994) (admitting DNA evidence in sexual abuse case); United States v. Coronado-Cervantes, 912 F. Supp. 497 (D.N.M. 1996) (admitting DNA evidence against Native American defendant in rape case); United States v. Galbreth, 908 F. Supp. 877 (D.N.M. 1995); Government of Virgin Islands v. Penn, 838 F. Supp. 1054 (D.V.I. 1993).

10 Paoli II, 35 F.3d at 741.

11 See, e.g., Bonds, 12 F.3d at 550; Chischilly, 30 F.3d at 1148; Johnson, 56 F.3d at 960.

12 Thus, if we know the base components of one strand of the DNA molecule, we also know the sequence of nucleotide bases of its complementary strand.

13 Sue Rosenthal, Note, My Brother's Keeper: A Challenge to the Probative Value of DNA Fingerprinting, 23 Am. J. Crim L. 195, 198 (1995).

14 1992 NRC Report at 3.

15 For example, the core sequence of VNTRs at segment 44 of chromosome 2, denoted as D2S44, may be "GTCA". Each individual has the same sequence of GTCA at D2S44; what varies from person to person is the number of times GTCA is repeated in a string before another sequence begins. The core sequence of VNTRs at segment 139 of chromosome 4, D4S139, may be TCA, repeated 100 times in one individual and 50 times in another.

16 Penn considers the collection of the sample from the crime scene as part of the laboratory stage, though it reasonably could be separated into a stage of its own. Although the gatekeeper hearing did not address the possible problems that may have occurred during the sample collection process, Fredericks preserved his right to challenge this process at a later date.

17 This process of comparing DNA fragments of different sizes are is known as Restriction Fragment Length Polymorphism ["RFLP"] analysis.

18 Dr. Budowle testified that, in some instances, the FBI increases the number of probes from four to six. Tr. at 85. The more VNTR probes used to analyze a person's VNTR pattern, the more distinctive that pattern will be and the lower the probability of a coincidental match.

19 From our earlier example, one probe may be "GTCA", the core sequence for VNTRs at segment 44 of chromosome 2, denoted as D2S44. Another radioactive probe may target VNTRs at a location denoted as D19S22, which is the 22nd segment of DNA on the 19th chromosome. Except for identical twins, no two people have VNTRs fragments of the same length at each chromosomal location.

20 See, e.g., David H. Kaye, DNA Evidence: Probability, Population Genetics, and the Courts, 7 Harv. J. L. & Tech. 101,104 (1993) ("Of all the technological and scientific issues in [the DNA] debate, the most difficult for the courts, and those that have generated the most disagreement within the scientific community, involve statistics.")

21 This assumption is premised on a theorem known as Hardy-Weinberg Equilibrium which holds, among other things, that independence can be assumed if there is random mating in the population.

22 The Product Rule can be illustrated using a deck of cards. If there are 52 cards in a randomly shuffled deck, the chances of drawing the ace of spades is 1/52. Once the card is replaced and the deck is reshuffled, the chances of drawing the ace of spades is again 1/52 because the act of drawing the ace of spades the first time does not affect the likelihood of drawing it a second time. The two events are said to be independent of each other. Assuming such independence, the Product Rule allows us to calculate the chances of drawing the ace of spades two consecutive times as (1/52 X1/52) orl/2704. See Tr. at 58.

23 1992 NRC Report at 149.

24 Richard C. Lewontin & Daniel C. Hartl, Population Genetics in Forensic DNA Typing, 254 Science 1745, 1750 (1991).

25 Hie import of Lewontin and Hartl's argument can be demonstrated by a simple analogy. If a population survey of Caucasian Europeans showed that 10% have blond hair, 10% have blue eyes, and 10% have fair skin, then — assuming that these physical traits behave independently, that is, there is no nonrandom factor linking them together, the frequency with which these traits occur in that population could be calculated as 1/1000 (1/10 X 1/10 X 1/10), using the Product Rule. Lewontin and Hartl argue that it would be misleading to apply this statistical method to a subgroup, such as Nordics, where these traits are not distributed independently but tend to be linked together. All things being equal, the frequency of blond, blue-eyed, fair-skinned people would be higher in Sweden than in Italy. Therefore, using the Caucasian European database and the Product Rule in Sweden would result in misleadingly low DNA frequency estimates. William C. Thompson, DNA Evidence in Criminal Law: New Developments, Trial, p. 36 (August 1994).

26 1992 NRC Report at 13.

27 Because of the disagreement over the validity of the product rule, several courts have either excluded DNA frequency estimates based on the product rule or required use of the ceiling principle. See, e.g., State v. Streich, 658 A.2d 38 (Vt. 1995) (admitting DNA frequency statistics only if ceiling principle used); State v. Bible, 175 Ariz. 549, 858 P.2d 1152, 1190 (1993), cert. denied, 128 L. Ed. 2d 221, 114 S. Ct. 1578 (1994) (finding DNA frequency estimations using the product rule inadmissible); Commonwealth v. Lanigan, 419 Mass. 15, 641 N.E.2d 1342, 1349-50 (Mass. 1994) (admitting DNA frequency statistics only if ceiling principle used); State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 494-95 (N.H. 1992) (excluding probability estimates based on product rule).

28 Federal Bureau of Investigation, VNTRs Population Data, A Worldwide Study (1993).

29 See, e.g., Bruce Budowle, Fixed-Bin Analysis for Statistical Evaluation of Continuous Distributions of Allele Data from VNTR Loci, for Use in Forensic Comparisons, 48 Am. J. Hum. Genet. 841 (1991); Ranajit Chakraborty, Statistical Interpretation of DNA Typing Data, 49 Am. J. Hum. Genet 895 (1991); Ranajit Chakraborty & Ken Kidd, The Utility of DNA Typing in Forensic Work, 254 Science 1735 (1991).

30 See, e.g., Kaye, supra n. 20 at 138-39; Richard Lempert, The Suspect Population and DNA Identification, 34 Jurimetrics J. 1 (1993); Tr. at 81.

31 Richard Lempert, DNA, Science and the Law: Two Cheers for the Ceiling Principle, 34 JuWEMETRICS J. 41 (1993).

32 Although Fredericks is actually from St. Kitts, tire suspect population is drawn from St. Thomas, because that is where the crimes took place.

33 The absence of procedures for implementing many of the NRC's recommendations is one of the FBI's main criticisms of the 1992 NRC Report. See, e.g., Tr. at 265.

34 For instance, there appears to be some divergence between the FBI and the 1992 NRC Report on the use of ethidium bromide (EB). Tr. 179-181. According to the defendant, use of EB may result in band shifting.

35 1992 NRC Report at 109.

36 See, e.g., Technical Working Group on DNA Analysis Methods ["TWGDAM"], Guidelines for a Quality Assurance Program for DNA Restriction Fragment Length Polymorphism Analysis, 16(2) Crime Lab Dig. at 40-45 (1989).

37 Congress of the United States, Office of Technology Assessment the Genetic Witness: Forensic uses of DNA Tests, 14 (1990) "Forensic DNA analysis has since been admitted into evidence in at least 185 cases by 38 States and the U.S. military as of January 1, 1990." Id.

38 See Alec J. Jeffreys et al., Hypervariable "Minisatellite" Regions in Human DNA, 314 Nature 67 (1985).

39 Gerald Lambourne, The Fingerprint Story 73 (1984).

40 See, e.g., Annotation, Admissibility, Weight, and Sufficiency of Blood Grouping Tests in Criminal Cases, 2 A.L.R. 4th 500 (1980); Randolph N. Jonakait, Will Blood Tell? Genetic Markers in Criminal Cases, 31 Emory L.J. 833 (1992).

41 In 1991, for example, scientist using the techniques of DNA profiling were able to unravel a 70-year-old mystery by establishing the identity of Czar Nicholas II, the last imperial ruler of Russia. See Associated Press, Royal DNA Matched with Russian Bones, The Boston Globe, December 11, 1992. Moreover DNA technology is increasingly being used to right past wrongs by casting doubt on criminal convictions many years old. See, e.g., Edward Connors, Thomas Lundregan, Neal Miller and Tom Mcewen, U.S. Dep't of Justice, Convicted by Juries, Exonerated by Science: Case Studies in the use of DNA Evidence to Establish Innocence After Trial 0une 1996); Larry Gerber, A "16-year nightmare" Comes to an End: DNA Tests Free Man Wrongly Imprisonedfor Attack on Wife, The Boston Globe, June 22, 1996.

42 Using the principles of genetics, scientist routinely screen for the presence of disease-causing genes. An expectant mother, for instance, can have her unborn fetus screened for certain debilitating diseases.

43 DNA analysis is used, for example, to detect genes for cystic fibrosis, fragile X syndrome, Huntingdon's disease, and various forms of muscular dystrophy and thalassemia to name only a few.

44 C.f. United States v. Ferri, 778 F.2d 985, 989 (3d Cir. 1985), cert. denied, 476 U.S. 1172, 90 L. Ed. 2d 983, 106 S. Ct. 2896 (1986) (admitting footprint evidence because it was capable of independent evaluation by the jury).

45 Jurors sometimes tend to confuse statistical evidence. One study found, for example, that jurors mistakenly reasoned that if a defendant's blood type is found in only ten percent of the population, there is a ninety percent chance that he is guilty. Thompson & Schumann, Interpretation of Statistical Evidence in Criminal Trials, 11 Human Behavior, 167, 170 (1987).

46 In Penn, for example, the probability of a random match was 1 in 41 million. Penn, 838 F. Supp. at 1065.

47 The accomplice, Reginald Sylvester Byers, Jr. has recently moved to withdraw his guilty plea and for new counsel to be appointed. If Byers is allowed to withdraw his plea, the Court may be required to reassess this Rule 403 analysis.