Opinion
Nos. 92-556 to 92-558.
June 1, 1994. Rehearing Denied August 25, 1994.
Appeal from the Circuit Court, Clay County, William A. Wilkes, J.
William J. Sheppard of Sheppard and White, and Michael R. Yokan, Jacksonville, for appellant.
Robert A. Butterworth, Atty. Gen., Gypsy Bailey and Giselle Lylen Rivera, Asst. Attys. Gen., Tallahassee, for appellee.
This is an appeal of judgments and sentences for multiple counts of burglary and sexual battery in three consolidated cases. Appellant was initially arrested on January 15, 1990, as he allegedly was attempting to break into a ground floor apartment. Based on similarities between this attempted burglary and two previous burglaries in which sexual batteries had been committed, a search warrant was issued to take a sample of appellant's blood for the purpose of conducting DNA profile testing. Based on the results of this testing, appellant was charged with the other burglaries and sexual batteries, to which he eventually pled nolo contendere, reserving the right to appeal the rulings on his motions.
On appeal, appellant raises four issues, contending the trial court erred in: (1) denying his motion to suppress evidence challenging the facial validity of the warrant pursuant to which his blood was drawn for analysis; (2) finding that alleged misrepresentations and omissions in the affidavit supporting the issuance of the search warrant did not entitle him to a hearing or require suppression of evidence; (3) denying suppression of evidence allegedly seized by an officer operating outside his territorial jurisdiction; (4) denying appellant's motion in limine as to the admission of alleged novel scientific evidence. We find no error as to the first two issues and affirm as to them. As to the third issue, we affirm but certify the question involved as one of great public importance. As to the fourth issue, we vacate the order on the motion in limine and remand for further proceedings, and certify the question as one of great public importance.
Appellant pled nolo contendere to a total of six counts of burglary and sexual battery involving three victims. He was sentenced to 22 years on the burglary counts and 15 years on the sexual battery counts, to be served concurrently. The parties stipulated below that the rulings on each of appellant's motions to suppress and the motion in limine were dispositive of the case.
The state argues on appeal that the rulings on the motions to suppress and motion in limine are not dispositive of the case despite having entered into the stipulation below. During the plea proceedings, the assistant state attorney stated that "if the court were to have granted any of those motions, that would have been dispositive of this case." In Zeigler v. State, 471 So.2d 172 (Fla. 1st DCA), review denied, 479 So.2d 118 (Fla. 1985), this court held "that a stipulation voluntarily entered into by all parties that an issue preserved for appeal by a defendant's nolo contendere plea will be so considered by this court."
After a hearing, the trial court denied appellant's motion to suppress in which he argued the officer who served the search warrant for the blood sample was outside his territorial jurisdiction when he did so, and thus in violation of section 933.08, Florida Statutes. There were two search warrants directed to the sheriff and deputy sheriffs of Duval County; one directed to appellant's body and one for his home. Officer Harris, a Clay County officer who was involved in the investigation of this case, went to Cecil Field in Duval County to pick up appellant. At the hearing, Officer Harris said when he initially picked appellant up, he gave appellant the option of going with him from Cecil Field to his home or having a Duval County officer come and get him, and that appellant went with him voluntarily. When confronted with his deposition testimony indicating he had read the warrant to appellant at Cecil Field, Harris said he could not recall having done so. He said he did not read the warrant to appellant until they were at appellant's home in Duval County, and then he read the warrant to appellant "in the authority of" Officer Baer, a Duval County sheriff's deputy. Officer Harris had arranged for Officer Baer to meet him and appellant at appellant's home, where both officers conducted a search. Harris testified further that appellant rode with him to the Sexual Assault Treatment Center for the blood to be drawn, and again Officer Baer met them there. Harris was in the room with appellant while the blood was being drawn; Baer was outside the room.
Section 933.08 provides:
Search warrants to be served by officers mentioned therein. — The search warrant shall in all cases be served by any of the officers mentioned in its direction, but by no other person except in aid of the officer requiring it, said officer being present and acting in its execution.
Appellant has discussed the following authorities in support of his argument that the search warrant directed to officers of Duval County was invalidly executed by an unauthorized Clay County officer: Morris v. State, 622 So.2d 67 (Fla. 4th DCA 1993); State v. Griffis, 502 So.2d 1356 (Fla. 5th DCA), review denied, 513 So.2d 1063 (Fla. 1987); Hesselrode v. State, 369 So.2d 348 (Fla. 2d DCA 1979), cert. denied, 381 So.2d 766 (Fla. 1980). In Morris, the district court concluded there had been an unauthorized search conducted entirely by the Auditor General's staff, who were not authorized to search pursuant to the Auditor General's investigative authority. A Fort Lauderdale police officer, who was authorized by search warrant, accompanied the staff members to the physician's office, but did not take part in, observe, or supervise the search; he waited in another room while the staff searched the physician's files. He signed an inventory sheet without verifying its accuracy or determining what had been seized.
The court in Morris noted that "[u]nder section 933.08, the persons authorized in a warrant to conduct a search and seize the items described must actually execute the warrant and conduct the search," 622 So.2d at 68, citing Hesselrode v. State, 369 So.2d 348 (Fla. 2d DCA 1979), cert. denied, 381 So.2d 766 (Fla. 1980). The court was concerned with the complete lack of participation in the search by the authorized police officer, stating:
Under the statute, the officer authorized by the warrant to conduct the search and seize the evidence designated must participate in or supervise the search even where he requires the assistance of others to do so. While the level of supervision and participation may vary depending upon the circumstances, it is absolutely essential that the officer authorized be present when and where the search is conducted and carry out his responsibility to see that the warrant is properly executed and that its authorization is not exceeded. It is not enough that the authorized officer wait in another room while the search is conducted by others.
In Griffis, a search warrant directed to Titusville police officers and Brevard County sheriff's deputies was executed by a Titusville police officer in the city of Cocoa, in Brevard County, outside his authorized territory. The court noted that, had a Brevard County deputy accompanied him it would have saved the search; although a Cocoa police officer was present, the search warrant was not directed to him. And in Hesselrode, a search warrant directed to Manatee County sheriff's officers was invalidly executed by Longboat Key police officers, who had applied for the warrant. According to the court, any Manatee officers present were mere passive observers "at best"; all of the investigative work had been done by the Longboat Key police, and they alone conducted the search and seizure. The court noted that the warrant could have named the Longboat Key officers, but did not.
Appellant contends the cases discussed above, particularly Morris, require a determination that the search warrant in the instant case was invalidly executed in that, although an authorized officer, Officer Baer, was present when Officer Harris read the warrant to appellant, Officer Harris drove him to the hospital where the blood was to be drawn, took custody of the blood sample, and signed the inventory. Officer Baer was also present at the hospital, although he was not present in the room when the blood was drawn. In this case, however, neither Harris nor Baer conducted the search at issue in the traditional sense; they both accompanied appellant to the hospital and were present while medical personnel collected the blood sample. On this record, we are not willing to say Officer Baer was simply "out and about the scene" as the Manatee County officers apparently were in Hesselrode, or that he did not in any way act in execution of the search warrant. Recognizing that the warrant named only Duval County officers, Officer Harris had made specific arrangements for Officer Baer to be present both at appellant's home where the search warrant was read and at the Sexual Assault Treatment Center where the blood was to be drawn. Finally, unlike the situation in Morris, where the actual search was conducted by Auditor General employees, only law enforcement officers were involved in the execution of the search warrant here, and unlike the situation in Griffis, an officer named in the search warrant was present. Therefore, we affirm the trial court's ruling on this motion to suppress. However, we certify the following question:
WHETHER A SEARCH WARRANT FOR A BLOOD SAMPLE IS PROPERLY SERVED AND EXECUTED IN THE PRESENCE OF AN OFFICER WHO IS WITHIN THE TERRITORY NAMED IN THE SEARCH WARRANT, WHEN THE OFFICER WHO READS THE WARRANT TO THE ACCUSED, TRANSPORTS THE ACCUSED TO THE HOSPITAL FOR THE BLOOD TEST, AND TAKES CUSTODY OF THE BLOOD SAMPLE, IS NOT WITHIN THE SCOPE OF THE WARRANT?
Appellant's final issue on appeal challenges the admissibility of DNA profile evidence. The specific basis for the challenge in this case involves the final stage of DNA profiling, when, after it is determined the defendant's DNA "matches" the DNA found in the sample taken from the crime scene, population statistics are applied to determine the likelihood that someone else in the population might also "match" the DNA found at the crime scene. Appellant, who is undisputedly of Puerto Rican descent, argues he is a member of a subgroup as to whom the data bases compiled by the FBI, and used by FDLE, to determine the probability that someone else in the population would also match, are not generally accepted in the scientific community as reliable for use at criminal trials. The specific reason for this alleged unreliability is the possibility of a substructure within ethnic groups whereby certain alleles would occur more frequently within that population, and this frequency is not reflected in the FBI data bases. As a result, appellant contends, the FBI data bases are not generally accepted as reliable for use in criminal trials as to such a defendant.
For an explanation of how distribution of alleles figures into the FBI's statistical calculations, and the racial/ethnic subgroup argument, see Commonwealth v. Lanigan, 413 Mass. 154, 596 N.E.2d 311, 315 (1992). "Population substructure, therefore, presents the possibility that using allele frequencies of larger population groups might produce an inaccurate frequency estimate for members of substructure groups." Id.
After an extensive hearing at which both sides offered expert testimony, the trial court ruled that: based on evidence of its admission in numerous criminal trials in many states including Florida, DNA evidence is not novel; the court would not need to further consider the standards of proof for admission of novel scientific evidence, but it would do so in order to explain that each standard has been satisfied; the standard for admitting novel scientific evidence in Florida is whether the evidence is relevant and reliable, based on Andrews v. State, 533 So.2d 841 (Fla. 5th DCA 1988), review denied, 542 So.2d 1332 (Fla. 1989), and Martinez v. State, 549 So.2d 694 (Fla. 5th DCA 1989); even if the court applied the test set forth in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923) and Stokes v. State, 548 So.2d 188 (Fla. 1988), DNA testing procedures and the application of population genetics to the test results are generally accepted as reliable in the scientific community, based on United States v. Jakobetz, 955 F.2d 786 (2d Cir. 1992); all of the experts who testified agreed DNA testing procedures and the application of population genetics to the test results are generally accepted in the scientific community; the dispute among experts addresses the sufficiency of the data bases used to calculate the probability that someone else in the population would have the same DNA profile as defendant; this dispute goes to the weight rather than the admissibility of the evidence.
Appellant contends the trial court erred in finding the evidence is not novel; in applying the relevance and reliability standard rather than the Frye general acceptance standard; in finding that the procedure is generally accepted in the relevant scientific community; and in determining that the dispute over the sufficiency of data bases goes to weight rather than admissibility of the evidence.
Appellant first argues, and we agree, that he is entitled to challenge the DNA profile evidence as novel scientific evidence because he complied with the requirement set forth in Correll v. State, 523 So.2d 562, 567 (Fla.), cert. denied, 488 U.S. 871, 109 S.Ct. 183, 102 L.Ed.2d 152 (1988), by making "a timely request for such an inquiry supported by authorities indicating that there may not be general scientific acceptance of the technique employed," even though the same type of evidence has already been received in a substantial number of Florida cases. Of the few Florida appellate decisions that discuss the admissibility of DNA profile evidence, none have considered a challenge to the adequacy of the data bases used to calculate the probability that someone other than the defendant might have the same DNA "fingerprints" as defendant.
In Andrews, the Fifth District Court of Appeal ("DCA") held that DNA profile evidence was properly admitted in a criminal trial. In that case, the court expressed doubt about the appropriate standard for admissibility, ultimately applying the relevancy/reliability approach rather than the Frye general acceptance analysis, but noting that the evidence would meet the Frye standard as well as the relevancy standard. But the controversy surrounding the data bases used in the probability calculations has arisen largely since Andrews was decided, and was not discussed in that case.
In Andrews, the court discussed appellant's argument that "the data base of 710 samples is too small to be statistically significant," and noted that "[t]he only evidence in the case supports the statistical value of the randomly selected samples." 533 So.2d at 850. The defense did not put on expert testimony in Andrews, although the defense did "skillfully and thoroughly" cross-examine the state's experts; the court concluded "[t]he scientific testimony indicates acceptance of the testing procedures," and that "evidence derived from DNA print identification appears based on proven scientific principles." The particular problem with DNA print identification at issue in the present case was not at issue in Andrews.
In Martinez, the issue presented was whether "the trial court erred in ruling admissible statistical data to show the probability of matching Martinez's DNA configuration with DNA material recovered from clothing of the victim of the crime." The defense argument was that "the expert's testimony was so overwhelming — one out of 234 billion — that if believed by the jury, it would establish Martinez's identity as the rapist beyond a reasonable doubt, and invade the province of the jury as the ultimate finder of facts." Again, the sufficiency of the data bases used to calculate the probability of someone else sharing appellant's DNA profile was not in controversy; the issue raised involved "the prejudicial effect of using such statistical data because of its conclusiveness. . . ." The Fifth DCA specifically noted that "[d]efense counsel in this case did not attempt to challenge or refute the scientific basis for the Lifecodes DNA print analysis, nor the mathematical and factual basis for the Hardy-Weinberg equilibria formula, upon which the statistical evidence was based."
In Robinson v. State, 610 So.2d 1288, 1291 (Fla. 1992), the court found no reversible error or abuse of discretion in the admission of DNA test results. The court noted that "Robinson did not . . . produce anything that questioned the general scientific acceptance of the testing." In the instant case, the defense made an extensive challenge below to the data bases upon which the probability calculations are based, putting on expert testimony and proffering extensive documentation to support its position. Thus we are faced with a new challenge to the admissibility of DNA profile evidence.
The analysis of admissibility of DNA profile evidence in Toranzo v. State, 608 So.2d 83 (Fla. 1st DCA 1992), was based entirely on Robinson, Martinez, and Andrews. The issue raised in the instant case was not raised there.
Appellant asserts that the correct standard for admission of the DNA profile evidence is whether it is generally accepted in the relevant scientific community, not simply whether it is relevant and reliable. Recently, in Flanagan v. State, 625 So.2d 827 (Fla. 1993), the supreme court said that in Florida, the correct standard for admitting expert profile testimony which relies on some scientific principle or test is the general acceptance standard established in Frye. Appellant further contends that, in determining whether the Frye test has been met, i.e., whether the scientific test is generally accepted in the relevant scientific community, this court should conduct a de novo review, whereas the state contends the standard of review is whether the trial court abused its discretion in determining the evidence is admissible. Our review of pertinent cases indicates that the correct manner of review is a de novo review of whether the evidence in question is generally accepted in the relevant scientific community, encompassing expert testimony, scientific and legal writings, and judicial opinions. See Flanagan v. State, 625 So.2d 827 (Fla. 1993) (reviewing relevant academic literature and case law in analyzing whether sexual offender profile evidence is generally accepted in the scientific community, and referring to Judge Ervin's concurring and dissenting opinion in Flanagan v. State, 586 So.2d 1085, 1112-20 (Fla. 1st DCA 1991) for its discussion of general acceptance); and Stokes v. State, 548 So.2d 188 (Fla. 1989) (court conducted de novo review of scientific research and literature to assess general acceptance of hypnotically refreshed recollection of witness to crime).
Appellee submitted a notice of supplemental authority, citing Daubert v. Merrell Dow Pharmaceuticals, ___ U.S. ___, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), in which the Court said the relevancy test set forth in the federal rules of evidence supersedes Frye as the standard for admission of expert testimony. The Florida Supreme Court noted Daubert in its opinion in Flanagan, but said that in Florida, Frye is the test for admissibility of scientific opinions.
To understand the significance of the expert testimony, it is necessary to have some basic understanding of DNA profile analysis. We have found the accounts of the process set forth in the case law instructive, and will not attempt to set forth a detailed explanation here. See e.g., Andrews, 533 So.2d at 847; State v. Vandebogart, 136 N.H. 365, 616 A.2d 483, 485-488 (1992); United States v. Porter, 618 A.2d 629, 631-33 (D.C.App. 1992); People v. Barney, 8 Cal.App.4th 798, 10 Cal.Rptr.2d 731, 734-37 (1st Dist. 1992). "DNA profiling can inculpate a criminal suspect by comparing the suspect's genetic material with genetic material obtained from human tissue left at the crime scene." Vandebogart at 486.
Stated very simply, there are three general steps in DNA testing:
1. Creating a DNA "print" or "profile" of a sample;
2. Determining whether the prints or profiles of different samples match; and
3. If samples match, computing the probability of a random match.State v. Bible, 175 Ariz. 549, 858 P.2d 1152, 1180 (1993), citing National Research Council, DNA Technology in Forensic Science (1992) ("NRC report"). A technique called restriction fragment length polymorphisms, or RFLP, analysis is used in determining whether there is a match between the suspect's DNA sample and the sample obtained from the crime scene. Vandebogart at 486. There appears to be little question that the technology used in determining whether there is a match, at step two above, is generally accepted in the scientific community. See Porter at 636, Barney at 739.
In the initial brief, appellant also asserted that the FBI's method of determining a match is not generally accepted in the relevant scientific community, based on Dr. Pollock's testimony that the use of a "5% window" of variation for proclaiming a match is not generally accepted because the consensus is that each laboratory should develop its own criteria. We are not persuaded by this argument in this case. In People v. Barney, 10 Cal.Rptr.2d at 739, the court noted that the National Research Council report "does not equate the absence of a standardized rule with a lack of general acceptance as to the matching step of DNA analysis." We also note that defense expert Dr. Wakeland testified that "the current controversy in the scientific community totally deals with the calculation of statistical probabilities." The record and materials submitted for our review indicate that the heart of the controversy in this case is the effect of population subgroups on the calculation of statistical probabilities, the third step in the DNA profiling process.
To determine the statistical probability of a random match, at step three above of the process, testing laboratories, in this case the FBI, use data bases they have compiled for Caucasian, Black, and Hispanic populations. Vandebogart at 488, Barney at 736. This step involves a comparison of the frequency with which certain DNA fragment patterns appear in the "target population," i.e., Caucasians, Blacks, or Hispanics, together with a series of mathematical calculations, Barney at 737. The current controversy, as presented to this court, centers around the construction and validity of these data bases.
This explanation of the process of determining the significance of a match is extremely simplified. For a detailed explanation, see Barney at 734-37.
One of the assumptions underlying the use of these data bases is that "members of the racial groups represented by the broad data bases — Caucasians, Blacks, and Hispanics — mate within their groups at random, i.e., without regard to religion, ethnicity, and geography." Barney at 740. The focus of the debate is the claim that certain ethnic subgroups exist for which this random mating assumption is not valid, so that "as a result, the subgroups may have substantial differences in the frequency of a given DNA fragment . . . identified in the processing step of DNA analysis." Barney at 740. Stated somewhat differently:
[t]he population genetics which are being utilized at the present time by many laboratories make the simplified assumption that all populations are homogeneously mixed without significant subgroups occurrences. Such an assumption is probably not correct. This is particularly true with regards to some segments of the African-American and Hispanic society. In order to apply Hardy-Weinberg equilibrium calculations, a homogeneous mixed population is necessary.
Mayersak, DNA Fingerprinting Problems for Resolution, 36 Medical Trial Technique Quarterly 441 (1990); see also Lewontin and Hartl, Population Genetics in Forensic DNA Typing, 254 Science 1745 (1991); Thompson Ford, Is DNA Fingerprinting Ready for the Courts, New Scientist, March 31, 1990, at 43. The alleged consequences "of genetic substructuring and subgroup differences in allele [DNA fragment] frequencies" are: "(1) it is inappropriate to use broad data bases to which all Caucasians, Blacks, and Hispanics may be referred for estimating frequencies, and (2) it is inappropriate to multiply frequencies together, for want of linkage equilibrium." Barney at 740. The assumptions of Hardy-Weinberg equilibrium and linkage equilibrium arguably are not valid for these subgroups. Id.
For explanations of linkage equilibrium and of how the lack of Hardy-Weinberg equilibrium affects linkage equilibrium, see Note, The Dark Side of DNA Profiling: Unreliable Scientific Evidence Meets the Criminal Defendant, 42 Stanford L.Rev. 465, 491 (January 1990).
Expert Testimony
In the present case, the state and defense each put on two expert witnesses during the hearing on the motion in limine. For the state, Dr. James M. Pollock, a forensic serologist for FDLE, and Dr. Martin Tracey, professor of biological sciences at Florida International University, testified. For the defense, Dr. Leslie Sue Lieberman, professor of anthropology and pediatrics at the University of Florida, and Dr. Edward Kittredge Wakeland, professor of pathology and laboratory medicine at the University of Florida, testified.
Dr. Pollock specializes in DNA analysis at FDLE. He received his training in DNA analysis from the FBI Forensic Research and Training Center in Quantico, Virginia. He has also attended seminars in DNA analysis, and has been a member of a technical working group sponsored by the FBI which meets three or four times a year at the FBI research and training center. Dr. Pollock said in his opinion, RFLP analysis, used by the FBI and FDLE, is generally accepted in the scientific community as reliable. Dr. Pollock explained in detail the RFLP process, and said he conducted the tests in each of the three cases involved here. In each case, appellant's DNA profile matched the DNA taken from the crime scene.
On conducting the population frequency portion of the analysis, Dr. Pollock used the FBI series of data bases, which he opined are generally accepted by members of the scientific community as being reliable. He obtained probabilities of one in 30 million in two cases pertaining to appellant, and one in 60 million in the third case pertaining to appellant (using an updated data base) of finding an unrelated individual having the same DNA profile.
On cross-examination, Dr. Pollock said it was generally accepted for most cases that substructure does not affect the population frequency calculations. He indicated he uses all three data bases, for Caucasians, Blacks, and Hispanics, in calculating the population frequencies, because often you don't know the suspect's background when conducting the tests; he called this the general population approach, which he said was the most conservative approach. He said there is no controversy in the scientific field over whether a profile for a member of a diverse ethnic group can be calculated using present data bases if one uses a general population frequency. He acknowledged that current scientific articles dispute his conclusion about ethnic substructure affecting frequencies, but said the individuals who raised the dispute were not forensic scientists. He stated DNA analysis has supporters outside the forensic community as well. Dr. Pollock acknowledged there could be substructure in isolated population groups, such as on Indian reservations or South Pacific islands. He had never studied the population history of Puerto Rico.
At oral argument, appellee suggested that the relevant scientific community in the instant case is limited to forensic scientists. We do not believe the relevant community is so narrow. See for example U.S. v. Porter, 618 A.2d 629, 634 (D.C.App. 1992), quoting the trial court to the effect that the relevant scientific community includes "`those whose scientific background and training are sufficient to allow them to comprehend and understand the process and form a judgment about it,'" and noting that there is no question that forensic scientists accept the evidence as reliable for use in criminal trials.
Dr. Tracey testified as an expert in the field of molecular biology and population genetics. He had worked and done research in RFLP analysis, and had reviewed work done at the FDLE lab, and said it is widely accepted in the general scientific community as a reliable testing method. Dr. Tracey had reviewed the analysis done in the three cases involved here, and he agreed with Dr. Pollock's conclusions. He stated the process of applying population genetics to DNA identification to ascertain the probability of the sample coming from someone other than the suspect is generally accepted in the scientific community.
When asked on cross-examination whether there is a great deal of controversy in the scientific community about whether substructure, especially among Hispanics, affects the accuracy of the FBI's probability statistics, Dr. Tracey said "there is a great deal of argument about whether . . . two data bases for subpopulations, principally within ethnic or within racial groups, are adequate to calculate accurate statistics." He opined that "there is a good deal of disagreement, principally because people are asking different questions. . . . [I]f you use a Mexican or Nicaraguan or Cuban or Puerto Rican data base, you are likely to get different numbers. Those numbers may differ — as much as in order of magnitude, it could go from 1 in a billion, down to 1 in millions. It would not reduce the number to 1 in 10 or 1 in 100s." He prefers to use a human data base because "it's impossible to equate the ethnicity or race of a suspect with a perpetrator until you're done."
Asked whether there is a significant controversy, considering recent publications questioning the data bases, Dr. Tracey said there was significant misunderstanding as to the appropriate questions to ask. Dr. Tracey acknowledged a data base compiled in Canada for Indians living in Ontario due to concern over the population frequency statistics, stating he was working on a similar study himself. He said the alleles differ, but the odds do not quantitatively differ. He said the solution to the substructuring problem suggested by Hartl and Lewontin was to use the highest frequencies; he acknowledged that was not the method used here, but he could sit down and calculate the frequencies using that method, and it would not change the qualitative answer.
See discussion and citations to scientific and legal writings, infra.
Dr. Lieberman was qualified as an expert in human evolutionary biology, biomedical anthropology, and the phenotypical makeup of the Caribbean population. She stated that for a variety of sociological, geographical, and historical reasons, based on her research and experience, she would not expect the Puerto Rican population to be in Hardy-Weinberg equilibrium.
Dr. Lieberman said she understood the FBI Hispanic data base was drawn from Texas, which would be largely Mexican-American Hispanics, and Miami, which would be largely Cuban Hispanics. She said she believed there would be greater variation in island populations. She said she believed the FBI data bases were generally accepted by forensic scientists, but that does not have anything to do with whether the Puerto Rican population is in Hardy-Weinberg equilibrium. She added that scientists outside the FBI might have different opinions. She acknowledged the controversy reflected in certain published scientific articles by Hartl and Lewontin and Chakraborty. Dr. Lieberman had not done any DNA analysis in a criminal context.
Dr. Wakeland testified as an expert in molecular biology, population genetics, and molecular genetics of polymorphisms. He had reviewed materials provided by FDLE as well as pertinent scientific articles. He stated the existing controversy in the scientific community relates "totally" to the calculation of the probability that someone else in the population could also match the crime scene DNA sample, because the frequency tables are poorly defined and poorly constructed. He said the substructuring issue is being hotly debated among population geneticists at this point. Asked whether DNA profiling is generally accepted among the scientific community of molecular biologists and population geneticists as reliable for use in criminal trials, Dr. Wakeland said at this point it is not clear that the calculated frequencies are accurate and there's a "tremendous amount of disagreement about that fact." He stated that "in order to make the probability calculations, one must assume both Hardy-Weinberg equilibrium and linkage equilibrium within the data base used to calculate the probability," this in turn requires the assumption that there is no population substructure, and the issue of substructuring as it relates to Hardy-Weinberg equilibrium "currently is being very heavily debated among population geneticists." On cross-examination, Dr. Wakeland said he was not an expert in forensic DNA analysis, and he had not done any forensic work. He agreed that the concept of applying population genetics to DNA profiles was widely accepted within the scientific community, and that a large number of labs use the FBI data bases.
Scientific and Legal Writings
Both appellant and appellee have cited a number of articles indicating the existence of a significant dispute in the scientific community as to the reliability of the FBI population frequency data bases due to the possible existence of subgroups along racial or ethnic lines. The majority of articles cited have been reviewed in the case law to be discussed below, and we will not review them again here. See Barney, 10 Cal. Rptr.2d at 740-41, and Porter, 618 A.2d at 636-38. "Our task . . . is not to choose sides in this dispute over the reliability of the statistical calculation process." Barney, 10 Cal. Rptr.2d at 743.
Appellant refers to the following publications in his brief: Note, The Dark Side of DNA Profiling: Unreliable Scientific Evidence Meets the Criminal Defendant, 42 Stanford L.Rev. 465 (1990); Lewontin and Hartl, Population Genetics in DNA Typing, 254 Science 1745 (1991); Mayserak, DNA Fingerprinting Problems for Resolution, 36 Medical Trial Technique Quarterly 441 (Summer 1990); Thompson Ford, DNA Typing: Acceptance and Weight of the New Genetic Identification Tests, 75 Vir.L.Rev. 45 (1989); and Thompson Ford, Is DNA Fingerprinting Ready for the Courts? New Scientist, March 31, 1990. Appellee refers to Risch Devlin, On the Probability of Matching DNA Fingerprints, 255 Science 717, February 7, 1992; and Imwinkelried, The Debate in the DNA Cases Over the Foundation for the Admission of Scientific Evidence: The Importance of Human Error as a Cause of Forensic Misanalysis, 69 Wn. U.L.Q. 19 (1991).
Judicial opinions
Appellant has directed our attention to a number of recent decisions from other jurisdictions questioning the general acceptance of DNA profile evidence because of inaccuracies in the data bases. See, e.g., State v. Bible, 175 Ariz. 549, 858 P.2d 1152 (1993); People v. Wallace, 14 Cal.App.4th 651, 17 Cal.Rptr.2d 721 (1 Dist. 1993); State v. Anderson, 115 N.M. 433, 853 P.2d 135 (App. 1993); Commonwealth v. Lanigan, 413 Mass. 154, 596 N.E.2d 311 (1992); State v. Vandebogart, 136 N.H. 365, 616 A.2d 483 (1992); People v. Barney, 8 Cal.App.4th 798, 10 Cal.Rptr.2d 731 (Cal.App. 1st Dist. 1992); United States v. Porter, 618 A.2d 629 (D.C.App. 1992); Commonwealth v. Curnin, 409 Mass. 218, 565 N.E.2d 440 (1991). Appellee directs our attention to United States v. Jakobetz, 955 F.2d 786, 800 (2d Cir. 1992) cert. denied ___ U.S. ___, 113 S.Ct. 104, 121 L.Ed.2d 63 (1992), arguing the issue goes to the weight rather than the admissibility of the evidence.
In State v. Bible, the Arizona Supreme Court, reviewing a trial court ruling that DNA profile testing was generally accepted in the scientific community under Frye, noted that "Cellmark, Lifecodes Corporation, and the FBI are the three major laboratories currently performing RFLP DNA testing in the United States. . . . Testing protocols for these laboratories are not identical." 858 P.2d at 1180. Accord State v. Anderson, 115 N.M. 433, 853 P.2d 135, 142 (App. 1993) ("Cellmark and Lifecodes laboratories use different data bases than the FBI."). Since the specific challenge in the present case involves only the FBI data bases, we have confined our review of the cases to those involving the FBI data bases analyzed under the Frye standard.
In People v. Wallace, the California court of appeal determined evidence of DNA analysis by the FBI was inadmissible under the Frye test of general acceptance in the relevant scientific community. Noting the existence of a "`bitter' and `raging' debate among population geneticists concerning the reliability of the process used to calculate the statistical significance of a match of DNA patterns. . . ." and that "[i]n the NRC report, a panel of experts drawn from a broad spectrum of the relevant scientific community has agreed there is a `substantial controversy' on this point," the court concluded that "[a]ny lay observer capable of comprehending the technique of DNA analysis is capable of perceiving the lack of general acceptance." 17 Cal.Rptr.2d at 725. Responding to the People's argument that a majority of experts accepted DNA profiling as reliable and that this majority was substantively correct, the court said:
the People persist in miscomprehending the issue before us. As we said in Barney, "the point is not whether there are more supporters than detractors, or whether . . . the supporters are right and the detractors are wrong. The point is that there is disagreement between two groups, each significant in both number and expertise." . . . Our inquiry must end with the perception of such disagreement and consequent lack of general acceptance.
17 Cal.Rptr.2d at 726. Ultimately, the court concluded the error in admitting the DNA evidence was harmless given the overwhelming evidence of guilt in the case.
In State v. Anderson, 115 N.M. 433, 853 P.2d 135 (Ct.App. 1993), analyzing general acceptance under Frye, the court noted that "[a] scientific principle must have the support of a clear majority of scientists. . . . If there is a significant body of scientific thought opposing the principle, there is no clear majority." 853 P.2d at 138. Noting that "[o]n balance, it appears that the cases that have considered the FBI's methods under the Frye test and have carefully reviewed an extensive record have refused to admit the DNA evidence," 853 P.2d at 144, and after reviewing the record, case law, and learned writings, the Anderson court also rejected the evidence due to flaws in the FBI population frequency statistics. "What appears to be of particular concern to many scientists is the FBI's use of a limited data base for comparison of the suspect's sample with that of others," and "the absence of reliable subpopulation data bases." 853 P.2d at 145, 146. The court quoted United States v. Yee, 134 F.R.D. 161, 181 (N.D.Ohio 1991): "[w]ithout the probability assessment, the jury does not know what to make of the fact that the patterns match. . . ." The court was careful to point out that it was only ruling "on the FBI's DNA analysis in the context of current scientific thought." 853 P.2d at 147.
In Commonwealth v. Lanigan, 413 Mass. 154, 596 N.E.2d 311 (1992), the court applied the Frye test in evaluating the FBI's DNA profile analysis of Lanigen. (Cellmark had performed the analysis as to other appellants.) The flaw alleged involved the population substructure theory, which "presents the possibility that using allele frequencies of larger population groups might produce an inaccurate frequency estimate for members of substructure groups." 596 N.E.2d at 315. The substructure groups "most likely occur along racial and ethnic lines." Id. The court concluded:
The Commonwealth is correct that it need not demonstrate absolute unanimity in the scientific world in order to demonstrate general acceptance of Cellmark's and the FBI's processes of estimating DNA profiles. In Commonwealth v. Lykus, 367 Mass. 191, 198, 327 N.E.2d 671 (1975), we stated that "neither infallibility nor unanimous acceptance of the principle need be proved to justify its admission in evidence." However, the lively, and still very current, dispute described above regarding the role of population substructure constitutes something much more than a lack of unanimity. We cannot say that the processes by which Cellmark and the FBI estimated the frequency of the defendants' DNA profiles has found "general acceptance" in the field of population genetics. Accordingly, evidence of the estimated frequencies of the defendants' DNA profiles is not admissible. Because the frequency estimates are inadmissible, evidence of a match between profiles is also inadmissible.
In State v. Vandebogart, 136 N.H. 365, 616 A.2d 483 (1992), again, the challenge was to the reliability of the FBI's data bases. According to the expert testimony and the NRC report, the existence of "population substructure would compromise the FBI's method for calculating the probability of a random match" because of its impact on the "product rule," a statistical tool used in estimating probabilities of a random match, the validity of which depends on "the absence of population substructure." 616 A.2d at 493. After describing the debate over the existence and possible significance of population substructure, the court concluded:
[i]n light of the conflicting expert testimony at the Frye hearing and the NRC's recognition of considerable debate among population geneticists concerning the possibility of significant population substructure, we conclude that the FBI's method for estimating population frequencies, which relies on the product rule, has not found general acceptance in the field of population genetics.
616 A.2d at 494. Ultimately, the court noted the suggestion in the NRC report of a "method for conservatively estimating population frequencies in order to account for population substructure" known as the ceiling principle, which could "account for any error caused by possible population substructure," 616 A.2d at 494, and remanded the case to the trial court for a determination whether the ceiling principle is a generally accepted technique.
In United States v. Porter, 618 A.2d 629 (D.C.App. 1992), the appellate court, passing on the admissibility of DNA profile evidence, affirmed the trial court's determination that there was no general acceptance in the relevant scientific community "as to the soundness of certain assumptions on which this calculation was predicated." 618 A.2d at 636. The court was again dealing with the issue of the FBI's probability calculations and the possibility and effect of population subgroups. The court rejected the government's argument that the defense objection to the probability calculation goes only to weight and not admissibility. "Since the probability of a coincidental match is an essential part of the DNA evidence, and since there is no consensus as to the accuracy of the FBI's calculation, we decline to hold that the defense objections to that precise calculation go only to its weight." 618 A.2d at 640.
The court in Porter, noting that "[t]he odds against a coincidental match do not have to be thirty million to one for evidence of the match to be admissible," remanded the case to the trial court for a determination of whether "the requisite consensus now exists for a conservative statement of the probability of a coincidental match and, if so, what the probability is in the present case." 618 A.2d at 642. The court did so because after the trial court ruled on the evidence, but before the appellate court came to its decision, the NRC came out with its report suggesting a "modified ceiling principle" for calculating the match probabilities, and based on the discussions of that principle in People v. Barney and United States v. Bridget, 120 Daily Wash.L.Rptr. 1697 (Super.Ct.D.C. 1992). 618 A.2d at 642-43.
In People v. Barney, the court ruled that "one element of current DNA analysis — the determination of the statistical significance of a match between a defendant's DNA and the DNA in bodily material at the crime scene — does not satisfy the Kelly-Frye test. . . ." 10 Cal.Rptr.2d at 732. The court reviewed the scientific debate and noted that "[t]he statistical calculation step is the pivotal element of DNA analysis, for the evidence means nothing without a determination of the statistical significance of a match of DNA patterns. . . . It is the expression of statistical meaning, stated in terms of vanishingly small match probabilities, that makes the evidence so compelling." The court said:
To end the Kelly-Frye inquiry at the matching step, and leave it to jurors to assess the current scientific debate on statistical calculation as a matter of weight rather than admissibility, would stand Kelly-Frye on its head. We would be asking jurors to do what judges carefully avoid — decide the substantive merits of competing scientific opinion as to the reliability of a novel method of scientific proof.
The court went on to say there must be a more conservative method of determining statistical significance that the general scientific community could agree on, and referred to the NRC report for its proposal to account for the possibility of population substructure, and its suggested interim methods for calculating the frequencies, including the "modified ceiling approach." But, the court said, these suggestions, while indicating hope for future admissibility of DNA evidence, did not solve the problem in the case on review, because the evidence admitted at trial was based on the unaccepted calculations, "and no amount of after-the-fact fine tuning of the statistical calculation process can cure the error." 10 Cal.Rptr.2d at 745. The court also concluded the admission of the DNA evidence was harmless error in light of the overwhelming evidence of guilt.
In State v. Alt, 504 N.W.2d 38 (Minn.App. 1993), the applicability of the modified ceiling principle was specifically litigated below, and figures were provided in accordance with that method of calculation when the hearing was reopened after the NRC report was issued. Therefore, the appellate court reversed the lower court's determination that the evidence was inadmissible, concluding "the statistical frequencies of individual loci should be admitted in this case if calculated according to the NRC modified ceiling principle." 504 N.W.2d at 51.
In our review of the pertinent cases, we come finally to United States v. Jakobetz, which the trial court cited in his order denying the motion in limine. In Jakobetz, the court applied an abuse of discretion standard of review, and assessed the admissibility of the DNA evidence under a less stringent relevance/reliability analysis rather than Frye, although the court expressed its view in dicta that the lower court's consideration of the issue would also satisfy Frye. See State v. Anderson, 853 P.2d at 142, and People v. Barney, 10 Cal. Rptr.2d at 744. Since it is now clear that the Frye test is applicable in Florida in determining whether novel scientific evidence is admissible, see Flanagan, we do not believe Jakobetz provides the pertinent framework for analysis.
Conclusion
Having reviewed the expert testimony in the instant case, as well as scientific and legal writings, and judicial opinions from other jurisdictions, we conclude appellant has demonstrated that the method by which FDLE arrived at population frequencies of one in 30 million and one in 60 million, using the FBI data bases, is not generally accepted in the relevant scientific community. Therefore, those population frequencies are not admissible. However, we cannot ignore the expert testimony presented by the state in this case indicating that it is possible to calculate more conservatively the population frequencies, and thereby account for the problems resulting from the possible effects of population substructure. We note, in particular, that Dr. Tracey's statement that the problem could be corrected by using the highest frequencies appears to correspond to the modified ceiling principle to the extent it "requires that the largest frequency from the current data bases be used for the defendant's allele frequencies." Porter at 643 quoting Bridget. The defense experts did not conclusively refute Dr. Tracey's testimony that a more conservative calculation could be made, or the corresponding inference that such an approach would address the substructure problem and, therefore, be generally accepted in the relevant scientific community. We agree with the court in Porter that it is not necessary to have population frequency probabilities of 30 or 60 million to one for DNA profile evidence to be admissible in criminal cases; more conservative numbers will suffice so long as the method of arriving at those numbers is generally accepted.
Specifically, Dr. Tracey testified as follows:
Q: Would you agree, in light of that statement, that there is not general acceptance within the scientific community, that the type of probability statistics that Dr. Pollock made in this case are not generally accepted for use at trial?
A: I'm only able to address that from a scientific perspective and I think that if you use the type of analysis that Dr. Pollock used and follow the recommendations of Hartl and Lewontin, who are the experts that you're citing, that it's perfectly appropriate and you'll get essentially the same number. It won't be 1 in 30 million. It might be 1 in 28 or 1 in 32, but it will be in the same range. And you can do that with FBI data base because they have separate eastern and western Hispanic data bases.
. . . . .
Q: Are you aware they've created a separate data base for Indians that live in Ontario versus Indians that live in —
A: Yes, I am.
Q: And they've done that because they're concerned about the frequency statistics, the frequency of a given allele might differ between those ethnic populations?
A: There's no question. I'm working on a similar study myself and the alleles differ. The odds do not qualitatively differ. That's the point.
. . . . .
Q: And that specifically what Dr. Hartl and Dr. Lewontin and some of the others would contest that you don't know what the substructuring is in the Puerto Rican population.
A: That's true. And their solution to that is at the end of the article and what they say is to use the highest frequencies.
Q: They say not to use the method that was used here, though, I believe.
A: I'd be happy to sit down and calculate using their method and I'm sure Dr. Pollock would to[o] and it will not change the qualitative answer.
Dr. Lieberman did state that her interpretation of certain proposed alternatives for calculating the probabilities was "that they do not put forth these alternative approaches as feasible; that is, they don't fix up the basic problems with the data base and assumptions as they currently stand," and that the proposals were not currently being used. Dr. Wakeland stated he believed changes could be made "as have been suggested by Lewontin and Hartl which probably ultimately could lead to an accurate data base . . ."
The discussion of the modified ceiling principle in the cases appears to confirm that a more conservative calculation may be possible, which would be generally accepted in the relevant scientific community, lending support to Dr. Tracey's testimony. See Barney, Vandebogart, Porter, Alt, discussed above; see also People v. Watson, 257 Ill. App.3d 915, 196 Ill.Dec. 89, 629 N.E.2d 634 (1994). In addition, we note that in the cases discussed above in which DNA profile evidence has been found inadmissible, there appears to have been no suggestion by the experts who testified that a more conservative estimate of population frequencies could be made which would account for the problems of possible substructure.
In People v. Watson, 257 Ill. App.3d 915, 196 Ill.Dec. 89, 629 N.E.2d 634 (1994), the appellate court remanded the case to the trial court, which had granted a defense motion in limine, for a determination whether the ceiling principle had gained general acceptance. "Because the match of DNA patterns is a matter of substantial significance . . . and because this case has the potential for becoming a significant precedent in this jurisdiction, we believe the trial court should be given the opportunity to determine whether the recently-promulgated ceiling principle is appropriate under Frye for calculating the probability estimate to be applied to a match declaration in the present case."
We recognize that the record before us does not contain the calculation of population frequencies using the more conservative approach alluded to by Dr. Tracey, however, we believe it is significant that there is evidence in this record to suggest that a more conservative calculation would minimize the possible substructure problem, since the NRC report and some of the cases suggest DNA evidence should be admitted with more conservative population frequency statistics; therefore, we are unwilling to say at this point that the trial court erred in ruling the DNA profile evidence admissible. In addition, we believe the trial court is in a better position to assess the significance of this information in light of our discussion in this opinion of the appropriate test to be applied in determining admissibility of this type of evidence. Therefore, we vacate the order on the motion in limine, and remand to the trial court for further proceedings to determine whether a more conservative method of calculating population frequencies, such as that alluded to by Dr. Tracey, is generally accepted; if so, the DNA evidence would be admissible.
We also certify the following question as one of great public importance:
IS THE FDLE (FBI) METHOD OF CALCULATING POPULATION FREQUENCIES FOR PURPOSES OF DETERMINING THE POSSIBILITY THAT SOMEONE OTHER THAN DEFENDANT MATCHES THE DNA TAKEN FROM THE CRIME SCENE IN DNA PROFILING GENERALLY ACCEPTED IN THE RELEVANT SCIENTIFIC COMMUNITY FOR USE IN CRIMINAL TRIALS IN FLORIDA; IF NOT, IS A MORE CONSERVATIVE METHOD OF ESTIMATING POPULATION FREQUENCIES GENERALLY ACCEPTED IN THE RELEVANT SCIENTIFIC COMMUNITY FOR USE IN CRIMINAL TRIALS?
AFFIRMED in part, VACATED in part, and REMANDED.
ERVIN, JOANOS and KAHN, JJ., concur.