The Science of Forensic Hair Comparisons and the Admissibility of Hair Comparison Evidence: Frye and Daubert Considered
Commentators in the legal community have expressed concerns that forensic microscopical hair comparisons may not be scientific and may not meet admissibility criteria set forth in the Federal Rules of Evidence and guiding legal decisions. Forensic microscopical hair comparisons are founded on the precepts of comparative biology, microscopy, zoology, histology, and anthropology. Empirical testing demonstrates that, while not absolute positive identification, hair comparisons are good evidence of association. The validity and acceptance of forensic hair comparisons are supported by guidelines for hair comparison methodology, papers in peer-reviewed journals, textbooks, and specialized training schools. Hair comparisons are a service offered by both public and private forensic science laboratories. Hair comparison testimony has been accepted in local, state and federal courts for decades. In the authors’ opinion, forensic hair comparisons meet every criteria required by the Federal Rules of Evidence, Frye, and Daubert.
A few legal commentators have revived old arguments about certain fields in forensic science in light of the Daubert decision, based on sparse legal rulings (Williamson v. Reynolds, 1995), legal publications (Rudolf, 1996; Saks, 1993; Smith and Goodman, 1996), and discussion among lawyers and forensic scientists (AAFS, Joint Session, 1996). Particularly, they are attentive to the concern that hair comparisons may not be scientific and thereby not meet the Daubert criteria. Similar challenges have recently been launched against forensic document examiners and fingerprint examiners. In the most serious challenges to date, courts have ruled that forensic document examination was not scientific and did not have to meet the Daubert standard; it was, however, admissible as expert witness opinion testimony (U.S. v. Starzecpyzel, 1995; U.S. v. Jones, 1997).
Forensic hair comparisons, however, meet the criteria required by the Federal Rules of Evidence (FRE 702) and both Daubert and its predecessor, Frye. Hair comparison testimony has been accepted in local, state, and Federal courts regularly for decades (U.S. v. Haskins, 1976; U.S. v. Cyphers, 1977; U.s. v. Brown, 1977; State v. Watley, 1989; State v. Faricloth, 1990; State v. Payne, 1991; State v. Bridges, 1992; Crawford v. State, 1992; People v. Wettese, 1992; Suggs v. State, 1995; McCary v. State, 1995; Beam v. State, 1995; U.S. v. Matta-Ballesteros, 1995). The confusion and misunderstanding by the legal community stems from a variety of sources, including popular notions about science in general and specifically forensic science, using forensic DNA examinations as a model for interpreting forensic hair examination results, and the difference between calculating probabilities and scientific reliability. Coupled with results that offer a range of certainty about the conclusions and an incomplete knowledge about forensic hair comparisons by the non-scientific members of the legal system and non-forensic scientists, this misunderstanding has led to isolated instances of hair comparison evidence being ruled inadmissible (Williamson v. Reynolds, 1995; McGrew v. State, 1996). Considering the relevant issues and the recent attacks, the authors think hair evidence should survive and remain a useful part of criminal investigations and trials. This is not only because forensic hair comparisons meet the requirements of FRE 702, Frye and Daubert, but primarily because they are a sound and useful forensic science.
The Science of Forensic Hair Comparisons
Forensic hair examinations are a comparative biological discipline grounded in the most basic ideas of microscopy, biology, anatomy, histology, and anthropology. The phenotypic characteristics of human hairs are assessed by observing the microscopic anatomy of a hair and estimating the body area, racial characteristics, disease effects, post-mortem effects, and trauma. The microscopic anatomy of hairs from representative samples is detailed and characterized based on its histological traits, such as the cuticle, medulla, cortex, pigment granules, cortical fusi, and others (Figure 1). The distribution and arrangement of these traits is important to the description of a hair or hairs.
Histology, the study of basic tissues and the microscopic anatomy of the body, is an excellent analogy to forensic hair comparisons (Ham, 1974). Histologists and pathologists identify tissue and cell types, cell anomalies, and other traits by their microscopic appearance.
Their identification of a cancerous cell does not turn on a probability; no statistic is calculated to determine the likelihood that a cell is either a neuron or a liver cell. It is visual identification based on known characteristics and comparison with known type samples. Likewise, forensic scientists visually identify human and animal hairs and determine body area origin, racial characteristics, and microscopic anatomy. Forensic hair examiners also compare questioned hairs to known hair samples to assess microscopic similarity. The growth and biology of hair as a part of the epithelial tissue comes under the disciplines of embryology and histology.
What, and How, Is Science?
The science of forensic hair comparisons, like all science, is a means of asking and answering questions. The methodology of science has been widely published both for scientists and non-scientists (Poundstone, 1988) and will be reiterated here only briefly. People often ask questions about the same subjects scientists themselves question, but they use different methods to answer them. Science, on the other hand, bases its existence on the possibility of alternative explanations and the ability to replace one theory with another. In that vein, science has a number of objectives as outlined by Ayala (1969):
- The systematic organization of knowledge to facilitate the discovery of relationships and patterns among phenomena and processes,
- The explanation for the occurrence of these phenomena and processes, and
- The proposal of these explanations in a format which is open to the possibility of rejection.
Forensic hair comparisons comply with each of these objectives. After a systematic collection and description of the physical evidence, in every case, the study begins as a tentative, working statement, called a hypothesis, and advances either to demonstrated disproof and rejection or to continuing support and acceptance. A hypothesis is a testable statement in science; it is either rejected or supported. Hypotheses are either descriptive or explanatory. A descriptive hypothesis is a statement of asserted fact. It may be true or false but it still does not explain anything: Saying it was 95o yesterday does not explain why it was that hot or that birds are migratory does not propose the underlying mechanisms of that event. Descriptive hypotheses (facts) are inherently non-explanatory, which is why no theory ever becomes a fact.
An explanatory hypothesis, however, does try to explain something. It must be testable, and is tested indirectly through facts. Explanatory hypotheses are frequently incomplete because the phenomenon is still under study and is not yet fully understood. This does not mean that an incomplete hypothesis is false and worthless. Most scientific hypotheses are ultimately incomplete because the fields from which they arise are still being explored. Hypotheses that are accepted today are accepted because they are supported by current knowledge and what accounts for that evidence. To criticize a hypothesis for being incomplete or not totally proven is to reveal ignorance about the nature of the scientific method. Only facts are directly verified: Theories never are.
Science, as a discipline, can be distinguished from other ways of knowing, such as religion or common sense, by two distinct characteristics: (1) the manner in which science establishes the correctness of conclusions, and (2) the integration of those conclusions into a systematic body of knowledge (Dunnell, 1982). Scientists apply a number of principles in the assessment of conclusions (e.g., symmetry, parsimony, coherence, correlation) but as Dunnell states,
“…the ultimate arbitrator is comprised by…’performance criteria’, i.e., how well a proposition works in an empirical context…The use of performance criteria permits definitive experimentation and definitive empirical testing. Scientific propositions must have definitive empirical consequences.” (1978: 199)
Generally accepted procedures and performance criteria have been established for forensic hair comparison (for example, see Proceedings of the International Hair Symposium, 1985). It is the testability of these published performance criteria that allow forensic scientists to assess the accuracy and precision of techniques applied, based on a particular theory, and used for a specific purpose, like hair comparisons.
The Tyranny of Numbers: Not All Science Is Physics
Much of the history of science is rooted in the ancient traditions of Greece, Rome, and the Middle East. The Greek philosophers, including Aristotle, were primarily rationalists, a branch of philosophy that posits the solutions to scientific problems simply by focused reasoning, involving what we would now call deduction. Historically, the success which these ancient scholars had in their explanations led to “an overrating of a purely rational approach” (Mayr, 1982) that culminated with the French philosopher and mathematician Renee Descartes. Descartes believed the most scientific conclusions and theories were those that had the certainty of a mathematical proof. This idea exists even today, particularly in the popular concept of science. It has persevered not only in the physical sciences, where mathematical proofs are often possible, but also in the biological sciences (Mayr, 1982). This tyranny of numbers obscures the potential many disciplines have of achieving an optimal level of resolution because a mathematical value is expected but may not be possible (Houck, 1999). In many cases, it may be impossible for biologists to provide proofs of pure mathematical certainty due to the complex nature of living systems.
The other way of approaching reasoning is induction, where we build up observations and generalize about the world from them (Poundstone, 1988). Common sense and much of what we know about the world comes from our essentially inductionist approach to life and learning. As Poundstone describes it,
“You see a raven. It’s black. You see other ravens, and they’re black too. Never do you see a raven that isn’t black. It is inductive reasoning to conclude that “all ravens are black.” (1988:14)
Induction is reasoning from “circumstantial evidence”, not at all uncommon in litigation, and it expands our knowledge of the world.
Which is more appropriate to the forensic hair comparison, deduction or induction? Although useful to the investigation, deduction alone is insufficient, just as facts alone are not explanatory. Induction provides the platform, the “real-world” units (“ravens”) and generalizations (“all of them [so far] are black”) that allow scientists to identify and construct deductive, testable arguments. To summarize a hair comparison by stating only that “the questioned and known hairs exhibit the same microscopic characteristics” with no interpretation is not useful to the trier of fact. This deductionist approach is only half of the picture. The results should then be put into a meaningful context, usually through induction: “And, accordingly, they could have come from the same original source.” The examiner comes to this conclusion based upon his or her experience, training, and laboratory protocols in hair comparisons.
In essence, deduction and induction constitute the scientific method applicable to hairs. The basics of biology provides the larger scientific context and the units of observation; the legal application of biology, whether DNA or microscopic hair comparisons, is merely one aspect of the greater process of science as a tool for understanding the world around us. These observations are made within a mental framework of theory that predetermines the structure into which they fit and against which they are tested. Theory drives observation, observation yields facts, and facts build and adjust theory.
The tyranny of numbers is a consequence of an over-reliance on deduction and mathematics that ultimately cheat a discipline; it could also be called “physics envy.” Equating quantification with science in an attempt to justify and validate its “science-ness” indicates that a faulty notion of science, or no notion at all, is at the heart of the tyranny (May, 1982). Additionally, some information simply does not lend itself to a mathematical approach. Historical certainty, for example, is different from, but not inferior to, mathematical certainty (Cleland, 2001). The existence of trilobites, dinosaurs, the cave paintings at Lascoux, and the Roman Empire are as certain as anything in mathematics but cannot be distilled into a formula.
What forensic scientists do is try to disprove hypotheses. If a hypothesis (e.g., the questioned hair exhibits the same microscopic characteristics as hairs composing the known sample) is tested and cannot be disproved, confidence is gained in it. The more it is tested by more methods and the more they fail to disprove it, the more confidence is gained that it is valid. In this sense, there is no such thing as an absolutely true theory or explanation. Something may come up tomorrow, like new evidence, to require an idea to be rejected; scientific truth is what we know up to this moment.
Simply observing the world is not enough. The late eighteenth century saw the first application of a method suitable for the study of diversity: The comparative method. The disparity between experiment and comparison is actually small. Mayr (1982) points out that in experimentation and comparison observed data plays a crucial role. The main difference between the observations is that in an artificial laboratory experiment, the conditions are chosen and the experimenter is able to test those specific factors that yield the outcome. Experimentation is often inapplicable to many biological problems. This does not make the comparative method inferior to experimentation; the conditions may vary but other controls can be imposed to provide useful scientific information.
Another factor adding to the legitimacy of the comparative method is that a side by side comparison process of detailed observations is different from a person making a subjective determination of the properties of an object, such as simple description. This “point-by-point” comparison, as is done with hair comparisons, is a powerful technique, one that suited biology, taxonomy, and zoology since it was developed by the French naturalist Georges Cuvier (1769-1832). Each discipline requires its own appropriate methodologies and the credibility of the hypotheses generated by a discipline should be evaluated relative to that discipline’s methods and body of knowledge (Hume, 1934). Just because the Kreb’s cycle cannot be explained by Heisenberg’s Uncertainty Principle does not mean it is not based upon a valid theory.
Why Can’t You Calculate a Probability Like DNA Does?
Clinical studies Bisbing and Wolner, 1984; Lamb and Tucker, 1994; Gaudette, 1976; Gaudette and Keeping, 1974; Strauss, 1983; Wickenheiser and Hepworth, 1990) and individual observation (experience) demonstrate that forensic hair comparisons also provide an excellent means of discriminating between individuals. For example, in 1974 Gaudette and Keeping found that in 366,630 comparisons of brown Caucasian head hairs, only 9 pairs of hairs were indistinguishable. Similarly, Wickenheiser and Hepworth (1990), in a related study, found that in 431,985 comparisons of brown Caucasian head hairs, only 6 or 7 pairs were indistinguishable. Performing comparisons with brown Caucasian pubic hairs, in 101,368 comparisons, only 16 pairs were unable to be distinguished (Gaudette, 1976). In a multiracial study, Strauss (1983) reported that 100% of assessments were correct and no incorrect inclusions or exclusions were made in 4,900 comparisons. In a review of 170 hair cases at the FBI Laboratory, only 9% of the positive microscopical comparisons were excluded by mitochondrial DNA testing (Houck and Budowle, 2002). These clinical studies, and others, indicate that forensic hair comparisons have a scientific validity and basis in demonstrable fact.
All of these numbers notwithstanding, to attempt to derive a population frequency of traits or to determine how likely it may be to encounter a given hair in a given population is fraught with complexity. Most experts testify that the likelihood of finding someone else with indistinguishable hair is remote, a rare event Bisbing and Wolner, 1984; Lamb and Tucker, 1994; Gaudette, 1976; Gaudette and Keeping, 1974; Strauss, 1983; Wickenheiser and Hepworth, 1990). They do not feel comfortable with any statements about the probability that a specific hair could have come from someone other than the person to which it was associated. The authors agree with that approach. The justification for that reluctance is based on the complexity of the probability question, difficulty choosing a population to which to assign the probability (for example, see Buckleton and Walsh, 1991), the lack of sufficient data where that question was addressed (Ogle, 1998), and court decisions excluding such statements of probability in the past.
Regardless of the legal community’s interest in a “simple” probability statement, a statistical calculation that a particular hair, in a particular case, which has been associated with an exemplar, could have originated from another unknown random individual is currently impossible. It may be possible to use error rates, frequencies of specific hair traits in a given population, or likelihood ratios to illustrate the significance of a hair comparison, but the data for these values are minimal or anecdotal. Besides, the Daubert decision requires the error rate for the procedure, and not the examiner per se. The applicability of a specific error rate (one sample, one examiner) to any one particular case is doubtful.
“In order to assess the probability of an incorrect inclusion or incorrect exclusion of a suspected source for a questioned hair, the examiner must first have frequency data for the set of varieties (the hair type) which forms the match constellation” (Ogle, 1998) combined with the likelihood given a whole host of other contextual facts related to the case at hand. Consider an example: If two people are riding in the front seats of a car involved in a head-on crash, the likelihood is very low that the hairs imbedded in the windshield, which exhibit the same microscopic characteristics as occupant A but different characteristics as occupant B, came from a third unknown person. The authors think the hair evidence from such a case would be assuming its relevance to the lawsuit. The reliability and admissibility is not measured by knowing the probability that a particular hair could have originated from another unknown person. Another example: if a questioned hair was found on a bank floor after a robbery, the likelihood that the hair originated from an unknown third party seems more important to consider.
In statistics, accuracy and confidence are inversely related. If the estimate of the number of coins in someone’s pocket is less than 500, one would have 100% confidence in that value. It is not, however, optimally accurate and not a very useful number. Uncertainty is a part of science and is recognized as such by the Daubert court in requiring error rates: The rate of error must be known but it does not have to be zero. As discussed previously, however, an over-reliance on strict statistical certainty reduces a discipline to an inflexible dogmatic approach instead of science. Additionally, comparison of a qualitative, descriptive science like hair comparisons with a quantitative method like DNA analysis is unfounded (Houck, 1999). Forensic hair comparisons deal with the characterization of continuous phenotypic variation while DNA renders results based on discrete, quantified genetic information. The two are no more alike in this respect than calculus and geometry.
Hairs and DNA
The advent of mitochondrial DNA (mtDNA) sequencing provides an additional test for assessing source association between a questioned hair and an individual. Neither the microscopic nor mtDNA analysis alone, or together, enables absolute positive identification; together, however, these methods can be complementary examinations. For example, mtDNA typing can often distinguish between hairs from different sources although they have similar morphological characteristics or insufficient characteristics); in contrast, hair morphology comparisons can often distinguish between samples from different individuals that are maternally related, where mtDNA analysis is uninformative. Moreover, the addition of mtDNA analysis provides for the assessment of the performance of microscopic hair comparisons.
A recent study has shown the utility of performing microscopic and DNA hair examinations (Houck and Budowle, 2002). The results of the 170 hair comparisons and subsequent mtDNA sequences were evaluated. For the microscopic comparisons, 58.2% of the analyses yielded conclusive results (80 associations and 19 exclusions). Comparatively, 94.7% of the evidence hairs yielded conclusive results by mtDNA sequencing (97 concordant and 64 exclusions). The greater success for mtDNA typing compared with microscopic analyses should not be a surprise—many morphological characteristics in an individual’s hair are expressed as a distribution. Additionally, mtDNA sequencing is a very sensitive technique that requires only 1-2 cm of hair and often only a single reference is required for comparison purposes.
Of the 170 microscopic examinations that were made in this study, 37 (21.7%) were inconclusive as to association or exclusion and 34 (20%) were not suitable for examination. However, mtDNA sequencing provided information for 35 inconclusive and 31 insufficient hairs, 66 in total; this information was exculpatory in slightly more than half of these “uninformative” hairs. A qualified hair examiner should still view the hair microscopically to determine its qualities and description. Mitochondrial DNA analysis is destructive and a photographic record of the hair should be made prior to submission to the DNA analyst.
Microscopic hair comparisons can exclude samples where the mtDNA sequences are the same, such maternal relatives or unrelated people with the same mtDNA sequences; by contrast, hairs that cannot be excluded due to a congruence of features will also exist. Of the 80 hairs that were microscopically associated, nine comparisons were excluded by mtDNA analysis. Many of these (4 or 44%) were defined as blond Caucasian head hairs. Blond hairs, by their coloration, have much less pigmentation than darker hairs and, because pigmentation is an important comparison characteristic, hairs with sparse or no pigmentation but from different sources could appear similar. It must be stressed that these nine mtDNA exclusions are not a false positive or error rate for the microscopic method or a false exclusion rate for mtDNA typing. These results display the limits of the hairs in this sample only and not for any hairs examined by any particular examiner in any one case. The microscopic comparison is not an absolute identification and therefore some small number of individual hairs that have a congruence of certain characteristics, even though they originate from separate individuals, may exist. In those rare instances where the same microscopic characteristics are exhibited in hairs from different individuals, the appropriate interpretation for the microscopic comparison is association.
There were no apparent differences in the exclusions obtained with both analytical methods. Of the 19 microscopic exclusions, 17 were confirmed by mtDNA sequencing. The other 2 hairs provided inconclusive or insufficient results with mtDNA sequencing: The exculpatory power of hair comparisons appears to be quite good. It therefore seems reasonable to continue to routinely examine hairs microscopically because a microscopical comparison can reliably exclude hairs, which is a principle aim of the forensic comparison. When the hairs are not excluded, the power and complementary value of mtDNA can be exploited.
The Admissibility of Hair Comparison Evidence
The admissibility of hair comparison evidence will depend to a slight extent on which rule is applicable and the interest by the trial judge in being a gate-keeper. Any problems with admitting hair evidence can partly be solved if the trial court is educated better by forensic scientists to the methodology and reliability of hair comparison science. Forensic hair comparison evidence fits the requirements of science and the rules of admissibility including Frye and Daubert.
Why Do Forensic Hair Comparisons Pass the Frye Test?
While Frye (U.S. v. Frye, 1923) is based on general acceptance in the scientific community, which occasionally has been difficult to define (Green, 1992; Thornton, 1994), it is still used by many jurisdictions in the United States. The microscopic examination of hair using the comparison microscope has been the accepted standard technique for the examination and comparison of hairs for approximately the past 60 years, and, as such, this technique has been accepted by both State and Federal courts throughout the United States, in all U.S. Territories, Canada and in many European and Asian countries.
The Frye standard usually involves a two-point question: Is the field in which the underlying theory falls generally accepted in the relevant scientific community? For hair comparisons, the answer is “yes.” Comparative biology, including medicine and physical anthropology, has a long history of microscopic identification and comparison dating back to the 18th century. Comparison is the cornerstone of the majority of biology, both past and present. Microscopic techniques, combined with studied experience, provide for a highly discriminating means to examine and compare hair (Bisbing, 1982). A long history of research in physical anthropology and forensic science detailing the differences between peoples’ hair supports the credibility of the relevant science (Bisbing, 1982; Hausman, 1925a; 1925b; Hausman, 1928; Houck, 2001; Kirk, 1994; Trotter, 1930; Trotter, 1938).
The reliability of the identification and association of human hair, assuming a competent comparison, is based on extensive experience by forensic laboratories around the world since about 1932, including the FBI, Royal Canadian Mounted Police, state, and local forensic laboratories, as well as private laboratories and a body of peer reviewed scientific literature.
Forensic hair comparison has a generally accepted theory and basis for a reliable scientific practice, based on literature in the Journal of Forensic Sciences, the Journal of the Forensic Science Society (now Science and Justice), the Canadian Society of Forensic Science Journal and other related peer-reviewed scientific journals and publications, forensic science textbooks (for example, Saferstein, 1996), and technical reference books (for example, Saferstein, 1982) used by forensic scientists. State and Federal forensic science agencies, university forensic science programs, and private research institutes regularly hold specialized training courses that teach hair comparison theory and methodology (Proceedings of the International Symposium on Forensic Hair Comparisons, 1985). An extensive body of literature exists for hair biology and comparisons (Houck, 2001).
There is also a collection of court decisions from all over the country where hair evidence has been admitted over objection and upheld on appeal. Numerous case precedents exist for admissibility of forensic hair examinations as a valid and reliable science (U.S. v. Haskins, 1976; U.S. v. Cyphers, 1977; U.S. v. Brown, 1977; State v. Watley, 1989; State v. Faricloth, 1990; State v. Payne, 1991; State v. Bridges, 1992; Crawford v. State, 1992; People v. Wettese, 1992; Suggs v. State, 1995; McCary v. State, 1995; Beam v. State, 1995; U.S. v. Matta-Ballesteros, 1995).
Are there procedures available that can produce reliable results and are they generally accepted in the relevant scientific community? Specific techniques for hair comparisons do exist based on previously cited literature, the curriculum of the courses taught, the protocols that exist in most forensic science laboratories, specifically those that are accredited or are undergoing accreditation, and the experience of two generations of forensic scientists. Hair comparison tests can be constructed to have discrete answers and are, therefore, testable. Forensic hair examiners undergo testing during training and many take proficiency tests once qualified (Peterson and Markham, 1995a, 1995b).
Several clinical studies and research projects with published and peer reviewed reports have demonstrated that given a limited number of questioned and known hair samples, correct inclusions and exclusions are the rule rather than the exception (Bisbing and Wolner, 1984; Lamb and Tucker, 1994; Gaudette, 1976; Gaudette and Keeping, 1974; Strauss, 1983; Wickenheiser and Hepworth, 1990).
Why Do Hair Comparisons Also Pass the Daubert Test?
Daubert (Daubert v. Merrell Dow, 1993), by comparison with Frye, has a four-pointed definition of admissibility: No single point must dictate the outcome of the Daubert scrutiny but any single point may dictate the outcome. For example, a reviewing court may decide that, because a forensic discipline has gained general acceptance, it will be admitted. Daubert’s four main points are as follows.
1. Can the procedure be tested and thereby validated?
As mentioned in the Frye discussion, hair comparison procedures can be tested by controlled research and proficiency testing. The fact that when two hair samples are randomly selected from the population and compared microscopically, it is very unusual that they cannot be distinguished (Hicks, 1977) has been borne out by the previously cited clinical trials. In individual cases, the associations can be tested by re-analysis by another qualified examiner (peer reviewed).
2. Is the procedure peer reviewed?
There is a wealth of peer reviewed literature regarding hair characteristics, variation between individuals, forensic hair comparison techniques, and reliability studies. Additionally, although perhaps not what the Supreme Court had in mind, many laboratories employ a peer review or confirmation process whereby the hair examiner, having determined that questioned hairs from items of evidence and known hair standards exhibit the same microscopic characteristics, then takes these hairs to another qualified examiner. This second hair examiner reviews the hairs microscopically, employing the same methodology as the first hair examiner. The second examiner is free to agree or disagree with the first examiner’s results. Only when both examiners agree is the hair association reported. The advent of mitochondrial DNA testing (Budowle, et al., 1990; Wilson, et al., 1993; Wilson, et al., 1995a, 1995b) of hair provides another test, one based on strictly genetic information. Mitochondrial DNA sequencing in itself does not provide absolute positive identification but when combined with a microscopic hair comparison can yield a very strong conclusion of association.
3. Is the rate of error for that procedure available?
The rate of error for hair comparisons is contingent upon the quality, calibration, and maintenance of the microscopes, the training and experience of the examiner and the quality of the known and questioned hair samples. Error rates have been reported in the previously cited clinical trials. Although these reports measurably increase our understanding of error rates, these rates can only be used as examples or touch points and cannot be used to predict the probability of error for the case at hand. As with any application of statistics, the error rate calculation, per se, is a known quantity based on mathematics; the actual error values depend on the sample data. If you change the data, you change the error values.
While proficiency testing can be a valuable training device for reducing errors, “it is exceedingly difficult to estimate relevant error rates from either industry-wide or laboratory-specific proficiency-test results” (Committee on DNA Forensic Science, 1996). However, one study of published proficiency test results indicates that over a 13-year period, forensic hair examiners erred in proficiency tests only 8% of the time (Peterson and Markham, 1995a, 1995b). It should be noted that production of hair proficiency tests is problematic simply because of the natural variation that human hairs exhibit: Providing mass-produced, uniform hair samples to many laboratories is just not feasible. In general, proficiency tests do not mimic actual cases. Every case is different with different samples, different questions, and different solutions to the problems at hand, making objective testing of the error rates for individual cases impossible. This difficulty, however, does not preclude the validation of forensic hair comparisons but the traditional large-scale testing of it as employed with other types of evidence.
The error rate at issue is the error rate of the procedure, not of the person performing the procedure. If the forensic procedure were one that totally consumed the evidence at issue, knowing the error rate of the procedure might be necessary to ensure a defendant receives a fair trial. If the evidence is not consumed or altered, and is available for review by a defense expert, the prerequisite of having an established error rate is less necessary.
Daubert requires that the method be reliable, not absolute. The record needs to reflect why hair comparisons are reliable. For example, the relevant scientific community has experienced and demonstrated that it is a rare event that the hair from two individuals cannot be differentiated microscopically. Additionally, the published clinical trials indicate that false positives are very rare; false exclusions are by comparison more common and this works to the defendant’s favor. Daubert, we think, does not state that the error rates must be zero, only that they must be understood.
4. Is the procedure generally accepted in the scientific community?
Since hair comparisons meet this requirement of the Frye standard, they likewise meet Daubert test for the same reasons.
When forensic evidence is introduced in a legal proceeding, it is offered to bolster one of the parties’ arguments. It should be no surprise then, that legal commentators, often with an agenda, will seek to exploit isolated, anecdotal legal opinions to support their argument (for example, see Starrs, 1997).
One opinion frequently cited to suggest recent, successful challenges to hair comparison testimony is State v. McGrew. However, the Indiana Supreme Court subsequently reversed the lower court’s ruling, and found that the hair comparison testimony was properly admitted (Williamson v. Ward, 1997). The other similar opinion is Williamson v. Oklahoma. Appellate review of that portion of the opinion rejected the criticisms of hair comparison testimony because the review was conducted under the wrong legal standard (State v. Fukusaku, 1997). Other jurisdictions have since found that hair comparison testimony is admissible. At present, there are no recent legal opinions that agree with the views expressed in the earlier McGrew and Williamson decisions.
Forensic examinations of human hairs have been performed for years. Depending on the microscopist’s training and experience, and the condition of the evidence, the potential rate of error of the technique is very low. The techniques are not novel and there exists a body of literature dealing with human hair characteristics and the reliability of forensic hair comparisons. It is true that hair comparisons, as do all other sciences, depend on the judgement and experience of the examiner; but hair examiners are scientists and not simply technicians. That is why they can reliably make the necessary value judgements that come from their scientific education, training, and experience.
Professional standards for the practice of forensic hair comparisons have been proffered through international cooperation supported by international symposia. Like other forensic techniques, such as firearms, fingerprints and handwriting comparisons, the forensic comparison of hair has been well-established in the forensic laboratories of the world. Quality assurance safeguards are in place in most forensic laboratories which make use of specialized training, proficiency testing and peer review. The finding of the same microscopic characteristics in the questioned hair as in the known sample is objective and demonstrable to the trier of fact; the nature and breadth of the inferences adduced can be explained easily by the qualified expert. If warranted, experts are available to re-evaluate the evidence because the method is non-destructive. If the hair evidence is considered important and valued, a careful microscopic examination will provide evidence that is reliable and probative in a criminal investigation and lawsuit. The forensic scientist should welcome the trial judge’s gatekeeper role and be able to explain the acceptability of the forensic hair comparison in detail.
American Academy of Forensic Sciences, Annual Meeting, 1996 Joint Session, New York City, NY.
Ayala, F. J. Biology as an autonomous science, American Scientist (1969) 56:207-221.
Beam v. State, 265 Ga. 853, 463 S.E.2d 347 (Ga., Nov. 13, 1995)
Bisbing, R. E. and Wolner, M. F. Microscopical discrimination of twins’ head hair, Journal of Forensic Sciences (1984) 29:780-786.
Bisbing, R. Forensic Hair Comparisons. In R. Saferstein (ed), Forensic Science Handbook, Volume I. Prentice-Hall, Englewood Cliffs, NJ, 1982.
Buckleton, J.S. and Walsh, K. A. J. Who is “random man”?, Journal of the Forensic Science Society (1991) 31:463-468.
Budowle, B., Adams, D. E., Comey, C. T., Merrill, C. R., Mitochondrial DNA: a possible genetic material suitable for forensic analysis, in: Lee, HC, Gaensslen, RE, (eds) Advances in forensic sciences. Year Book Medical Publishers, Chicago, pp. 76-97, 1990.
Cleland, C. 2001.
Committee on DNA Forensic Science, An Update: The Evaluation of Forensic DNA Evidence, National Academy Press, 1996.
Crawford v. State, 840 P.2d 627 (Okla.Crim.App., Oct 01, 1992)
Daubert v Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)
Dunnell, R. C. Science, social science, and common sense, Journal of Anthropological Research (1982) 38:1-25.
Dunnell, R. C. Style and function: A fundamental dichotomy, American Antiquity (1978) 43:192-202.
Gaudette, B. D. Probabilities and human pubic hair comparisons, Journal of Forensic Sciences (1976) 21:514-517.
Gaudette, B. D. Strong negative conclusions: A rare event, Canadian Society of Forensic Science Journal (1985) 18:32-37.
Gaudette, B.D. and Keeping, E.D. An attempt at determining probabilities in human scalp hair comparison, Journal of Forensic Sciences (1974) 19:599-606.
Green, M. D. Expert witnesses and sufficiency of evidence in toxic substances litigation: The legacy of Agent Orange and bendectin litigation. Northwestern University Law Review (1992) 86:643-___.
Ham, A. W. Histology, Philadelphia, Pennsylvania, A.J. Lippincott Company, 1974.
Hausman, L. A. A comparative racial study of the structural elements of human head hair, American Naturalist (1925) 59:529-538.
Hausman, L. A. The pigment granules of human head hairs: A comparative racial study, American Journal of Physical Anthropology (1928) 22:273-283.
Hausman, L. A. The relationship of the microscopic structural characters of human head hair, American Journal of Physical Anthropology (1925) 8:173-177.
Hicks, J. W. Microscopy of Hairs, Federal Bureau of Investigation, Washington, D.C., 1977.
Houck, M. M., A bibliography of hair articles for the forensic scientist, Forensic Science Communications (2001), in press.
Houck, M. M., Statistics and trace evidence: The tyranny of numbers, Forensic Science Communications (1999) V1, N3, Available:
Houck, MM, and Budowle, B. Correlation of Microscopic and Mitochondrial DNA Analysis of Hairs, Journal of Forensic Sciences (2002) V45, N5: 1-4.
Hume, D. Treatise of human nature: An attempt to introduce the experimental method of reasoning into moral subjects, J. M. Dent & Sons Ltd, London, 1934.
Kirk, P. Crime Investigations, J. Thornton, ed. New York, New York, John Wiley and Sons, 1994.
Knoll v. State, 55 Wis. 249, 12 N.W. 369 (1882)
Lamb, P. and Tucker, L. G. A study of the probative value of Afro-Caribbean hair comparisons, Journal of the Forensic Science Society (1994) 34:177-179.
Mayr, E. The Growth of Biological Thought, Cambridge, Massachusetts, Harvard Belknap Press, 1982.
McCary v. State, 904 P.2d 110 (Okla.Crim.App., Sep 12, 1995)
McGrew v. State , Ind. App., 673 N.E. 2d 787 (1996)
Moore v. Commonwealth, 211 Va. 569, 570, 179 S.E. 2d 458 (1971)
Nicholas v. State, 99 Tex, Cr.R. 504, 270 S.W. 555 (Crim. App. 1925)
Ogle, R. R. Individualization of human hair: The role of the hair atlas, The Microscope (1998) 46:17-22.
People v. Kosters, 175 Mich App 748 , 438 N.W. 2d 651 (App.Ct. 1989)
People v. Wettese, 195 Mich.App. 235, 489 N.W.2d 514 (Mich.App., Aug 03, 1992)
Peterson, J. L. and Markham, P. N. Crime laboratory proficiency testing results, 1978-1991, I: Identification and classification of physical evidence, Journal of Forensic Sciences (1995) 40:994-1008.
Peterson, J. L. and Markham, P. N. Crime laboratory proficiency testing results, 1978-1991, II: Resolving questions of common origin, Journal of Forensic Sciences (1995) 40:1009-1029.
Poundstone, W. Labyrinths of Reason, New York, NY, Doubleday, 1988.
Proceedings of the International Symposium on Forensic Hair Comparisons, FBI Academy, Quantico, Virginia, June 25-27, 1985.
Rudolf, D. and Widenhouse. Hair comparison evidence: If it doesn’t fit, you can’t admit. Criminal Defense Newsletter, Michigan Appellate Defender Office (1996)19:1-2.
Saferstein, R. Criminalistics: An Introduction to Forensic Science, Englewood Cliffs, NJ, Prentice-Hall, Inc., 1996.
Saferstein, R. Forensic Science Handbook, Volume I, Englewood Cliffs, New Jersey, 1982.
Saferstein, R. Forensic Science Handbook, Volume II, Englewood Cliffs, New Jersey, 1988.
Saferstein, R. Forensic Science Handbook, Volume III, Englewood Cliffs, New Jersey, 1993.
Saferstein, R. Trace Evidence in Transition, FBI Symposium, San Antonio, TX, June, 1996.
Saks, M. Implications of the Daubert Test for forensic identification science, Shepard’s Expert and Scientific Evidence (1993) 427-434.
Skinner v. Commonwealth, 212 Va. 260, 183 S.E. 2d 725, 727 (1971)
Smith, C. A. S. and Goodman P. D. Forensic Hair Comparison Analysis: Nineteenth Century Science or Twentieth Century Snake Oil? Columbia Human Rights Law Review (1996)27:227.
Starrs, J. From bad to worse: Hair today – scorned tomorrow, Scientific Sleuthing Review (1997) 21:1-11.
State v. Bridges, 107 N.C.App. 668, 421 S.E.2d 806 (N.C.App., Oct 20, 1992).
State v. Faricloth, 99 N.C.App. 685, 394 S.E.2d 198 (N.C.App., Aug 07, 1990)
State v. Fukusaku, 85 Hawaii 462, 946 P.2d 32 (1997)
State v. McGrew, 682 N.E.2d 1289 (Ind. 1997)
State v. Payne, 328 N.C. 377, 402 S.E.2d 582 (N.C., Apr 03, 1991)
State v. Watley, 109 N.M. 619, 788 P.2d 375 (N.M.App., Dec 28, 1989)
Strauss, M. A. T. Forensic characterization of human hair I, The Microscope (1983) 31:15___ .
Suggs v. State, 907 S.W.2d 124 (Ark. 1995)
Thornton, J. I. Courts of law v. courts of science: A forensic scientist’s reaction to Daubert. Shepard’s Expert and Scientific Evidence Quarterly (1994)1:475-____.
Trotter, M. A review of the classifications of hair, American Journal of Physical Anthropology (1938) 24:105-126.
Trotter, M. The form, size, and color of head hair in American whites, American Journal of Physical Anthropology (1930)14:433-445.
U.S. v. Frye, 293 F.2d (D.C. Cir. 1923).
U.S. v. Matta-Ballesteros, 71 F.3d 754, 43 Fed. R. Evid. Serv. 388, 45 Fed. R. Evid. Serv. 255, 95 Cal. Daily Op. Serv. 9042, 95 Daily Journal D.A.R. 15, 853 (9th Cir.(Cal.), Dec 01, 1995)
United States v. Brown, 557 F.2d 541 (6th Cir. 1977)
United States v. Cyphers, 553 F.2d 1064, 1072-73 (7th Cir. 1977)
United States v. Haskins, 536 F.2d 775, 779 (8th Cir. 1976)
United States v. Jones , U.S.App. Lexis 3695 (6th Cir.) (1997).
United States v. Starzecpyzel, 880 F.Supp. 1027 (S.D.N.Y. 1995).
Wickenheiser, R.A. and Hepworth, D.G. Further evaluation of probabilities in human scalp hair comparisons, Journal of Forensic Sciences (1990) 35:1323-1329.
Williamson v. Reynolds, 904 F.Supp. 1529 (E.D. Okla. 1995).
Williamson v. Ward, 110 F.3d 1508 (10th Cir. 1997)
Wilson, M. R., DiZinno, J. A., Polanskey, D., Replogle, J., Budowle, B. Validation of mitochondrial DNA sequencing for forensic casework analysis. International Journal of Legal Medicine (1995) 108:68-72.
Wilson, M. R., Polanskey, D., Butler, ___., DiZinno, J. A., Replogle, J., Budowle, B. Extraction, PCR amplication and sequencing of mitochondrial DNA from human hair shafts. Biotechniques (1995)18:662-668.
Wilson, M. R., Stoneking, M., Holland, M. M., DiZinno, J. A., Budowle, B. Guidelines for the use of mitochondrial DNA sequencing in forensic science. Crime Laboratory Digest (1993) 20:68-78.