|
|
|
A Review of the (Published in Fingerprint Whorld, 21 (82), October, 1995, and in the Journal of Forensic Identification, 46 (1), January/February, 1996 - the article was submitted simultaneously to Fingerprint Whorld and the Journal of Forensic Identification by mutual agreement)
I. W. Evett1 Abstract: The fingerprint service of England and Wales works to the requirement that a fingerprint identification should be based on at least 16 points of comparison before evidence may be given in court. In 1988--89 the authors carried out a review of the need for this requirement. The review included: visits to bureaus in the U.K. and in various other countries; a study of the statistical aspects of fingerprint identification; a historical review; and a collaborative study in which fingerprint experts from many different bureaus at home and abroad examined ten sets of comparisons. This paper describes the conduct of the review and its conclusions. Introduction Fingerprint experts throughout the world know that their colleagues in Britain work to a 16 points standard for court quality identifications. This is an exceptionally high number and in 1988 Home Office commissioned a review to establish whether the 16 points standard was necessary. The review was carried out by the authors who were advised and assisted by two senior British fingerprint experts: Ian Lucas, Deputy Head of the New Scotland Yard Bureau, and Raymond Broadstock, Acting Head of the Midlands Bureau. This paper describes the review and its conclusions. The next section is a brief description of the organization of fingerprint bureaus in England and Wales. This is followed by an explanation of the origin of the 16 points standard with a sketch of the history of how current practices have emerged. The conduct of the review is then summarized, including a brief discussion of the relevance of statistical theory to criteria for identification and a description of a collaborative study which included international participation. The paper concludes with a summary of the findings and the recommendations made for the future development of the fingerprint service.
1_The Forensic Science Service, Priory House, Gooch Street North, Birmingham
B5 6QQ, U.K. The fingerprint service of England and Wales The policing of England and Wales is carried out by 41 forces for the provinces and two forces for the capital. Of the latter, one -- the City of London Police -- is small whereas the other -- the Metropolitan Police -- is by far the largest in the country. The Metropolitan Police, whose headquarters is known throughout the world as New Scotland Yard, has the largest fingerprint bureau in the country and that, too, has an international reputation. Not all of the provincial forces have their own fingerprint bureaus, because some share the services of regional bureaus. At the time of the review there were 34 provincial bureaus. The review was not concerned with the practice of using ten print fingerprint forms to establish the identity of individuals against known records. Neither was it concerned with the processes for developing and collecting crime marks from scenes. It was concerned with the process of identifying a friction ridge impression associated with a crime with an impression from a known individual. In the USA this is termed latent print identification. This paper employs the British shorthand terminology: the impression associated with the crime is called a mark, whereas the inked impression taken from the individual is called a print.
All of the fingerprint bureaus maintain collections of prints from known
offenders and searching the collections against marks from scenes of crimes
is a major and productive commitment. In 1988, over 42,000 identifications
were made in England and Wales of which nearly 16,000 resulted from “cold
searching”. At the time of the review, New Scotland Yard alone had an automatic
fingerprint recognition (AFR -- or AFIS as it is known in the USA) facility:
all of the other bureaus relied solely on manual searching. The provincial
bureaus employed around 460 staff on crime mark identification and in New
Scotland Yard there were roughly 150 staff devoted to this activity. History of the 16 points standard The first fingerprint bureau in the U.K. was established at New Scotland Yard in 1901. The standard for identification in those early days was that mark and print should exhibit at least 12 points (characteristics, or minutiae) of agreement and, of course, no points of disagreement. However, in 1924 there was a significant development when an fingerprint officer in the New Zealand Police sent to New Scotland Yard a copy of the illustration which is reproduced in Figure 1, though the present reproduction appears to be of considerably better quality than the one which was originally sent, as will be explained later. It shows prints from two different individuals who were born in the mid--nineteenth century and originated from a paper by M. A. Bertillon in a journal which has long ceased publication [1]. The prints had been masked in a manner which apparently was adopted to obscure dissimilarities but, within the revealed regions, 16 points of similarity were marked up. When the illustration was examined at New Scotland Yard, it was concluded that 6 of the points were not close enough to be considered similarities but that the remaining 10 were. Given an instance where two fingers from different people apparently revealed such a number of points of similarity it was decided that the 12 points standard was not high enough and it was raised to 16. A few other police forces in England and Wales had bureaus at that time and they were informed of the decision. Although 16 points had been adopted as a standard there is evidence that, during the ensuing 30 years, fingerprint experts were prepared to give identifications on less than 16 points. In 1953, an expert from a provincial bureau went to court to testify in a murder trial on marks from two different fingers associated with the crime: both of these were identified with the defendant, one having 12 points of similarity, the other 15. In that trial the defendant was convicted but the Home Secretary was alerted to the events and expressed concern that, in cases where identifications were given on less than 16 points, the defense might successfully challenge the evidence and this would tend to discredit fingerprint evidence as a whole. A meeting of representatives of the major bureaus, the Director of Public Prosecutions, and the Home Office was therefore held. Among the conclusions of the meeting was the following: “It was desirable that a common standard should be observed by all forces ... in order that there should be little risk of such evidence being challenged. “In the case of a single print this standard should be a minimum of sixteen points of resemblance.” At that meeting it was also agreed that an annual meeting should be held of representatives of all fingerprint bureaus and this became known as the National Conference of Fingerprint Experts. Between 1953 and the time of this study, the need for 16 points had been questioned in the light of the fact that no other major country had such a high numerical standard. The issue had been referred on several occasions to the National Conference for debate but the experts had always agreed that adherence to a national standard of 16 points was necessary. In 1984, however, one concession was made by the conference. It was agreed that, if a case was of particular importance, then it would be permissible for an expert to give evidence of identification on less than 16 points. The conference stipulated that the expert should be of long experience and high standing, and recognized that it might be necessary for him to say that, although he was personally satisfied with the identification, it did not meet the national standard. There was another development in practice which must be explained. The following quotation is taken from the minutes of the 1978 National Conference: “There was general agreement amongst fingerprint officers that if there were 8 points of resemblance present, that was sufficient to establish identity beyond all reasonable doubt.” Over the years, the practice has evolved of giving “non--provable” identifications. If an expert finds 8 to 15 points (10 to 15 in New Scotland Yard) then he reports the identification with the proviso that it is insufficient for court purposes. Thus the result may provide a valuable investigative lead but the expert will not give evidence to the identification in court -- except in a case of particular importance, in line with the 1984 ruling of the National Conference. Current practices and performance At the time of the review, the training period for fingerprint experts was set at five years. Each bureau had its own training regime based primarily on informal apprenticeship. The one common element in all programs was that, before qualification, a would--be expert must attend a three week advanced course, including evidence preparation and mock court exercises, at New Scotland Yard. Following that course, the expert's name was recorded in a register at New Scotland Yard. Verification procedures varied to some extent between bureaus but, in general, each identification must have been agreed upon by three people, at least two being of expert status. Shortly before the review, the performance of the provincial fingerprint bureaus had been analyzed by the Audit Commission for Local Authorities in England and Wales using statistics which are collected annually for the National Conference [2]. The Audit Commission found a wide range in the recorded figures. For example, in 1987, some bureaus recorded over 140 identifications per fingerprint officer whereas for some bureaus the figure was less than 50. In the upper quartile, bureaus recorded over 130 identifications per officer per year, whereas the lower quartile recorded less than 70. The report concluded that the overall variation in performance between bureaus was very wide and it explained that, if the performance of all provincial bureaus were at the same level of the current best 10%, then the total number of identifications per year would be doubled. The statistics analyzed by the Audit Commission made no distinction between provable and non--provable identifications. In response to inquiries made in anticipation of the current review, it was found that in around one third of the provincial bureaus over 20% of the identifications were classed as “non--provable” (i.e. 8 to 15 points). Of 35,500 identifications recorded by the provincial bureaus in 1988, approximately 6,000 were non--provable. At New Scotland Yard the proportion was far smaller -- less than 2% of the identifications were non--provable. Activities of the review team The authors assigned a high priority to meeting fingerprint experts. They visited 10 bureaus in the U.K. and talked to several experts at each. They also traveled abroad to bureaus in Europe and North America. In this section some of the impressions from the visits are described. There is also a brief summary of the relevance of statistical theory to the matter under review. Finally, some of the results of a collaborative study which was carried out as part of this review are described. Meetings with British Fingerprint Experts Most of the British fingerprint experts that the review team met strongly supported the 16 points standard, arguing that it protects the system from challenge and ensures the high regard in which fingerprint evidence is held by the courts. They expressed pride in the fact that in over 80 years there had been no known incorrect identifications given in court in the U.K. The overwhelming view was that the 16 points standard means that fingerprint evidence in the U.K. is of better quality than it is in other countries. Experts appeared to have a particularly poor regard for the fingerprint profession in the USA where there is no national numerical standard. Cases of wrongful identification which had been made by small bureaus in the USA were cited as being symptomatic of a poor system and the dominant view was that such unfortunate events would not have occurred had there been a 16 points standard in operation. When heads of bureaus were asked about the consequences of changing to a lower numerical standard they argued that this would lead to the possibility of incorrect identifications being made. Yet each of the heads was emphatic that such a mistake would not occur in his/her own bureau: the fear was always that it would occur in another, less well run, bureau. Experts in the provinces were not aware of the origin of the 16 points standard. None of those spoken to had seen the Bertillon prints. Visits to Overseas Bureaus The teams' overseas visits are listed in Table 1. While in Canada and the USA the team also had opportunity to speak to a judge and to prosecuting and defending attorneys. Table I Overseas bureaus visited by the review team
Holland Central Recherche Informatiedienst, Germany Bundeskriminalamt, Wiesbaden France Police Prefecture, Paris
Canada Royal Canadian Mounted Police, Ottawa
USA New York Police Department,
Federal Bureau of Investigation, Illinois State Police, Carbondale, IL
United States Secret Service,
City of Lakewood Police Department.
California Department of Justice,
Contra Costa County Sheriff's Department,
Los Angeles Police Department,
Los Angeles County Sheriff's Department, Holland has a 12 points standard and was a good starting point because the head of the national bureau was at that time chairing a committee which was reviewing quality standards for fingerprint evidence. Germany has a 12 points standard, but with the proviso that if the mark exhibits a clear pattern (e.g. loop, whorl) then 8 points suffice. The visit to Paris, where the national standard of 12 points is observed, was undertaken on behalf of the authors by police colleagues. The visits in Canada and the USA were particularly important because of the 1973 resolution of the International Association for Identification (IAI) that there was not, at that time, a basis for deciding on any particular minimum number of points to guarantee an identification. The team was impressed by the enthusiasm and dedication of the fingerprint officers they met. The standards in the major organizations were particularly impressive and it was instructive to learn of the strides that the IAI has taken in its certification program for fingerprint experts. It was useful to learn about the court experiences of experts in North America. It appears that, in spite of the known cases of wrongful identification which have occurred, fingerprint evidence is held in high regard. Theoretical Considerations It is tempting to believe that the problem of deciding on a numerical standard for identification can be solved by statistical methods. Surveys and statistics can provide insight and background, it is true, and there have been several interesting and informative studies. The review of Stoney and Thornton [3] and the D. Crim. thesis of Kingston [4] are invaluable sources for this topic. The simplest statistical argument is analogous to pulling balls of different color from an urn. What is the chance that the first ball will be red? -- the second blue? -- the third green? ... and so on. There are training manuals which show calculations based on raising a probability to a given power. The New Scotland Yard manual, for example, bases calculations on a probability of 1/4 which, raised to the power of 16, becomes a probability of approximately one in 4 billion. However, it is recognized by all that such arguments are overly simplistic. Of crucial importance in an identification is the spatial relationship between the points -- called the `coincident sequence' in the U.K. It is the establishment of the coincident sequence which is one of the hallmarks of the expert: even the most sophisticated of computer systems cannot emulate the expert's skill of matching and discriminating between these two dimensional patterns. Just as important is the consideration that the marks which arise in crime cases are not of the copybook standard which a simplistic statistical approach demands. They may be incomplete, contaminated with background materials, smudged, blurred, and, because of the elastic nature of the skin surface, will exhibit distortion of one kind or another. A comparison is then a series of decisions where questions such as “is this a significant difference?” can only be answered by skilled judgment. Likewise, deciding on whether a particular configuration of gray/white patches in a small area is a ridge ending, bifurcation, or continuous ridge calls for the same skilled judgment. Indeed, were it not for the need for these skills there would be no call for highly trained experts and criteria for identification could be established by statistical methods alone. However, given the complexity of the real world of fingerprint identification, to determine whether or not 16 points makes a good standard requires some understanding of what variation, if any, exists between experts in exercising their skills. For this reason the review team agreed that a collaborative study was necessary in which a large number of experts would be asked to carry out the same sets of comparisons independently of each other. Collaborative Study The study material comprised 10 mark and print comparisons. Nine of these, labeled pairs A to I, were from past casework at New Scotland Yard. Marks A to F were typical of casework but were of varying quality. The two team experts agreed that these six comparisons should result in full identifications. Marks G, H, and I were selected as borderline cases: both experts thought that G and H were non--provable identifications. One expert thought that I,. too, was a non--provable identification; the other thought that it was a full identification. The tenth comparison arose from an experiment using AFR which was carried out at New Scotland Yard for the review team. An expert carried out an automated search of 8,000 records for the nearest match to a small portion surrounding the delta of one of his own thumb prints. The database did not include his prints. He took the nearest retrieved record and used a dimpled beer mug to select the region of his thumb which gave the closest match to it. This was included in the study as a control to confirm the ability of experts to discriminate between close matches. About 140 sets of photographs of the 10 comparisons were sent to the bureaus of England and Wales with the request that experts of 10 or more years' experience should undertake the 10 comparisons independently of each other. Results were returned in sealed unmarked envelopes to the headquarters of West Mercia Police and sent in batches to one of the authors, then at Aldermaston, Berkshire, for analysis. Only at Aldermaston were the envelopes opened, thus guaranteeing anonymity to the participants. In relation to each of the ten comparisons, each expert was asked to decide whether the result was: (a) a full identification; (b) a non--provable identification; (c) not identical; (d) insufficient detail for an opinion. If the response was (a) or (b), the participant was asked to record the number of points of similarity between mark and print. Replies were received from 130 participants: a response level of over 95%. The overall picture is adequately represented by the following summary. Figures 2a, 2b, and 2c are histograms for the number of points of similarity reported for three of the comparisons which the team's experts had considered to be full identifications. Figure 3 shows the histogram for set H which had been expected to be classified as a non--provable identification. There are 71 responses represented in this figure because the remainder of the participants recorded either “insufficient detail” or “not identical” (see figure 4d). The variation in each of the four histograms confirms the subjective nature of points of comparison. It is interesting to observe, particularly from 2a and 2b, that participants tended to avoid returning 15 points. Figure 4a to 4d are pie charts which summarize the levels of identification for the same four comparisons. The responses for set J, the fabricated example, were all either “not identical” or “insufficient detail”. A puzzling aspect of set H was that 8% of participants returned “not identical”. This is a complete contrast from the 54% who returned a non--provable identification, bearing in mind (vide infra) that experts consider any identification, provable or non--provable, as a moral certainty. , the fabricated example, were all either “not identical” or “insufficient detail”. A puzzling aspect of set H was that 8% of participants returned “not identical”. This is a complete contrast from the 54% who returned a non--provable identification, bearing in mind (vide infra) that experts consider any identification, provable or non--provable, as a moral certainty. Figure 5 summarizes the distribution of the number of full identifications. The experts on the study team considered that there were 6 or 7 full identifications, yet 16% of the participants recorded 3 or fewer. Statistical analysis did not suggest any association between the number of full identifications made by an expert and his/her length of experience. International Comparisons Most of the overseas bureaus that were visited were pleased to examine the study and gave their opinions according to their own standards. The international variation was quite considerable. Holland represented one extreme. There, the three members of the committee that was setting standards were unanimous that: one of the comparisons was a court quality identification; set J was not an identification; and the remaining eight comparisons were inconclusive. At the other extreme, four experts at the FBI were unanimous in deciding that there were 9 court quality identifications, the tenth comparison being not identical. Most of the North American experts decided on 8 or 9 full identifications. Other countries took intermediate positions. Findings It should be emphasized that opinions expressed in this section are those of the authors alone and were not necessarily shared by the fingerprint officers in the review team. Exact science There is a misconception which appears to be widespread among U.K. fingerprint experts that fingerprint identification is an “exact science”. The essence of science is inductive inference based on data from observations, experiments, and measurement. Whereas data may be exact up to the limits of precision of the measurement process, the inference is a mental process which cannot be exact. Indeed the phrase “exact science" contains an internal contradiction which renders it meaningless. The root of the misconception appears to be the precision implied by the number 16. But the precision is illusory because the determination of the individual points is subjective and the collaborative study showed that experts vary widely in their judgments of individual points. The “exactness” concept may also be said to stem from the practice of only giving categorical identifications. But this is a stricture imposed by the profession upon itself by deliberately shunning corroborative evidence. This issue is discussed later. International comparisons There is a widespread view among U.K. fingerprint experts that, because they work to a higher number of points than most of the rest of the world, their “standards” are, in some sense, higher. The collaborative study showed this to be a mistaken view. If standards are measured by the avoidance of mistaken identifications, then experts in Holland and Germany are more cautious than their U.K. colleagues -- even though they declare a lower number of points for identity. If standards are measured by the readiness to give court identifications on limited quality marks, then North American experts are more confident than their U.K. colleagues. The question of what is meant by “standards” is returned to later. The Bertillon Prints It has been explained that the original decision taken at New Scotland Yard to adopt a 16 points standard was based on a pair of photographs. These, which had apparently been copied from a book on legal medicine which was published in 1924, had been sent to New Scotland Yard for comment by a fingerprint officer in the New Zealand Police Bureau. However, their original source was a 1912 paper by Alphonse Bertillon in a journal, Les Archives de l'Accasagne [1]. The review team was given valuable assistance by Professor Margot of the University of Lausanne who had a copy of the original journal in his departmental library. When it became possible to scrutinize better copies of the prints than those which were originally sent to New Scotland Yard, it was clear that they had been touched up [5]. Professor Margot confirmed that five of the points which had been considered by New Scotland Yard in 1924 to be similarities had, in fact, been converted from differences by the addition of ink lines. However, this discovery was far from original: Locard [6] had written as long ago as 1914, in another article located for the review team by Professor Margot, in relation to this illustration (translations from French by Professor Margot): “... the similarities are extremely crude and approximate, and do not withstand even superficial examination. Further, certain of the points are patently fabricated." It is not clear why Bertillon produced the illustration, because elsewhere in the same article he says: “We consider that ten to fifteen peculiarities are sufficient to enable one's conclusions to contain a degree of likelihood bordering on certainty.” And: “ ...confirming identity does not so much depend on the number of common peculiarities but the definite absence of any dissimilarity.” The original decision to adopt a standard of 16 points at New Scotland Yard was clearly based on incomplete and misleading information. Chance Marks Between Prints from Different People During meetings with U.K. fingerprint officers the team heard, in support of the 16 points standard, anecdotes -- often second hand -- of how experts had seen more than 8 points of comparison in prints from different individuals. These anecdotes were followed up and all bureaus were invited to send the team the best examples they had. Most of the examples thus received would not cause a qualified expert a moment's difficulty. The best was a masked region around a delta which demonstrated several ridge endings in similar positions. This best example was shown to experts in many bureaus and the response was invariably: “Yes, there are similarities -- but they are not identical”. This is further confirmation of the fact that there is far more to an identification than a mere counting of points. Corroborative Evidence When the original decision was taken at New Scotland Yard to adopt the sixteen points standard it was envisaged that experts might still be prepared to give opinions of probable origin in cases where less than sixteen points were found. However, this practice has not been implemented and U.K. fingerprint experts are strongly opposed to the idea of giving opinions which fall short of certainty. Most of the overseas experts that the team encountered were of the same view. Forensic scientists are accustomed to viewing evidential strength as something that varies along a continuum stretching between complete neutrality and virtual certainty. Whereas, for example, a handwriting expert may in one case express a categorical opinion that two sets of handwriting are by the same hand, in another he/she may give a qualified opinion of corroborative value only. The dominant view among fingerprint experts is that the adoption of such practices within their discipline would tend to weaken the impact of full identifications. The authors do not share this fear and hope that fingerprint experts worldwide will debate the issue anew in coming years. Non--provable Identifications It is important to understand that a fingerprint expert regards any identification as a certainty. This makes the status of the non--provable identification difficult to justify and this can cause tension between investigators and experts. The non--expert could understandably argue that if an expert is certain about an identification then he has an obligation to testify to that opinion if necessary. The expert's response is likely to be that he is unable to do so because the mark falls below the national standard: but this response is specious because the National Conference has agreed that identification evidence can be given on marks with less than sixteen points provided the case is of particular importance. There is an extraordinary inconsistency here: if such evidence can be given in important cases then why not in less important cases? Indeed it could be argued that decisions about whether or not evidence may be given are not even the expert's responsibility. The National Conference of Fingerprint Experts had no statutory professional standing: indeed, since the review it no longer exists, having been absorbed into a National Conference for Scientific Support. Experts may argue that non--provable identifications can provide useful investigative leads. That may well be true but they should not be attributed the status of certainty; neither should they, as they were at the time of the review, be given the same status in the national statistics as provable identifications. Teasing the Points Out Probably because of the sixteen points standard, a practice has grown in the U.K. service which the team did not find in the other countries visited. A fingerprint expert will generally reach an inner conviction about the correctness of an identification long before he has found 16 points. His or her subsequent activity will center on establishing that features which are clearly visible in the print can also be seen in the poorer quality mark. The print is used as a guide for scrutinizing the mark. This is called, in some quarters, “teasing the points out”. This contrasts sharply with the practice in Holland, for example, where the expert must decide on all of the usable features that are present in the mark before before Probably because of the sixteen points standard, a practice has grown in the U.K. service which the team did not find in the other countries visited. A fingerprint expert will generally reach an inner conviction about the correctness of an identification long before he has found 16 points. His or her subsequent activity will center on establishing that features which are clearly visible in the print can also be seen in the poorer quality mark. The print is used as a guide for scrutinizing the mark. This is called, in some quarters, “teasing the points out”. This contrasts sharply with the practice in Holland, for example, where the expert must decide on all of the usable features that are present in the mark before before comparing it with the print. Whereas the U.K. procedure does not necessarily entail risk, it is questionable from a scientific viewpoint: to search for features in an evidential item of unknown origin based on features observed in one of known origin would not be considered best practice in other areas of forensic science. An identification based on a small number of clearly demonstrable points in the mark would be preferable to one based on a larger number of unclear points which have only been established after reference to the print of the suspect. Sixteen points and the attainment of quality It is necessary for any organization within the criminal justice field to maximize the quality of its output within the constraints of caseload, time, finance and personnel. In the case of fingerprint bureau there are two high priority objectives: (1) A zero rate of incorrect identifications. (2) A high rate of court quality identifications. In relation to the first, the available measure is known incorrect identifications. The British experts attribute to the 16 points standard the fact that no known incorrect identifications have issued from their bureaus. So Objective I can be claimed to have been achieved. But Britain is not alone in this. Undoubtedly there are many other countries which can claim a similar record. The major organizations visited by the review have a zero rate but in the case of those such as the FBI, RCMP, LAPD, NYPD, and Illinois State, they are organizations which either have no numerical standard or which have guidelines in the 8 to 10 points range. So, whereas a 16 points standard may be a sufficient condition to achieve Objective 1, it is not a necessary condition. The 16 points standard has little to do with the attainment of Objective 2. Whereas some of the British bureaus were achieving high rates of court quality identifications at the time of the review, many had much lower rates. The inevitable conclusion to be drawn from the Audit Commission report (ibid.) and from the collaborative study was that a good proportion of bureaus were underachieving in relation to Objective 2. On the other hand, the collaborative study suggested that experts in the major North American bureaus are higher achievers in relation to Objective 2 than most of the experts in England and Wales. All of those bureaus claim to have achieved Objective 1. As a means of achieving quality of service, the 16 points standard is poor. Indeed, insisting on any particular minimum number of points is not an effective way of ensuring quality. Sixteen points is an organizational standard to guard against weaknesses in professional standards. It is the authors' view that priority should be given to ensuring that professional standards throughout our fingerprint service are of the highest. To this end they recommended a range of measures, all of which would contribute to the achievement and maintenance of high professional standards. None of these measures were original in either concept or implementation: they draw from the practices of the wider scientific world and from practices which were already in operation elsewhere in the fingerprint world. Recommendation The review recommended that a unified scheme of quality management should be instituted for the fingerprint bureaus of England and Wales. As components of the scheme the following measures were proposed. Training Apprenticeship will always be the dominant means for achieving expertise. Nevertheless there should be a minimum degree of structuring of the program for every trainee according to a national set of targets and milestones. The progress of each trainee should be centrally monitored and, if necessary, tailored to fit individual needs. Certification Testing The culmination of the training program should be an examination which includes a series of comparison tests. Examination success should be recognized by the conferring of a certificate. The lead which the IAI has taken in the USA in this area should be applauded and the U.K. can profit from that initiative. The testing should be demanding, but failure by an individual at any stage should not be seen to be a disaster, rather as an indication of what further training is necessary for that individual. On the other hand, there will be trainees who will progress at a rate which will enable them to pass the certification testing well before the five year training period that was in place at the time of this review. There is no reason why such people should not be certified. Performance Testing of Experts It is also necessary to monitor regularly the performance of practicing experts. Setting up ten comparison tests per year, along the lines of the collaborative study, would only require a small allocation of resources. Every expert would then see his/her own opinions tested against those of his/her peers throughout the country. Further, there would be a steady flow of management information. The performance tests could include marks made under carefully controlled conditions and would provide benchmarks for comparing identification criteria among participants. AFR could be used to set up more “near misses” such as the one included in this study. The authors noted and applauded the initiative taken by Collaborative Testing Services in creating performance tests in North America. File Audits To monitor overall standards within each bureau a national system for auditing bureaus files should be initiated. The officers carrying out these audits would, of course, need to be high caliber experts: perhaps seconded to the audit team for limited periods on rotation. Blind Trials The idea of blind testing of laboratories is familiar to those in the forensic science world. A blind trial is, essentially a fabricated case, submitted to the examining agency through normal police channels as an actual case. Such an exercise, while often difficult to set up, provides an invaluable test of the agencies' procedures from reception through to statement preparation. A continuing scheme of blind trials would provide valuable information on the performance of the fingerprint bureaus. The sixteen points standard There was widespread expectation that the review's recommendations would be couched in terms of a number of points of comparison. On the contrary, the review confirmed what had been determined by the IAI in the USA after detailed deliberation in 1973: that point counts do not make a lot of sense. There is no statistical justification for 16 points. There is no logical justification for 16 points. It is not even a “standard” as its supporters maintain because the determination of individual points is poorly defined. Its origin was a pair of imperfect, poorly understood, misleadingly retouched photographs which surfaced at New Scotland Yard in 1924. Settling for any particular number of points is not a necessary measure for the attainment of quality. The measures outlined in the previous section offer much more effective means for improving the service to the criminal justice system than would any arbitrary change in the required number of points. Furthermore, it is the firm view of the authors that it would be quite wrong to attempt such a change from outside the profession. To minimize misunderstanding, it is worth stating unequivocally that the authors did not recommend a change to the 16 points standard. What the authors did, and continue to, believe is that the developments in the profession which would follow from the implementation of their recommendations would lead to the kind of atmosphere in which its members would come to debate the issues more freely than in the past. In particular, they believe that the training process should encourage a questioning attitude rather than a doctrinaire obedience to dogma. More junior, yet progressive, members should have the freedom to put their views forward in discussion arenas which, in the past, appear to have been dominated by a small number of highly influential senior individuals whose overpowering objective has been maintenance of the status quo. “It has stood the test of time” is not a valid reason for avoiding change. A fingerprint identification is a matter for personal expert opinion. Given a system where it is known that each expert is operating above a minimum level of competence, there is no need for a national numerical points standard. Some set of guidelines may be desirable but these should be no more than general recommendations within which the expert has the freedom to exercise his/her own professional skills. In these circumstances, a rigid numerical points count is not only unnecessary, it is irrelevant. Acknowledgments The authors wish to thank all of the fingerprint officers who gave their time willingly and courteously to talk to us both at home and overseas. Wherever the team visited as part of this review, members were always received in a spirit of hospitality and co--operation. The authors also wish to thank all of those experts, at home and overseas, who participated in the collaborative study. A particular word of thanks is due to the two experts on the team. They worked hard to give us our basic grounding in this fascinating subject and participated in a number of debates with the authors --— debates which were often lively, but never rancorous. The authors thank Professor Pierre Margot of the University of Lausanne who located the articles by Bertillon and Locard and translated all of the passages which were relevant to this inquiry. The authors also wish to record that valuable interchanges occurred with two men who have made significant contributions to the advancement of fingerprint identification as a science: David Grieve, of the Illinois State Police, and David Ashbaugh, of the RCMP. Finally, this review was carried out as part of a much wider review of scientific support services to the police and the authors worked in close co--operation with Chief Superintendent William Hannis and Superintendent Alan Scrafton. They both contributed significantly in terms of organization, support, and ideas as well as providing the police officer's perspective. For further information, please contact:
Ian W. Evett, D. Sc.
References 1 . Bertillon, M A., “Les Empreintes Digitales”, Archives d'Anthropologie Criminelle de Lacassagne, 27, 1912, pp 36--52. 2. Anon., “Improving the Performance of the Fingerprint Service”, The Audit Commission for Local Authorities in England and Wales: Police Papers, 2, 1988. 3. Stoney, D. A.; Thornton, J. I., “A Critical Analysis of Quantitative Fingerprint Individuality Models”, Journal of Forensic Science, 31 (4), 1986, pp 1187--1216. 4. Kingston, C. R., “Probabilistic Analysis of Partial Fingerprint Patterns”, D. Crim. Dissertation, University of California, Berkeley, 1964. 5. Champod, C.; Lennard, C.; Margot P., “Alphonse Bertillon and Dactyloscopy”, Journal of Forensic Identification, 43 (6), 1993, pp 604--625. 6. Locard, E., “La Preuve Judicaire Par Les Empreintes Digitales”, Archives D'Anthropologies Criminelle, No. 245, 1914, pp 321--348.
(Editor--— This article should be required reading for all examiners, and
thoroughly studied by those who testify. It provides a detailed background
and explanation for the U.K.'s “16 point standard”. In the article's recommendations
it confirms the I.A.I.'s 1973 deliberations on the topic. I.A.I.'s conclusion
in 1973 was the well known resolution “there is no valid basis...” In 1995,
an international assembly [in Ne'urim, Israel] of thirty examiners agreed
on a slight variation of the wording in IAI 1973's resolution and stated
that: “no scientific basis exists for requiring that a pre--determined
minimum number of friction ridge features must be present in two impressions
in order to establish a positive identification. [JFI pp 580--581 / 45(5)].
The histograms and pie charts are extremely important in illustrating “the
subjective nature of points of comparison” or point counting. The variations
in the total number of points counted is demonstrative evidence of the
various levels of detail at which examiners may be making an examination.
One examiner looks at mark “F” and reports fourteen points, while another
examiner reports fifty--six points --— illustrating the extreme differences
achieved when merely counting points.
This article was reprinted in “THE PRINT” |