The admissibility of fingerprint evidence: An African perspective

Abstract Fingerprint analysis involves the comparison of a latent print and an exemplar using the standard ACE-V methodology. The uniqueness and persistence of fingerprints form the basis of their use as unique human identifiers. Despite its usefulness in criminal investigation, fingerprint analysis has been criticised for its likelihood to, sometimes, occasion avoidable miscarriages of justice. The causes of error in fingerprint analysis include cognitive bias, non-conforming regulatory standards, and ethical misconduct. This article analyses the types and causes of error in fingerprint analysis vis-à-vis the common law requirement in Nigeria and other Commonwealth countries to prove criminal charges beyond reasonable doubt. Finally, it discusses the peculiar challenges of using forensic fingerprint analysis in the criminal justice system in developing countries and explores possible ways of solving these problems.


Introduction
, eleven bombs went off in four passenger trains in Madrid, killing 193 people, and injuring 2400 others in the process. Investigations led to the arrest of Brandon Mayfield after the FBI had alleged that his fingerprints matched those recovered from the crime scene. Against the run of events, the Spanish National Police correctly identified Ouhnane Daoud, an Algerian immigrant as the owner of the latent prints. Consequently, Mayfield was released from custody and eventually got a $2 million settlement after a series of lawsuits went in his favour.
In another prominent case, Shirley McKie, a Scottish Detective Constable was accused of perjury during the trial of a murder case in 1997 involving Marion Ross. Although Ms. McKie insisted on not being present at the murder scene, she was linked to the crime scene by fingerprint analysis. Consequently, she was dismissed from the Police and charged to court. She was, however, discharged and acquitted by a Scottish High Court after fingerprint experts from the United States proved her innocence. This led to a series of lawsuits, and she eventually accepted a £750,000 out-of-court settlement.
As a result of the miscarriages of justice due to errors in fingerprint analysis, experts have called for a review of fingerprint analytical techniques and its admissibility in courts. In support of this, the US National Research Council noted that except for nuclear DNA analysis, most other forensic disciplines lacked sufficient scientific validation. The Law Commission in England and Wales has also proposed an amendment of the criteria for the admissibility of expert witnesses because of the prejudice "bad science" may have on the court. During the past decade, however, the number of research aimed at addressing these challenges has grown appreciably and the recommendations replete in the scientific literature, when applied to case work, will significantly improve the scientific validation of fingerprint analysis. As discussed in greater detail in this article, the persisting problem with fingerprint analysis relates to the non-universality of standards and this could become problematic particularly in cases where international collaboration is required in evidence analyses.
The Nigerian criminal jurisprudence gives pre-eminence to three ways of establishing the identity of the offender in a criminal trial namely, visual identification of the offender by the victim or a witness to the offence; circumstantial evidence which proves the identity of the offender; and the opinion of an expert who uses real or documentary evidence to prove the identity of the offender. 1 However, with the absence of a functional fingerprint database in many developing countries in Africa, it is challenging to identify offenders using latent prints recovered from crime scenes in these countries. This article will review the fingerprint examination procedure and its potentials in the Nigerian criminal justice system. It will analyse the types and causes of error in fingerprint analysis vis-à-vis the common law requirement in Nigeria and other Commonwealth countries to prove criminal charges beyond reasonable doubt. Finally, it will highlight the peculiar challenges of using forensic fingerprint analysis in the criminal justice system in developing countries and explore possible ways of solving these problems.

Fingerprint analysis -an overview
Fingerprint analysis is a systematic process that involves the comparison of a known fingerprint (also known as the exemplar) with a latent print (finger mark) retrieved from a scene [1]. This process begins with the analysis of the finger mark. In this step, the examiner assesses the overall presentation (quantity and quality) of the print and decides on its suitability for further inspection. If the examiner determines that the finger mark has sufficient details to establish an identification, they then compare the print with the exemplar provided and record any similarities or differences observed. The examiner then evaluates the comparison and declares a match if they find that both samples arose from the same source, and an exclusion if sufficient dissimilarities exist to exclude a match. They may also declare an inconclusive result if they are unable to form an opinion about the samples (insufficient clarity), the area needed for comparison is not recorded in the known, or where there was agreement or disagreement but the amount was insufficient to reach an identification or exclusion opinion. Once a decision is made, the prints are then passed on to independent experts who verify the results by repeating the procedure described above [2][3][4]. If the experts have discordant views, a discrepancy meeting is held to resolve their differences. This loop describes the ACE-V method for fingerprint analysis proposed by Huber and Headrick [5] and subsequently developed by Ashbaugh [6].
Fingerprint examiners typically study print details at three levels. The first level details involve the macroscopic examination of the print. The examiner locates the cores and deltas and classifies the friction ridge skin based on the Fauld's classification system. However, the ambiguity of the Fauld's system led to the introduction of other classification systems now used in some centres [7][8][9][10][11][12]. At the second level, the examiner studies specific ridge characteristics (also known as minutiae) such as ridge endings, bifurcations, islands, dots, lakes, and scars. Third level details involve the examination of more intricate ridge details such as the shape and thickness of individual ridges and the location of sweat pores. Since some fingermarks may be smudged and of very poor quality, third level details may not be readily visible in all caseworks [13]. However, when present, they give more credence to the conclusions reached by the examiner.
Like other forensic science disciplines, fingerprint analysis has also had its fair share of research and development. Early work carried out by Henry Faulds and colleagues in the nineteenth century focused on fingerprint classification methods. Other significant improvements in this field include the introduction of the automated fingerprint identification systems in the 1960s [14], and successive reviews of the number of point details required to match fingerprints. Recently, the sequential processing approach and the GYRO colour-coding documentation techniques have been introduced in some centres to minimise errors in fingerprint comparisons [15]. The impact of bias on fingerprint examinations have also been very well researched [16][17][18][19]; and some of the outcomes of these studies have already been gazetted in fingerprint analysis technical manuals.
Although improvements in fingerprint recovery methods have helped provide more working material for the analyst, the impact of human error in fingerprint analysis significantly limits the examiner's ability to conduct reliable fingermarks comparisons. This gap in knowledge birthed the discovery of new areas of research in fingerprint analysis such as cognitive psychology and the use of computational and mathematical models for latent print examination [29][30][31][32][33]. These new fields have permitted a better understanding of the errors that characterise fingerprint analysis, and in some cases, proffered plausible solutions to these problems.

Fingerprint evidence under the Nigerian law
Fingerprint analysis is an integral element of Nigeria's civil and criminal justice system. In the sphere of criminal justice, the Nigeria Police, the Economic and Financial Crimes Commission (EFCC), Independent Corrupt Practices and Other Offences Commission (ICPC), National Drug Law Enforcement Agency (NDLEA), and the Nigeria Customs Service among others collect, store, and employ fingerprints in discharging their law enforcement duties. 2 While the admissibility of fingerprint evidence in criminal trials in Nigeria is not in doubt, the weight to be attached to fingerprint analysis in convicting an accused person is less settled. 3 The guilt of an accused person must be proven 'beyond reasonable doubt' to secure a conviction, whereas fingerprint analysis could be vitiated by cognitive bias, ethical failures, and uneven standards, rules and practices. 4 It is not clear that the margin of error tolerable to forensic scientists is shaped by the standard of proof in criminal trials, but courts across common law and continental European jurisdictions have demonstrated significant creativity and pragmatism in employing forensic analysis in criminal trials. It is not every error that will affect the probative value of forensic reports tendered in evidence. In the case of Miller v. Minister of Pensions, 5 Denning J. (as he then was) observed that: 'The law would fail to protect the community if it admitted fanciful possibilities to deflect the cause of justice. If evidence is so strong against a man as to leave only a remote possibility in his favour which can be dismissed with the sentence "of course it is possible but not in the least probable" the case is proved beyond reasonable doubt but nothing short of that will suffice.' Sections 52 and 55 of the Nigerian Evidence Act 2011 presumes the accuracy of fingerprint evidence, and recognition that 'proof beyond reasonable doubt' is not 'proof beyond any shadow of doubt' bolsters the probative value of fingerprint evidence despite the potential errors highlighted above. Section 68 of the Evidence Act, 2011, which deals inter alia with the evidence of a science expert accommodates expert evidence on the identity of finger impression and DNA analysis of the blood of a human being. Section 149 of the Evidence Act further presumes the genuineness of forensic reports generated by foreign officials if the same would be admissible by the courts in that country. This provision is especially relevant in that forensic analyses in Nigeria, given deficits in expertise and facilities, typically involve collaboration with foreign law enforcement bodies. As aptly stated in the English case of R. v. Buckley, 6 per Rose, L.J., fingerprint evidence, like any other evidence, is admissible as a matter of law if it tends to prove the guilt of the defendant.
The prevailing judicial thinking in Nigeria is that provided the forensic report is relevant and tendered by a Forensic Officer who is familiar with the investigation, the report does not qualify as documentary hearsay evidence. 7 Sections 39, 40-50 of the Evidence Act 2011 provides that where the investigating police officer is unable to attend court proceedings in person or their attendance will cause unreasonable delay or expense, the forensic report can be tendered through any expert familiar with the investigation in the absence of any objections from the defence and if the court consents to the admission of the statement. Thus, a combined reading of Sections 39(d), 42, 52, 55, and 68 of the Evidence Act 2011 allows this evidence and cures any defect therein if any. 8 This does not mean that there are no guardrails. In the United Kingdom, for example, conditions for the admissibility of expert witnesses and training/recognition of fingerprint examiners are vigorously debated. Like in Nigeria, fingerprint experts are required to be employees of the Police or Government and must have completed the UK National Fingerprint Learning Programme (NFLP) [34]. Whereas foreign training confers extra credibility on forensic scientists in Nigeria, the UK does not recognise experts trained overseas [35]. Unlike the UK where fingerprint examiners are required to be affiliated to the Government, a policy believed to increase errors in latent print analysis due to confirmation bias [36,37], Nigeria allows an expert uninvolved in producing a fingerprint report to tender same in court provided they are familiar with the investigation. 9

Challenges of fingerprint analysis
In the context of this report, errors describe mistakes or deviations from truth or accuracy. Measurements are seldom perfect and are characterised by varying degrees of inaccuracy. As a result of this, scientific measurements are usually accompanied by uncertainty measurements which determine the allowable error limit for the given measurement [29]. The margin for error depends on the sensitivity of the data produced, analytical instruments/techniques used and the available knowledge in the field at the time of measurement. For measurements to be acceptable by the scientific community, they must give room for error and uncertainty considerations. Uncertainties are unavoidable and constitute the omnipresent noise that causes measurements to be imprecise [38].
Courts require expert evidence to be impartial, reliable, clear, and relevant to command credence and respect [39]. Where experts are for hire or public sector experts are tainted with corruption, dishonesty and incompetence, forensic evidence could easily become duplicitous. Courts in Nigeria have acknowledged the relative novelty of fingerprint evidence in criminal justice practice, even as its theoretical premise is now dogma and trite principle since the mid-1980s. 10 Advances in forensic science mean that the margin of error is at its narrowest in historical terms [40,41], but in developing countries, courts remain cautious in basing conviction solely on forensic evidence, preferring, often to take refuge in the 'totality of evidence' adduced by the prosecution. In Shonubi v People of Lagos, 11 the court noted, per Obaseki-Adejumo, JCA. thus: 'I must state that the use of forensic analysis has not really taken firm ground in Nigeria especially in proof of murder cases, except for drug and fingerprints in fraud and election malpractices. ' The reasons for this attitude include a. The dearth of expertise in forensic science within the country and the complexity it presents for judging; b. Inadequate facilities for the conduct of forensic investigation in Nigeria [42]; c. The ubiquitous practice of effecting the arrest of suspects often before an investigation has commenced and the resultant incentive to ratify such arbitrary arrest by securing a 'false' conviction [43]; d. The low level of trust in the core institutions of criminal justice in Nigeria [44]; and e. A pattern of late arrival of law enforcement officers to crime scenes mean that innocent bystanders and civil volunteers/rescuers may contaminate the scene. Consequently, judges seldom convict on forensic evidence alone in the same manner that they convict exclusively on eye-witness accounts.
Several factors account for experimental errors in fingerprint analysis including: a. Accidental, stupid, or intended mistakes. While accidental errors can be prevented by checking and double-checking analytical data and procedures, stupid mistakes are accidental errors that have been overlooked. Intended mistakes on the other hand, deliberately mislead the audience perhaps, by over-or understating the truth. Intended errors may be a result of insufficient knowledge about a particular procedure or irrational deductions made from analytical data. Irrespective of the cause, such deliberate mistakes have been described as scientific crimes, and are strongly discouraged [38,[45][46][47]. The practice of forensic science in Nigeria is replete with accounts of accidental, stupid, and intended errors ranging from incomplete fingerprint and DNA analysis to shabby crime scene investigations. The apparent explanations for accidental and stupid error are incompetence and inadequate staffing, while intended errors are sometimes aggravated by bribery and corruption [48]. b. Systematic errors are non-random errors that distort the result of a measurement.
They may arise from lack of proper calibration, inaccurate measurements (such as parallax, zero-point deviations), impurities in materials, or from other extrinsic sources. Such errors are frequently noticed (and corrected) during independent peer verification of results. c. Random errors or Uncertainties describe unpredictable imprecisions. They are ultimately caused by physical noise (i.e., natural fluctuations) which may occur during measurements. Since such errors are unavoidable and unpredictable, they are more correctly defined as uncertainties and reflect the acceptable error margin for measurements [29,49,50].
Another view of errors places them in three categories: pre-analytical, analytical and post-analytical errors [51]. Pre-analytical errors describe extra-laboratory errors that occur before the analysis had begun. In fingerprint analysis, for example, these may include misplaced, mismatched or wrongly labelled latent prints submitted for analysis. Analytical errors refer to mistakes that occur during the actual analysis of the evidence. They are usually associated with non-conformance to the ACE-V protocol for fingerprint analysis. They may also arise from personality differences or biases that may affect objectivity in fingerprint analysis. Post-analytical errors occur after print analysis and frequently involve transcription errors. They may also occur during expert testimonies where facts may be distorted, exaggerated, or quoted out of context during a presentation to the jury.

Cognitive bias
Cognitive bias is a tendency in the reasoning process which leads to a systematic deviation from standards of rationality and formulation of correct judgements and assessments. It arises when factors unconnected to a task affect the task itself. These extraneous factors may be connected to the case under review, but frequently have no relationship with the forensic evidence itself [19,52,53]. Four types of cognitive bias have been identified namely, confirmation bias, anchoring bias, halo effect, and overconfidence bias [54]. These biases are intuitive and may be developed from experience, culture, or the environment. They essentially constitute a fundamental part of human development. In some cases, cognitive bias may improve judgement (affect heuristic), especially when speed is valued over accuracy, while in many others, it increases the margin for error [55,56]. The different sources of cognitive bias and their influence on scientific observations and interpretation have been well researched [57,[58][59][60][61]. These influences (see Figure 1), which constitute a basic part of human nature also include case-specific variables such as expert competency level, institutional differences, and contextual information.
Case-specific information has the most effect on the conclusions reached by examiners. For example, latent print experts, irrespective of their level of experience tend to alter their results when contextual information about examined prints become available [57,[62][63][64]. Some authors have suggested that irrelevant contextual information be withheld from print experts in forensic casework, however, divergent views [65][66][67][68][69], exist on what constitutes 'irrelevant contextual information' .
Finger marks are usually smudged and of such poor quality that knowledge of their recovery techniques may provide useful background information that will assist examiners in making correct decisions. On the other hand, details about the criminal profile of the suspect, religion, race, gender, progress report of police investigations and other non-technical facts about the case may cloud the expert's judgement [53]. Moreover, fingerprint experts have been known to tweak their reports to agree with the outcomes of adjunct forensic analysis such as DNA tests [19,70]. Since fingerprint examiners are frequently police officers; it becomes almost impossible to limit their access to contextual information. However, this problem may be solved by recruiting examiners who are independent of law enforcement. In addition to eliminating confirmation bias, it will also ensure examiners serve the case and not their employers [34,36].
Individual differences among experts also contribute significantly to cognitive bias. Although experienced latent print experts are less likely to make mistakes when compared with rookie examiners [61], personality differences [58], and stereotypes [71] play a significant role in the accuracy of decisions made by professional latent print examiners. Orthogonal visual search predictors, cognitive and personality assessments, target-specific training, and competency evaluations have been proposed as workable solutions [72,73]. To a large extent, the adoption of technical procedures governed by the scientific method will curb the influence of cognitive bias on fingerprint analysis [74].
The Linear Sequential Unmasking (LSU) procedure has also been put forward as a means of minimising cognitive bias in fingerprint analysis [75]. This model requires the analyst to examine the unknown print in isolation from the reference material. This procedure does not only make it impossible for experts to tweak their initial findings to agree with the known print but ensures that the expert works from the evidence to the suspect rather than the other way around [18]. Furthermore, the use of the GYRO documentation system where colour codes are used to annotate prints minimises bias during fingerprint comparisons [76]. A modification of the LSU approach proposed by Dror and colleagues allows experts to amend their initial analysis but restrains them from completely removing or deleting it [77]. However, this model still leaves some questions unanswered as it does not define the number of changes that can be made to verified results after the introduction of the reference print. In addition to limiting bias, the Linear Sequencing Unmasking -Expanded approach also reduces the noise and appreciably improves the accuracy of decisions made by forensic experts [16]. Nonetheless, these approaches are still limited by their complexity and overall cost implications [78].
Another major cause of cognitive bias is the base rate effect. The cognitive architecture of the human brain (see Figure 1) tends to make judgements from experiences gathered over time rather than based on hardcore evidence. For example, immigration officials are inclined to rely on their intuition to identify travellers bearing weapons at airports rather than use X-ray monitors. Similarly, doctors in intensive care units (ICU) are likely to rely on their experience to identify emergencies rather than depend on medical monitors that frequently raise false alarms [79,80]. In the same way, fingerprint verifiers develop a base rate expectation to agree with the identification declared by the first examiner if their conclusions are consistent over time.
Overreliance on the ordering of candidates on automated fingerprint identification systems may introduce bias as examiners are likely to give more attention to items on the top of the list where they expect to find identifications than those lower down the list where matches are unlikely. One way to solve this problem is to introduce fake/dummy cases (fillers) into a batch of prints submitted for analysis. This filler model, which has also been described as the 'line-up approach' [18] will not only minimise bias due to the base rate effect but will also provide a good quality control measure in fingerprint analysis [81,82]. However, this approach proposes a trade-off between enhanced accuracy and other technical variables such as added cost, skilled workforce, and an increased turnaround time which most centres may be unwilling to accept.

Ethical concerns
Ethics describe a set of rules that govern professional conduct. Irrespective of personal beliefs and values, members of a particular profession are expected to abide by the rules that govern their trade [83]. Neglecting these guidelines amount to professional misconduct and may attract sanctions depending on the prevailing circumstances around the case. In fingerprint analysis, for instance, these rules relate to the objectivity of analysis, adherence to latent print examination protocols, professionalism, confidentiality, and truth statements. Forensic experts represent a critical part of the criminal justice system since they provide empirical evidence that guide legal proceedings. Consequently, experts have a duty of care to not mislead the jury, State, or defence while declaring the outcome of their investigations [84].
Truth statements constitute a significant factor in ethical considerations. Expert witnesses agree to "tell the truth, the whole truth, and nothing but the truth", which in reality, is never as straightforward as it seems. Problems with truth statements include the definition and objectivity of truth. Is the expert obligated to tell the whole truth even if it contradicts his employer's expectations? To what extent should truth be told? Is the expert expected to declare exculpatory information even if they are unnoticed by the defence (or prosecuting) team? In 1933, Tarski proposed the criteria for defining truth statements. These criteria require truth statements to be formally correct and materially adequate. In 1956, the model and model-theoretic reviews which considered the correlation between constants and variables in truth statements were published [85].
Even with a clear definition of truth statements, experts are still divided about what extent of truth statements are ethically acceptable [86]. Truth statements that only express facts without giving details about the deductions that produced these truths are described as subjective truth-telling. Objectivity of truth statements, on the other hand, require experts to not only tell the truth but recognise and declare the limitations of the methods and procedures used in arriving at their conclusions [87]. Unfortunately, this is rarely done as it exposes experts to 'unnecessary' scrutiny in court which they would rather avoid. Notwithstanding, expert witnesses are strongly encouraged to give their reports in detailed, yet easy-to-understand ways so that the jury is neither confused nor misled by their testimonies [61,[88][89][90]. Furthermore, experts must state the truth within the confines of their knowledge and skill levels and refrain from overstating facts. Gennari recommends that legal practitioners undertake training in the rudiments of science as it will help ease communication between experts and the jury and demystify the complexity of scientific reports [37]. While Gennari's submission may seem effective, the benefit-cost ratio is still uncertain.
In criminal trials in Nigeria where the stakes are high, e.g. murder, the presumption of regularity provided by sections 52 and 55 (1-2) of the Evidence Act 2011 in favour of forensic reports tendered by investigating officers or through an expert witness might generate ethical questions in practice if not principle. Since crime victims often bear the financial burden of investigation in a majority of criminal matters in Nigeria, the high costs of obtaining forensic reports from overseas creates inequities in the criminal justice process and could undermine the cause of justice where a crime victim cannot afford to bear the cost of such investigation [48,91]. Even where forensic analysis can be completed locally, the fact that they are tendered through an expert witness and can only be rebutted by an expert witness can present a conundrum for indigent accused persons. Expert witnesses are typically paid to give testimony in court as professionals, and an accused person unable to procure (a good) one (for example, those represented by Legal Aid) may have no viable means of rebutting even a flawed forensic report, raising questions of fairness. Of course, this point is by no means peculiar to forensic evidence as in the adversarial justice system, the party that can afford the most expensive lawyers often win high-stake contests. 12 The economics of forensic evidence even cuts deeper in that where the playing field is level and all parties can procure an expert to offer and rebut forensic evidence, the net effect may be zero-weight attached to the evidence such that the trial ultimately turns on independent evidence adduced before the court. This scenario renders futile some of the huge investments in forensic analysis where conviction or acquittal will probably rest on the totality of evidence. This, however, does not diminish the overall importance of forensic science in criminal trials for as the court held per Uwais, JSC in Nasiru v. State,13 '…because in criminal cases the possibility of always proving the offence charged by direct and positive testimony of eyewitness is rare… it is therefore permitted to infer from the facts proved, other facts necessary to complete the elements of guilt or establish innocence.' Another key ethical concern is the professionalism of forensic experts. Forensic scientists are frequently drawn from different independent professions to lend their knowledge and skills to criminal investigation and other forensic matters. Due to the wide range of expert disciplines, it is challenging to adopt a unified ethical code of conduct that will apply fairly to all forensic experts [90]. In a generic sense, however, forensic experts are required to live up to the expectations their clients have of them. These ideals include professional competence and credentials, honesty, confidentiality, and scientific knowledge [92].
Misleading expert testimonies and deviations from standard analytical protocols constitute a sizeable proportion of errors in fingerprint analysis. Smit and colleagues observed that incorrect fingerprint analysis comprised a third of the overturned verdicts given in England and Wales between 2010 and 2016 [90]. Although most of the errors in fingerprint analysis are due to cognitive bias and the poor quality of latent prints, non-conformance to the standard ACE-V protocol also accounts for some mistakes in fingerprint analysis. 'Dry labbing' describes scenarios where analysts report tests that were not conducted. It also includes deviations from standard test procedures without legitimate reasons.
Besides reporting tests not done, fingerprint experts may also be culpable of a breach of public trust when they arrogate to themselves knowledge or skills which they do not truly have [93]. In the UK, for example, an appeal court quashed a homicide conviction due to inconsistencies in reports given by latent print experts [35]. In addition to the Shirley McKie case [94], the R. v. Smith case [35] also brought fingerprint examination in the UK under severe criticism as all five examiners had different opinions about the same exhibit. Fingerprint experts may also unwittingly mislead the jury due to a lack of experience or sheer incompetence. One plausible way of mitigating against this is to ensure that the relevant regulatory bodies conduct frequent proficiency testing programs to determine that the skill levels of fingerprint examiners remain at the desired levels of competence.
Inadequate laboratory facilities and a shortage of expertise can cause a rise in 'dry labbing' especially in developing countries [95]. Until recently when two more laboratories were established, there was just one Police Forensic Laboratory in Nigeria at FCID Alagbon, Lagos State [42]. Objective accounts of this Laboratory noted its level of dilapidation and nothing close to state-of-the-art. 14 While the EFCC now has a relatively more modern forensic laboratory, the impartiality of that agency is greatly imperilled by public perception of it as a highly politicised witch-hunting agency whose outputs can only be salvaged by the presumption of regularity. The cost of disproving this presumption is prohibitive. Ultimately, the principle that it is better for 99 guilty persons to escape justice than to convict one innocent person must be weighed against the cost of unsolved crimes in society, especially violent crimes.
Several proposals have been put forward to minimise errors in fingerprint comparison due to ethical misconduct. Flanagan [96] and Charlton [97] for example, posit that the introduction of quality standards such as the ISO 17025 and participation of the academia in the validation of fingerprint processes, technologies and competencies will significantly reduce the margin for error. Other authors have recommended blind and double-blind modifications of the ACE-V protocol. These models aim to minimise contact between experts and permit independent verification procedures [75]. Fagert and Morris have suggested a framework for rationalising distortions in 1:1 fingerprint comparison [98]. Montani and colleagues have also introduced a protocol for resolving conflicting expert opinions using a predetermined threshold [99]. This proposal does not only reduce the tendency of experts to manipulate their results but also increases the probative value of fingerprint evidence.

Non-uniformity of standards
The sovereignty of States implies that they have independent security architectures, and by extension, different criminal investigation policies. Although the International Association for Identification (IAI) accommodates fingerprint experts from around the world, its proposed standards of education and training for fingerprint examiners are not universally enforceable since membership is optional. As a result of this, the workflow, training standards, certification criteria and conditions for the admissibility of experts differ globally. While this may reaffirm the political independence of one state from another, it may also become problematic when criminal investigation requires international collaboration.
In Shonubi, 15 a highly experienced Assistant Superintendent of Police with extensive training in forensic science from Nigeria, Britain and France could only manage preliminary tests on collected specimens in Nigeria due to the lack of equipment and had to base his report on the Appellant's shirt, CCTV and DNA of the deceased, fingernail clippings, on analysis sent to him from the London Forensic Homicide Service, London Police, UK (Metropolitan Police). In the 3-stage test on the Appellant's stained shirt, investigators had to answer the following questions: i. Was the stain blood? ii. Does this blood belong to a human being? iii. Does the human blood belong to the deceased? While questions 1 and 2 were answered in Nigeria, question 3 could only be answered by the Metropolitan Police which confirmed that the stain on the shirt was blood that matched that of the deceased. Tensions in non-uniform standards was brought to sharp relief in the build-up to the arrest of Ounane Daoud, the mastermind of the Madrid bombing in 2004 [100], the FBI and Spanish National Police were unable to reach a consensus on the identification of finger marks recovered from the crime scene as both groups of examiners based their judgements on different criteria.
The outcome of fingerprint analysis is based on the number of similarities (or unexplainable dissimilarities) shared by the exemplar and latent print. Initially, there were no standards on the number of similarities required to establish a match as identifications were solely at the discretion of the examiner. In 1953, the 16-point standard proposed by the Metropolitan Police replaced the 12-point standard adopted 39 years earlier. In 2001, however, the Home Office and the Association of Police Officers resolved that there was no logical reason for maintaining a numerical standard and recommended a non-numerical standard for fingerprint examination which is currently in use in the UK [101]. This is also the practice in Israel and the United States.
Other European countries, however, still use the 12-point standard while Italy and South Africa use the 16-and 7-point standard respectively [102,103]. Discrepancies in the application of the ACE-V protocol also exist. For example, while some centres have adopted the blind, and double-blind verification steps, others are unable to implement them due to their high cost and personnel requirements. Unlike other forensic disciplines, such non-conformities in fingerprint analysis make international collaboration extremely subjective and prone to errors.
Another cause of errors in fingerprint analysis is the incorrect interpretation and presentation of reports, especially in scenarios where close non-matches (CNMs) are encountered in casework [104]. The 12-point standard first proposed by Sir Francis Galton in 1892 was refuted by Edmund Locard who had suggested an 8-point standard following his discovery of third level ridge details. He went on to propose a 'probabilistic' template for declaring identifications if fewer than eight points of similarities exist in two prints. This proposal was, however, ignored by the fingerprint community for the next century.
Nonetheless, science has now evolved to the point where measurements which do not consider uncertainty and error rate analysis are scarcely accepted. This development birthed research on the use of mathematical and computational models in fingerprint analysis which essentially, makes a case for probabilistic reporting [31,[104][105][106][107][108][109]. The popularity of this view has continued to increase since the 2009 NAS document [110] alluded that fingerprint analysis may only be accepted as 'true science' when error calculations are included in reports, as is the case with DNA. Indeed, many agencies now include comments on error rate measurements in their reports.
Standards provide models which ensure that products, services, and systems are dependable and perform consistently to an expected level. They are also periodically reviewed to ensure that they keep pace with new technologies and client expectations. In essence, standards are set up to maintain the confidence users of a particular service have in the process, which in the case of forensic sciences, are law enforcement and the jury. These standards, which are provided by regulating and accreditation agencies use different templates around the world.
In the UK, for example, this is achieved through the United Kingdom Accreditation Service (UKAS), British Standards Institution and the Forensic Science Regulator in England and Wales. In the US, this function is carried out by the National Institute of Standards and Technology (NIST), in Canada, by the Royal Canadian Mounted Prince (RCMP), the European Network for Forensic Science Institutes (ENFSI) in Europe and the National Association of Testing Authorities (NATA) in Australia. In South Africa, the practice of forensic science is regulated by the South African Academy of Forensic Sciences (SAAFS) [111]. Besides having different templates, the recommendations provided by these institutions are not intended to be enforceable, and this constitutes a significant threat to the quality of the system [112][113][114].
In the United Kingdom, for example, conditions for the admissibility of expert witnesses and training/recognition of fingerprint examiners are vigorously debated. Fingerprint experts are required to be employees of the Police or Government and must have completed the UK National Fingerprint Learning Programme (NFLP) [34]. This policy also does not recognise experts trained overseas [35] The compulsory affiliation of fingerprint examiners to the Government has been criticised by several experts as it is believed that this will introduce errors in latent print analysis due to confirmation bias [36,37].

Conclusion
Dactyloscopy, a systematic way of grouping and classifying fingerprints was first discovered by Juan Vucetich, an Argentinian police official. Since Juan first used fingerprint analysis to uncover a murder case in Argentina in 1892, this technique has continued to play a vital role in criminal investigation around the world. At its best, fingerprint evidence can be exculpatory in vindicating an innocent accused person. However, the cost of a flawed fingerprint identification can be devastating especially in capital offences. Errors in fingerprint analysis due to cognitive bias, insufficient research, ethical misconduct, and inconsistent regulatory policies have led to avoidable miscarriages of justice. The reliability of fingerprint analysis in Nigeria is inexorably impacted by the higher tendency for these errors to occur.
In the light of these shortcomings, this article recommends independent latent print verification protocols, probabilistic reporting of fingerprint analyses, the introduction of cognitive classes into the curriculum for training fingerprint experts, unification of regulatory standards, increased frequency of proficiency testing programs, and independence of latent print examiners from law enforcement as plausible solutions to these problems.

Notes Funding
Mark Ezegbogu received a Commonwealth Scholarship funded by the UK government (NGCS-2018-427).