Monthly Archives: October 2026

The Hedgepeth trial, when DNA statistics put the interpretation of evidence to the test

A likelihood ratio of 170 septillion was presented to the jury in Durham, North Carolina, at the trial of Miguel Salguero-Olivares, who is accused of murdering Faith Hedgepeth in 2012 [1]. Derived from the analysis of a DNA mixture recovered from the victim’s body, this figure provides extremely strong statistical support for the proposition that the defendant is one of the contributors to the mixture. It is not, however, proof of his guilt. The testimony given on 14 and 15 October 2026 illustrates a central difficulty of forensic DNA evidence, namely understanding exactly what the statistics can establish and which questions they cannot answer.

Nine years of investigation before a genetic identification

On 7 September 2012, Faith Hedgepeth, a 19-year-old student at the University of North Carolina at Chapel Hill, was found dead in her apartment. The investigation generated a large number of biological samples, but the initial DNA comparisons failed to identify a suspect. In 2021, investigative genetic genealogy led detectives to Miguel Salguero-Olivares [2]. The method uses genetic matches with relatives to direct enquiries towards an individual who may be the source of a trace. In this case it identified distant relatives of the suspect, who were then interviewed and provided DNA samples [2]. A sample obtained without his knowledge during a traffic stop, two days before his arrest was announced on 16 September 2021, allowed a direct comparison with his profile [2].

Fourteen years after the events, his trial opened in Durham. The trial sessions held in October 2026 gave a prominent place to the DNA results. Erin Ermish, a forensic scientist at the North Carolina State Crime Laboratory, presented the analyses carried out on samples from the victim’s body and on several items recovered from the scene [1]. Her testimony drew particular attention to the use of likelihood ratios to assess the strength of the DNA comparisons.

What 170 septillion actually means

According to WRAL, a US news outlet, the scientist stated on 15 October 2026 that the results obtained from a sample taken from the victim’s body were 170 septillion times more probable if the mixture originated from Faith Hedgepeth (the victim) and Miguel Salguero-Olivares (the defendant) than if it originated from the victim and an unknown, unrelated individual [1]. As one septillion is 1 followed by 24 zeros, the value can be written 1.7 × 10²⁶ [1].

This figure is a likelihood ratio (LR). It compares the probability of observing the scientific findings under two alternative propositions, usually framed with reference to the positions of the prosecution and the defence. The higher the ratio, the more the findings support the first proposition over the second. It does not mean that the defendant’s guilt is 170 septillion times more probable than his innocence, nor does it represent the probability that the DNA came from him. This distinction is fundamental, since the ratio concerns the probability of the findings given the propositions considered, not the probability that those propositions are true [5].

The value also depends on the alternative proposition chosen. The one used here refers to an unknown, unrelated individual and therefore does not cover the possibility that a close relative of the defendant, who shares part of his DNA, is the source of the mixture. The question arises all the more naturally because the identification was made through relatives, and a ratio calculated against the proposition that a brother of the defendant was the contributor would be considerably lower.

To rule out this possibility, investigators do not rely on the statistical calculation alone. They also examine the suspect’s family circle to identify relatives who could plausibly be alternative sources of the DNA, such as a brother or father living in the area at the time of the offence. Where such individuals exist, it must then be established whether they can be eliminated, either by direct comparison of their DNA profiles or by evidence showing that they could not have been at the scene, such as a verified alibi. If no relative comes into consideration, or if all have been excluded, an unknown unrelated individual is indeed the appropriate alternative for the evaluation.

The other analyses presented at trial also illustrate the range of results. A second sample from the body produced a value of 114 septillion, calculated under the same propositions [3]. For the top of a rum bottle that the prosecution regards as the murder weapon, the scientist reported a likelihood ratio of 48.6 tredecillion, roughly 4.86 × 10⁴³ [3]. This figure, far higher than those obtained for the body samples, does not mean that the bottle provides stronger support for the defendant’s involvement, because the propositions compared were not the same. For the body samples, the alternative proposition paired the victim with an unknown individual, whereas for the bottle it assumed two unknown contributors [3]. The calculation therefore also reflects the rarity of the victim’s own profile, even though the presence of her DNA on an item found in her home was hardly surprising. The defence did ask the scientist whether this choice could affect the value obtained [4].

By contrast, the DNA recovered from the main body of the bottle, an area that had not been analysed initially, could not be associated with either the defendant or the victim [4]. The profiles obtained from beneath the fingernails of the victim’s left hand neither associated the defendant with the sample nor excluded him [3]. These results are not necessarily contradictory, since the quantity of DNA recovered and the composition of mixtures can vary from one area of an object to another. Their interpretation nevertheless requires a clear distinction between the samples concerned and the conclusions each one actually supports.

Procès de Hedgepeth

By Monica Casey, WRAL Durham reporter

Identifying a source does not explain a contact

A DNA match, even one accompanied by an extremely high likelihood ratio, does not necessarily show how the DNA came to be on a surface. The evaluation may address the biological origin of a trace, which forensic science refers to as the source level [6]. It may also seek to discriminate between different activities that could explain its presence, such as direct contact or indirect transfer. This second level, the activity level, requires additional information and, where the scientific data allow, a specific evaluation [7]. A likelihood ratio calculated for the source question therefore cannot be carried over as it stands to the activity question [6].

DNA can persist after contact and be transferred indirectly through a person or an object [7]. Its presence alone therefore does not allow the deposit to be dated precisely or the action that produced it to be established. The nature of the biological material does, however, bear on this evaluation. In the Hedgepeth case, the detection of spermatozoa in intimate swabs considerably narrows the range of plausible activities, without establishing when the deposit occurred or its connection with the killing [3]. The reports presented in court thus provide information on the possible contributors to the DNA mixtures, but they should not automatically be read as demonstrating the circumstances of the murder.

This distinction is central to the guidance of the European Network of Forensic Science Institutes (ENFSI), which recommends that scientific findings be evaluated against clearly defined propositions suited to the issues in the case [5].

The risk of transposing the conditional

One of the best-known errors in the interpretation of scientific evidence is the prosecutor’s fallacy, a term coined by Thompson and Schumann in 1987 [8]. Despite its name, it is not confined to the prosecution and can be committed by anyone involved in a trial. It consists in confusing the probability of observing a result if a proposition is true with the probability that the proposition is true given the result. An everyday example shows that the two are far from equivalent. The probability that a judge in France speaks French is close to 1, whereas the probability that a French speaker is a judge is minute. Nobody would make this confusion here, but it becomes much harder to detect when it involves numerical values derived from a DNA analysis. Reading the 170 septillion as a probability of guilt, an interpretation rejected above, would be one example.

Bayesian reasoning places the contribution of the expert evidence in its proper context. The likelihood ratio measures how strongly the findings support one proposition rather than the other, without indicating on its own the probability that either is true. It acts as a multiplier applied to the view that could be taken of these propositions before the DNA analysis, on the basis of the other evidence in the case, such as the whereabouts of the person concerned or the witness statements obtained. That assessment is a matter for the fact-finder, not the expert [9]. The same likelihood ratio can therefore lead to very different conclusions depending on whether the person was identified through a body of corroborating evidence or solely through a hit on a DNA database such as the FNAEG, the French national DNA database. When the ratio reaches values comparable to those in the Hedgepeth trial, this prior assessment carries much less weight on the source question, and the debate shifts to the choice of propositions, the risk of error and the circumstances of the deposit.

The Sally Clark case in England illustrates the consequences of this confusion. Tried for the murder of her two sons, who died when only a few weeks old, she was convicted in November 1999 after the paediatrician Roy Meadow told the jury that the probability of two sudden infant deaths in a family like hers was 1 in 73 million [10]. The figure had been obtained by squaring an estimate of 1 in 8,543, as if the two deaths were independent, although genetic or environmental factors shared within a family make that assumption highly questionable [11]. Above all, it was understood, including by part of the press, as the probability that the deaths were accidental, in other words as the probability that the mother was innocent [12]. Yet the murder of two infants by their mother is also an extremely rare event, and the only relevant question was the relative probability of the two explanations in the light of the findings [11]. The convictions were quashed in January 2003, mainly because of microbiological results that had not been disclosed to the defence [10]. As early as October 2001, the Royal Statistical Society had publicly condemned the misuse of statistics in the case [11].

The growth of probabilistic genotyping makes this vigilance all the more necessary. Software such as STRmix or EuroForMix can interpret complex mixtures and low-template traces by modelling the phenomena that affect profiles, in particular allele drop-out and drop-in [13]. These systems frequently produce likelihood ratios in the billions, and sometimes values of the same order as those presented in Durham. The value obtained nevertheless depends on the propositions compared, the number of contributors assumed, the model parameters and the population data used, so two analyses of the same profile may yield different values without either necessarily being wrong.

Nor does a likelihood ratio of 170 septillion mean that the risk of error is 1 in 170 septillion. The calculation assumes that the profile was correctly generated and attributed to the right sample. The risks of contamination or sample mix-up, even if small, are generally far greater than such values and, where they are not negligible, can limit the probative value of the DNA evidence much more than the rarity of the profile [14].

Expert evidence that must be understood and open to challenge

Erin Ermish’s cross-examination focused on the analytical methods and on the possibility that samples stored for several years had degraded [3]. The defence also questioned the sampling choices and the results obtained from different parts of the rum bottle [4]. These exchanges show how important it is to make scientific conclusions understandable without altering their scope. A striking figure can give an impression of absolute certainty, whereas its interpretation remains limited to a specific question.

ENFSI recommends in particular that expert reports set out the propositions considered and the limitations of the evaluation [5]. The aim is to enable those who receive the report to understand the strength of the findings without confusing this scientific assessment with the judicial decision. This requirement applies directly to experts, who must explain their results in accessible language without losing the necessary nuance. It also applies to legal practitioners and investigators, whose questions can help determine whether a conclusion genuinely addresses the propositions at issue in the case.

In the Hedgepeth trial, understanding the ratio of 170 septillion therefore requires knowing which samples it relates to and which propositions were used to calculate it. The results obtained from the other exhibits must be examined for what each contributes in its own right, without being overshadowed by the most striking figures.

Interpretation and the true value of the evidence

The scientific testimony heard in Durham is a reminder that a DNA analysis does not necessarily answer every question raised by an investigation. It can provide extremely strong support for a proposition about the origin of a trace without establishing the exact circumstances of its deposit. A likelihood ratio replaces neither the examination of the other evidence in the case nor the assessment of propositions about the facts, which is a matter for the court. For judges and lawyers, the reported value only makes sense in the light of the conditions under which it was calculated. Knowing which sample was analysed, which propositions were compared and whether the evaluation addresses the source of the DNA or the activity that led to its deposit makes it possible to determine what the expert evidence actually establishes. The clarity with which the expert sets out these elements, both in the report and in court, is an integral part of the quality of the work [5].

Références

[1] WRAL. 170 septillion times more likely, Expert testifies about DNA in UNC student murder trial. 15 octobre 2026.
[2] ABC11. Court documents reveal how investigators linked suspect in Faith Hedgepeth’s murder. 24 janvier 2022.
[3] 9th Street Journal. Day 14 of the Faith Hedgepeth trial, Scientist testifies on DNA evidence. 15 octobre 2026.
[4] The Daily Tar Heel. Forensic scientist connects defendant’s DNA to crime scene during Faith Hedgepeth murder trial. Octobre 2026.
[5] European Network of Forensic Science Institutes (ENFSI). ENFSI Guideline for Evaluative Reporting in Forensic Science. Version 3.0, 2015.
[6] Gill P., Hicks T., Butler J.M. et al. (2018). DNA commission of the International society for forensic genetics, assessing the value of forensic biological evidence, guidelines highlighting the importance of propositions, Part I, evaluation of DNA profiling comparisons given (sub-) source propositions. Forensic Science International, Genetics, 36, 189-202.
[7] Gill P., Hicks T., Butler J.M. et al. (2020). DNA commission of the International society for forensic genetics, assessing the value of forensic biological evidence, guidelines highlighting the importance of propositions, Part II, evaluation of biological traces considering activity level propositions. Forensic Science International, Genetics, 44, 102186.
[8] Thompson W.C., Schumann E.L. (1987). Interpretation of statistical evidence in criminal trials, the prosecutor’s fallacy and the defense attorney’s fallacy. Law and Human Behavior, 11(3), 167-187.
[9] Aitken C., Roberts P., Jackson G. (2010). Fundamentals of probability and statistical evidence in criminal proceedings, guidance for judges, lawyers, forensic scientists and expert witnesses. Royal Statistical Society, Practitioner Guide No 1.
[10] Court of Appeal (England and Wales). R v Clark [2003] EWCA Crim 1020, 11 avril 2003.
[11] Royal Statistical Society (2001). Royal Statistical Society concerned by issues raised in Sally Clark case. Communiqué de presse, 23 octobre 2001.
[12] Royal Statistical Society (2002). Lettre au Lord Chancellor relative à la présentation des preuves statistiques dans les procès pénaux, janvier 2002.
[13] Gill P., Gusmão L., Haned H. et al. (2012). DNA commission of the International Society of Forensic Genetics, recommendations on the evaluation of STR typing results that may include drop-out and/or drop-in using probabilistic methods. Forensic Science International, Genetics, 6(6), 679-688.
[14] Thompson W.C., Taroni F., Aitken C.G.G. (2003). How the probability of a false positive affects the value of DNA evidence. Journal of Forensic Sciences, 48(1), 47-54.

Two divers recovered after 30 years on the Mediterranean seabed

Lost in the same diving accident in December 1993, two men were recovered 30 years later at a depth of more than 100 metres, just 13 metres apart. Yet after an identical period of immersion in the same sea, the two bodies were not in the same state of preservation. In publishing the case in the International Journal of Legal Medicine, the forensic medicine team at Nice University Hospital reminds practitioners that in an underwater environment, the appearance of a body is not necessarily a reliable guide to how long it has been in the water [1].

The facts

The two divers disappeared together during an accident in the Mediterranean in December 1993. Their remains were not located until 30 years later, lying at depths of 103 and 113 metres, 13 metres from one another [1]. Taken into the care of the forensic medicine department at Nice University Hospital, both bodies underwent whole-body post-mortem computed tomography, external examination, autopsy, toxicological analysis and genetic identification [1]. Identification was achieved for both men, from muscle tissue in the first and from femoral bone in the second. Toxicological analyses were negative [1].

Two bodies, two taphonomic trajectories

Both bodies showed changes of the same kind, namely advanced skeletonisation affecting mainly the upper body, together with the persistence of adipocere-like tissue in the pelvic region and lower limbs [1]. On CT, these tissues showed attenuation values of between −100 and −40 Hounsfield units, consistent with adipocere [1]. The difference lay in the quality of what remained. In the first diver, the residual tissues were largely disorganised and reduced to yellowish-to-brown aggregates. In the second, the skin, hair and still recognisable muscle masses of the lower limbs had been preserved [1]. Adipocere, a waxy substance formed from body fat in moist, oxygen-poor conditions, is known to slow the breakdown of the tissues it permeates for long periods.

The diving equipment offers one line of explanation.

The first diver wore a two-piece semi-dry wetsuit, the second a full dry suit [1]. The authors relate these differences to the characteristics of the equipment and to the conditions of the immediate micro-environment surrounding each body [1]. A dry suit limits exchange with seawater and access by scavenging fauna, which may have favoured the formation of adipocere and then its preservation over 30 years.

What this means for expert practice

The authors draw a warning that applies to forensic pathologists and investigators alike. The visible state of decomposition does not, on its own, allow a reliable estimate of the postmortem submersion interval (PMSI) [1]. Two bodies immersed at the same time and in the same place may, 30 years on, display different post-mortem features, enough to seriously distort an estimate based on morphological criteria alone.

The case also illustrates the weight of individual factors in underwater taphonomy, a field still far less documented than its terrestrial counterpart. Diving equipment, clothing, body build or the immediate micro-environment can profoundly alter the evolution of a submerged body. Finally, it confirms that genetic identification remains possible after 30 years of deep immersion, including from soft tissue where it has been preserved, a valuable finding for investigations into long-standing disappearances in aquatic settings [1].

This finding bears directly on the concerns of the LETHE project (Localisation des épaves, traitement hydrographique et enquête, or wreck location, hydrographic processing and investigation), launched by the French Gendarmerie nationale in the summer of 2025 and led from Conflans-Sainte-Honorine by the detachment of the Pôle national de gendarmerie nautique [2]. The project grew out of the discovery, in October 2023, of skeletal remains inside a car that had lain for some 40 years in the Seille, a small river in Côte-d’Or. Its aim is to map every vehicle resting on the bed of the rivers, lakes and ponds of mainland France, in order to search for possible victims and to resolve cases that have remained unanswered [2]. By February 2026, three bodies had already been discovered, and the scheme is of close interest to the Gendarmerie’s cold case division, the Division des affaires non élucidées (DIANE), and to the serial and unsolved crimes unit of the Nanterre judicial court [2]. A public appeal has since been launched inviting anglers equipped with sonar, divers and archaeologists to report any wrecks they locate [3]. When such wrecks yield human remains, those remains have often spent several decades under water. The Nice case shows that their identification remains feasible, and that any estimate of how long they have been in the water will call for the greatest caution.

Sources

[1] Bernardi C., Amoretti N., Isaac A., Nogueira L., Temma J., Leccia C. et Alunni V. (2026), Two divers submerged in the Mediterranean Sea for 30 years, persistence of adipocere and implications for postmortem submersion interval estimation, International Journal of Legal Medicine, publié en ligne le 9 octobre 2026, DOI 10.1007/s00414-026-04038-y.

[2] Faure A. (2026), LETHE, un outil cartographique pour recenser les épaves immergées, Gendinfo, Gendarmerie nationale, 20 février 2026.

[3] Gendarmerie nationale (2026), Épaves immergées, le projet LETHE lance un appel à contribution citoyenne, Gendinfo, août 2026.

How forensic handwriting and document examination serves the search for truth

Forensic handwriting and document examination belongs to the field of forensic science. In the French nomenclature of court-appointed experts it falls under heading G, “Criminalistique – science criminelle médico-légale”, sub-heading G.12, “Investigation scientifique et technique” (criminalistics and forensic science, scientific and technical investigation)[1]. It treats the written word as an object of study in its own right, on the same footing as a fingerprint, a biological stain or a DNA trace. Within this framework, handwriting is analysed as a set of physical traces left by the writing movement. Every piece of writing leaves observable clues, such as the pressure applied to the paper, the speed of execution, the stroke sequence and direction (what French examiners call the ductus, the way the strokes are executed and linked to form the letters) or the writer’s unconscious graphic habits, all of them deeply rooted in the act of writing.

Taken together, these graphic features form traces that can be exploited through a rigorous methodology based on observation, comparison and interpretation. Unlike graphology, which seeks to read personality in handwriting, forensic handwriting examination does not set out to interpret a personality. It aims to establish facts from objectively observable graphic evidence, in order to identify or eliminate the writer of a document and to help determine whether that document is genuine. It is in this sense that forensic handwriting examination now plays a decisive role in the administration of evidence, by providing a scientific, rigorous and objective analysis of the written document.

Pour mieux comprendre la technicité de cette discipline, il est nécessaire de revenir brièvement sur le processus d’écriture. L’écriture manuscrite est un acte beaucoup plus complexe qu’un simple mouvement de la main. Intrinsèquement liée à l’activité du cerveau, elle résulte d’une combinaison d’impulsions motrices, physiologiques et neurologiques mobilisant de nombreux mécanismes cérébraux.

Forensic handwriting examination, a discipline grounded in the neurology of writing

To appreciate the technical nature of this discipline, it is worth looking briefly at the writing process itself. Handwriting is far more than a simple movement of the hand. Intrinsically bound up with brain activity, it results from a combination of motor, physiological and neurological impulses that draw on many cerebral mechanisms.

The brain plans the writing movement and coordinates the fine movements of the hand and fingers. It processes visual and spatial information, retrieves the stored forms of letters and words from memory, and ensures the fluency and rhythm of the writing. From the intention to write, whether conscious or not, to the finished text, many cerebral areas are called upon, among them the motor areas of the frontal lobe, Broca’s area, the parietal cortex, the cerebellum and the hippocampus. Any neurological impairment is therefore liable to disrupt the organisation and execution of the writing movement. As Professors Georges Serratrice and Michel Habib put it in their book,

“It is thus deep within the brain, as much an instrument as a programmer and a motor, and a seat of motivation and emotion, that this astonishingly human faculty we call writing has its source”[1].

Figure 01 : Main brain regions involved in handwriting

The joint involvement of several cerebral networks goes some way to explaining the highly individual character of the writing movement and the persistence of graphic habits specific to each writer. This neurofunctional organisation also helps to explain why handwriting displays both stable characteristics, particular to the individual, and the natural variation inherent in any graphic act.

Fundamental principles of individuality and variability in handwriting

Forensic handwriting examination rests precisely on these two fundamental principles. On the one hand, handwriting is individual, which is what makes it unique. On the other, every writer displays a natural range of variation. It is the interplay between these two characteristics that allows the examiner to identify or eliminate the writer of a handwritten document.

The individuality of handwriting stems from the progressive automation of the writing movement. Although school instruction tends to impose a common copybook model, each individual, by virtue of his or her neurological, psychological and motor make-up, develops over time kinetic habits that are entirely his or her own. Identical twins may therefore share certain handwriting features, but their personal histories and neurological functioning introduce traits specific to each of them. Handwriting is thus the expression of a neuro-psychomotor automatism peculiar to each person. Stable yet individual, this acquired mechanism makes it possible to carry out comparisons for identification purposes.

This stability does not, however, mean that the writing movement is reproduced identically every time. No one signs or writes in exactly the same way on each occasion. Handwriting is naturally subject to variation arising from factors internal to the writer or from external causes. Physiological or psychological state, fatigue, age and emotional context, as well as writing position, the quality of the paper and the nature of the writing instrument, are all liable to affect the stability of the writing. These perfectly normal variations influence the written line without undermining the fundamental constants of the individual’s graphic automatism.

The examiner’s task is therefore to distinguish natural variations compatible with a common origin from divergences that reveal an inconsistency, which may in particular result from simulation or forgery. In other words, the examiner must separate normal fluctuations from incompatible discrepancies. Such an assessment requires a structured analytical methodology, based on rigorous criteria of observation, comparison and evaluation.

The ACE-V method applied to forensic handwriting examination

Forensic handwriting and document examination now follows a methodological approach modelled on that used in other forensic disciplines, notably the ACE-V method. Developed in the 1980s by David Ashbaugh, a Canadian fingerprint expert of international standing, this scientific method provides a rigorous analytical framework designed to structure the comparison process and lend it greater objectivity.

It comprises four successive stages, namely Analysis, Comparison, Evaluation and Verification.

Analysis:

This first stage involves analysing the questioned documents and the known specimens separately. The examiner assesses the quality of the writing and the quantity of graphic material available, that is, the volume of writing that can usefully be examined, while also examining the physical components of the document, such as the paper, the inks and all the technical and graphic characteristics of the writing. The paper and inks are examined using non-destructive optical techniques. Documents are viewed under light sources in the visible spectrum (380 nm to 780 nm) and in the ultraviolet and infrared ranges. Optical filters and varying angles of incidence, including oblique lighting, are also used to reveal features invisible to the naked eye.

This analytical phase seeks to identify the specific features of the writing movement, their intrinsic variability and their recurrence. Its main purpose is to assess the homogeneity of the writings under examination, in other words to determine whether the writings within a single body of material are consistent with a single writer.

To this end, a series of parameters are analysed and measured in turn: arrangement and layout, line density, baseline alignment, slant, the size and proportions of the letterforms, connecting strokes, line quality and dynamics, writing habits and letter design. All of these features are studied methodically, rigorously and systematically.

This stage of observation also serves to look for possible signs of disguise or simulation, such as unusually slow and unexplained execution, hesitations, pen stops, patching or retouching, and breaks in rhythm or fluency.

Figure 02 : Examination of pen pressure (MEPRI method). The perceptible stop in the pressure pattern reflects a hesitation in execution, frequently observed in forged writing.

Comparison:

Before comparing, the examiner must first assess the suitability of the reference material on which the comparison will rest. A usable body of comparison material must meet several essential requirements, both qualitative and quantitative. The specimens must be originals, contemporaneous with the questioned writings, sufficient in number, attributed with certainty to the writer concerned, naturally executed and written on a surface similar to that of the documents under examination. The strength of the conclusion depends directly on the suitability of the comparison material submitted for examination.

The characteristics observed in the questioned writings are then methodically compared with those found in the known specimens. This comparative analysis follows a strict, standardised protocol in which the observations progress from the general to the particular. The observations focus in particular on the specificity and the fine detail of the written line. These are individual, automated movements, most of them imperceptible to the naked eye, which an unskilled forger finds difficult to reproduce. It is precisely these distinctive features that the examiner sets out to bring to light.

They include the pressure applied to the writing instrument, which is a particularly strong identifying feature, and the barely perceptible micro-movements on the surface of the paper that can be observed at the beginning and end of the strokes, or between the strokes themselves. The examiner looks simultaneously for similarities and differences, assessing their nature, their frequency and their discriminating value.

Figure 03 : Homothetic comparison of the internal structure of signatures (NEGA software). The homothetic profile of the questioned signature (red dotted line) is inconsistent with the reference signatures, indicating a graphic incompatibility that supports the hypothesis of a forged signature.

Evaluation:

The evaluation stage consists in weighing the significance of the findings. The examiner determines whether the similarities observed can objectively be attributed to the same writer. In a case of simulation, for example, similarities may be the result of deliberate copying and carry no weight. Conversely, where a writer denies authorship of a piece of writing, those same similarities may be relevant evidence of identification. In this respect, and following Edmond Locard, it bears repeating that “the clue must be interpreted”, since a trace only acquires meaning at the end of a contextual and reasoned analysis.

Differences, for their part, are interpreted in the light of the natural variation of handwriting. Attributing a piece of writing to its author rests on the absence of significant differences. The examiner therefore assesses the natural variation of the writing under study and sets it against the questioned documents, in order to determine whether the discrepancies observed are significant or not. This phase calls for a critical appraisal that takes into account the limitations of the material examined, as well as the quantity and quality of the comparison specimens available, so as to assess rigorously the scope and reliability of the findings.

Figure 04 : Examination of the writing movement, clockwise in the questioned signature, anticlockwise in the genuine signatures. Analysis of the marks produced by the rotation of the pen, in particular the striations running from the inside outwards in the curved forms, reveals a writing movement that the forger neither perceived nor reproduced.

Verification:

The ACE-V method calls for a verification stage in which the case is re-examined by another qualified examiner, who applies the same methodology to the same documents with the same analytical means. This stage is intended to strengthen the reliability of the analysis and to limit the cognitive biases that can affect interpretation.

Applying the ACE-V model to forensic handwriting examination thus helps to anchor the discipline in a structured scientific approach. The methodology rests on the methodical observation of graphic traces, their reasoned comparison and the critical evaluation of the results obtained, all in the interest of objective expert conclusions. Like any scientific discipline, however, forensic handwriting and document examination depends on specific operating conditions. Where these are not met, the features observed may prove insufficient, much as an incomplete or damaged fingerprint cannot support a reliable identification.

The limitations and constraints of forensic handwriting and document examination

Like the other forensic disciplines based on the comparative analysis of traces, forensic handwriting examination has methodological and material limitations that oblige the examiner to frame conclusions with caution and restraint.

The main difficulties lie first and foremost in the quality and quantity of the writing submitted for examination. Badly degraded writing or paper, writing with little individual character, as is the case with some anonymous letters written in block capitals, or documents containing only a small number of characters, all considerably restrict the scope for analysis and identification. The use of copies, photographs or digital reproductions also entails the loss of information essential to the analysis. Certain decisive features, such as the speed of the writing movement, the pressure applied to the paper or the dynamics of the line, may become difficult or even impossible to assess objectively.

While an examination remains possible from a digital document, it is limited in scope and cannot rule out every form of falsification, in particular a composite produced by inserting the image of a genuine signature into a document. Some alterations can nonetheless be detected by examining the source file, notably through its metadata, the document properties or the consistency of the pixels.

A good-quality, first-generation photocopy can also yield useful evidence. Examining it makes it possible to assess the consistency of the printing, the quality of the images and the presence of any graphic anomalies. The absence of any detectable anomaly, however, does not permit a conclusion that the document is genuine, since some alterations may remain undetectable without the original.

In both cases, the digital document and the photocopy, the detection of technical or graphic inconsistencies is objective evidence capable of revealing a falsification. Examples include a low-resolution signature on a document that is otherwise of high image quality, or a signature which, under the microscope, proves to have been produced by colour inkjet printing while the rest of the document comes from a laser (toner-based xerographic) printer.

Non-original documents can also distort the true dimensions and proportions of the letters, particularly when the photographs were taken from an unsuitable angle, for example a tag written on the ground and photographed at 45 degrees. They can thus misrepresent the original graphic pattern by masking the writing movement, its direction, whether clockwise or anticlockwise, right to left or the reverse, and the number of pen lifts made during execution. Finally, reproductions can cause fine strokes to disappear or, conversely, artificially create spots, breaks or irregularities absent from the original document, compromising the reliability of the examination.

Examination of a non-original document, all the more so when it has passed through several generations of reproduction, does not always make it possible to determine whether a graphic element has been manipulated, for instance by cut-and-paste or image insertion.

The absence of the original document also deprives the examiner of several indispensable technical examinations, in particular the optical examinations systematically used to reveal signs of falsification such as retouching, added ink, mechanical erasures or chemical erasures.

The quality and relevance of the comparison material are decisive in any forensic handwriting examination, since they directly determine the scope of the comparative analysis and, in turn, the conclusions that can be drawn. Inadequate or quantitatively insufficient comparison material inevitably narrows the scope of the examination. How is one to compare a questioned signature when only a single known specimen is available? In practice, it cannot be done. The same applies when an examination has to rely on signatures that were not naturally written, or were even deliberately disguised, in which case the scope of the analysis is severely limited.

In such situations, a body of comparison material that is both plentiful and of suitable quality is essential in order to gauge the natural variation of the writing movement and to identify the graphic automatisms specific to the writer. A forger generally reproduces the most visible elements of the writing or signature, chiefly its overall form and certain morphological details. The same is true of a person seeking to disguise his or her writing, whose attention is focused mainly on the morphological aspects of the script.

Yet the most discriminating characteristics often lie in less perceptible elements, linked to the automation of the movement, its rhythm, its dynamics, its proportions (the relationships between its component parts) or the spatial organisation of the writing. The latter is strongly influenced by cues external to the writer, such as a box, a line or a typed entry, against which the writer spontaneously positions his or her writing or signature.

Only the examination of a sufficient number of genuine writings makes it possible to assess the permanence and variability of these characteristics and, consequently, to weigh the evidential value of the similarities and differences observed.

To these material constraints must be added intrinsic and extrinsic factors capable of significantly altering the writing movement, sometimes to the point of making it barely usable. Age, certain neurological conditions or tremors, fatigue, medication and postural constraints can all markedly change a person’s handwriting. Here again, strictly contemporaneous reference material, collected under conditions as close as possible to those of the questioned writings, is essential in order to judge whether the differences observed fall within the natural variation of the writing movement or reveal inconsistencies pointing to forgery.

An inconclusive result caused by qualitative or quantitative shortcomings should not be read as a weakness of the discipline, but as the expression of a rigorous scientific approach that respects the objective limits of the material examined. As Paul L. Kirk observed, the credibility of the forensic sciences rests as much on their ability to recognise their limits as on their ability to identify a trace or a writer.

Conclusion

Handwriting is thus a true forensic trace, the product of a complex and individual human process, on a par with the other physical evidence used in technical investigations. Its analysis relies on a rigorous methodology, grounded in scientific principles and in a structured comparative evaluation of the graphic characteristics observed.

Forensic handwriting and document examination is a constant balancing act between method, experience and interpretive caution. While frameworks such as the ACE-V method strengthen the reliability and objectivity of the examiner’s reasoning, the limitations inherent in the documents examined and in the comparison material call for a measured approach. The value of the discipline therefore lies in the examiner’s ability to consider the graphic evidence as a whole, in order to bring to the judicial debate an analysis that is reasoned, nuanced and scientifically founded.

Bibliographie :

Alain Buquet, L’expertise des écritures manuscrites, Masson, Paris, 1991
Pr. G. Serratrice et M. Habib, L’écriture et le cerveau. Mécanismes neurophysiologiques, Masson, Paris, 1993
Harralson H., Miller L., Huber and Headrick’s Handwriting Identification: Facts and Fundamentals, 2e éd., CRC Press, Boca Raton, 2018, ISBN 978-1-4987-5130-8
Hicklin R. A., Eisenhart L., Richetelli N. et al., « Accuracy and reliability of forensic handwriting comparisons », Proceedings of the National Academy of Sciences, 2022, vol. 119, n° 32, e2119944119, doi.org/10.1073/pnas.2119944119
Crot S., Marquis R., « A comparative review of error rates in forensic handwriting examination », Journal of Forensic Sciences, 2024, vol. 69, p. 2127-2138, doi.org/10.1111/1556-4029.15589


[1] French Order (arrêté) of 5 December 2022 on the nomenclature provided for in Article 1 of Decree No. 2004-1463 of 23 December 2004, https://www.cours-appel.justice.fr/sites/default/files/2023-02/arrete.pdf

[2] Pr. SERRATRICE et HABIB, L’écriture et le cerveau. Mécanismes neurophysiologiques, p. 1, MASSON.

[3] Mise en Évidence de la Pression avec Radiation IR. AIEED-2001

Canine DNA: a British database to make dog hair talk

A dog hair on a car seat or a trace of saliva on a pair of trousers still goes unexamined all too often in forensic casework. In late September 2026, researchers at King’s College London and the Royal Veterinary College announced that they are building a DNA database of the British canine population, with the aim of indicating a dog’s likely breed, and ultimately identifying an individual animal, from traces recovered at crime scenes [1][2].

The project: mapping the genetics of British dogs

The project is led by Denise Syndercombe Court, Professor of Forensic Genetics at King’s College London, and Dr Federica Giangaspero, who developed the database as part of her doctoral research (the “DogID” project), with archival sample support from Professor Brian Catchpole at the Royal Veterinary College [1]. The team’s latest findings have been published in Forensic Science International [1].

The database compiles genetic profiles from a wide range of breeds present in the United Kingdom (Labradors, Dachshunds, Cocker Spaniels, Pugs and others) in order to establish a reference population [1]. This is what gives a match its statistical weight: without knowing the frequency of a given profile within the canine population, an expert cannot tell a court what a “match” is actually worth. “Dogs are such a big part of our lives – but often overlooked in forensics. They can leave traces of their DNA at crime scenes,” explains Denise Syndercombe Court [1].

Why canine DNA matters to investigators

Animal hairs transfer readily onto clothing, car seats and carpets. At a scene, however, they are frequently classified as “non-human” and set aside without analysis [3]. Part of the evidence vanishes with them, even though sequencing the mitochondrial DNA of a single hair, as practised at the King’s Forensics laboratory, can be enough to place a suspect at the scene of a kidnapping or an unlawful detention [3]. The team is also working to estimate a dog’s likely breed from samples too small or too degraded for commercial canine ancestry testing [3].

Applications range from suspect-to-scene association to the formal identification of a dog involved in an attack on a person or on livestock, as well as dog theft and trafficking [1][2]. A precedent already exists, in cats: in 2013, hairs recovered from the curtain wrapped around a victim’s body, compared against a feline mitochondrial DNA database built by the University of Leicester, helped secure David Hilder’s manslaughter conviction in Portsmouth [5].

Non-human DNA: a field still underused

If animal DNA remains marginal in forensic laboratories, it is above all for want of reference databases, validated kits and interpretation standards comparable to those established for human DNA. A documented, published population database removes part of that obstacle: it makes it possible to quantify the value of a match and to defend that conclusion under adversarial scrutiny.

France is not starting from scratch. Since 1992, the IRCGN, the French gendarmerie’s forensic science institute, has operated a wildlife forensics department, the Département faune et flore forensiques (DFFF), attached to its human identification division. Its experts already examine traces of animal origin, including hair, feathers, blood and soft tissue, to identify a species by molecular methods and to link places or people through an animal, notably in cases of wildlife trafficking, poaching and animal cruelty [6]. The British project goes a step further in aiming at individual identification and breed prediction, both of which require dedicated population databases.

Whether such databases will be built for local canine populations, in France and elsewhere, and under what conditions these analyses will be accepted in court, remains to be seen.

Sources

[1] King’s College London — Cracking crimes with canines: forensics experts build database of dog DNA, communiqué du 28 septembre 2026
[2] The Guardian — DNA database of UK dogs could help police collar criminals, 25 septembre 2026
[3] King’s College London — Dog bites to kidnapping: how King’s Forensics uses canine DNA to solve crime, avril 2023
[4] Forensic Magazine — Forensics Experts Build Database of Dog DNA to Help Solve Crimes, 28 septembre 2026
[5] NBC News — Cat of the Baskervilles: Pet DNA helps convict British killer, août 2013
[6] Gendarmerie nationale — Département faune et flore forensiques (DFFF), IRCGN

Alcohol after death, when post-mortem fermentation misleads toxicology

After death, some micro-organisms produce ethanol of their own, so an alcohol concentration measured in a body does not always prove that the person had been drinking [1, 2]. That figure nonetheless weighs heavily on how the circumstances of a death are understood, particularly after a road traffic collision [2]. A review published in September 2026 in the International Journal of Legal Medicine takes stock of the specimens, markers and storage conditions that help separate genuine consumption from post-mortem production [1]. Its recommendations echo those set out by Olds and Jones in 2024 in the Journal of Analytical Toxicology, which add orders of magnitude useful to anyone who has to read a toxicology report [2].

Ethanol made by microbes after death

Within hours of death, gut bacteria spread into the vascular system and then to the organs [2]. Some of them, along with several yeasts, ferment whatever glucose remains and produce ethanol, much as in a wine vat [1, 2]. This post-mortem neoformation depends chiefly on the time elapsed before autopsy, the ambient temperature and the degree of putrefaction [2]. It is considered negligible when the autopsy takes place within 24 hours on a body showing no sign of decomposition, and in bodies recovered frozen [2]. It is more pronounced in heavily traumatised bodies, where open vessels expose blood to environmental microbes, and more so still after immersion in water or burning [2]. Road traffic victims, who often have multiple injuries, are therefore among those most exposed to this artefact, in cases where much hangs on the blood alcohol result [2].

The amount of available glucose also matters, which explains the attention paid to people with diabetes. Often hyperglycaemic at the time of death, they offer more substrate for fermentation, and neoformation may set in sooner [1, 2].

Usually low concentrations, occasionally high ones

The concentrations generated after death usually fall between 0.1 and 0.5 g/L of blood (0.01 to 0.05 g/100 mL), but values of 0.6 to 2 g/L have been reported [2]. A theoretical calculation gives a sense of scale, since a blood glucose of 1 g/L fermented in full would yield about 0.5 g/L of ethanol [2]. These orders of magnitude overlap the statutory drink-driving limits, which range from 0.2 to 0.8 g/L depending on the country [2]. In France, driving with 0.5 g/L of alcohol in the blood or more is a minor offence, and with 0.8 g/L or more a criminal one [6].

A high blood alcohol concentration in a putrefied body is not necessarily an artefact, however. In an autopsy series cited by Olds and Jones it mostly reflected genuine drinking, to which neoformation may have added 0.1 to 0.6 g/L [2]. These authors suggest, without claiming the force of a standard, raising the cut-off above which a post-mortem result is reported as positive from 0.1 to 0.2 g/L [2]. Where the body is obviously decomposed, they also propose subtracting 0.5 g/L from the mean result, a correction they acknowledge to be arbitrary and which massive trauma or high blood glucose may render insufficient [2].

Comparing blood with alternative specimens

Allane Patricia Santos da Paz and colleagues at the Federal University of Pará reviewed 10 years of literature, from 2014 to 2024 [1]. Of 414 articles identified in Scopus, Web of Science and PubMed, 54 studies were retained and used to draw up recommendations on collection, storage and interpretation [1]. Like earlier reviews, this one urges that a blood concentration should never be interpreted in isolation when the circumstances make production after death plausible [1, 2, 3]. Peripheral blood, drawn from the femoral vein before evisceration, remains the specimen of choice [1, 2]. Cardiac blood is more vulnerable, because ethanol still in the stomach at the time of death diffuses towards the heart, liver and lungs [2].

Vitreous humour, the fluid inside the eye, is the most useful complement. Remote from the gut and hard for bacteria to reach, it remains usable in a body that is already putrefied [1, 2].

It has its limits, notably after prolonged immersion or in a severely dehydrated body, from which it sometimes cannot be collected at all [1, 2]. Urine provides similar information, because it contains almost no glucose in healthy individuals and the intact bladder delays the arrival of bacteria [2]. That protection is lost in people with diabetes or a urinary tract infection, when yeasts such as Candida albicans find in the urine the sugar they need to ferment [1]. For badly degraded bodies, the review also mentions cerebrospinal fluid, preferably taken from the spinal canal, and matrices that are still exploratory, such as costal cartilage or the larvae of necrophagous insects [1]. Studies do not all agree on which matrix to favour in putrefaction, and some rank urine ahead of vitreous humour [1].

Interpretation then rests on whether the results are consistent. After genuine consumption, vitreous humour, which contains more water than blood, holds slightly more ethanol, with a mean ratio of 1.10 to 1.20 [2]. A positive blood result alongside negative urine or vitreous humour points instead to post-mortem production within the vascular compartment [2]. The 1989 explosion aboard the battleship USS Iowa illustrates the point, since of the 47 servicemen killed, 23 had a positive blood alcohol concentration, one of them at 1.9 g/L although neither his urine nor his bile contained measurable ethanol [2]. It would be unwise, on the other hand, to back-calculate a blood alcohol concentration from a vitreous or urine result, because the scatter between individuals is too wide [1, 2].

Looking for evidence of drinking before death

The second approach is to look for metabolites that glucose fermentation does not produce. Ethyl glucuronide (EtG) and ethyl sulphate (EtS) are formed in trace amounts when the body metabolises ingested alcohol, and their presence in blood, urine or vitreous humour is evidence of consumption before death [1, 2, 4].

EtG is the more sensitive of the two, but it can be degraded during putrefaction and disappear from badly decomposed blood, with a risk of false negatives [1]. EtS is more robust, including to heat, and less prone to being formed after sampling in infected urine [1, 2]. The Brazilian review regards it as the more reliable marker when the two disagree [1]. In a putrefied body, EtG is better measured in vitreous humour than in blood [1]. These markers call for two caveats. Their detection window, in urine especially, far exceeds that of ethanol, so a positive result establishes recent drinking without in itself establishing a blood alcohol concentration at the time of death [1]. Nor do they reveal how much of the measured concentration came from drinking and how much formed after death, since the two can add up in the same individual [2].

The laboratory also has indicators of microbial activity. The alcohols 1-propanol and 1-butanol, which micro-organisms produce alongside ethanol, cast doubt on the result when they appear on the chromatogram [2, 3, 5]. For that reason 1-propanol, the usual internal standard when blood from living subjects is analysed, is replaced by tert-butanol for autopsy specimens [2]. Acetaldehyde is harder to interpret, as it arises both from fermentation and from the oxidation of ethanol in a poorly preserved sample [1]. Only abnormally high concentrations, found together with higher alcohols, suggest neoformation, and some groups question the reliability of 1-propanol itself [1]. In urine, a raised ratio between two serotonin metabolites, 5-HTOL and 5-HIAA, likewise indicates drinking before death [2, 3].

Securing specimens from autopsy to analysis

Neoformation can continue in the tube after sampling, during transport and storage [1, 2]. Adding sodium or potassium fluoride at 1 to 2%, a higher concentration than in routine clinical tubes, inhibits glycolysis and curbs microbial activity [1, 2]. It should go into every specimen, vitreous humour and urine included [2]. Fluoride does not undo what happened in the body before autopsy, and it is less effective against yeasts such as Candida albicans [2]. Temperature matters just as much. Refrigeration at 4 °C is a minimum, and freezing becomes preferable beyond 48 hours of storage, at −20 °C in routine practice and down to −80 °C for long-term storage [1, 2].

The ethanol present in a sample can also fall over time, notably through oxidation to acetaldehyde, and all the more so when the tube is underfilled or has already been opened [1]. In fluoride-preserved autopsy blood frozen at −20 °C for 6 to 15 months, the mean decrease is of the order of 8 to 10% [2]. The stability studies compiled in the Brazilian review report losses of up to 15 or 16%, but several were carried out on blood from living subjects, whose stability is not necessarily that of autopsy blood [1, 2]. The authors accordingly recommend duplicate samples, in airtight tubes filled at least halfway, so that a never-opened specimen remains available for reanalysis [1]. These precautions are a matter of laboratory technique, yet they directly determine how much confidence can later be placed in the result.

The difficulty is that a result can be perfectly accurate analytically and still be hard to interpret in medico-legal terms.

A figure that has to be read in context

On 28 February 1975, a London Underground train failed to slow as it approached Moorgate station and ran on into a closed-off, unused section of track, killing 43 people including its driver [2]. His body was not recovered for 4 days, from a tunnel where the temperature had risen to 49 °C [2]. Blood taken from several sites gave concentrations of 0.2 to 0.8 g/L, and the figures reached the press before they had been interpreted, with no alternative specimen available to corroborate them [2]. The coroner’s jury returned a verdict of accidental death, and the crushing of the body, the heat, the delay and the fourfold difference between sampling sites all point to post-mortem production [2].

An overestimated post-mortem blood alcohol concentration can lead to a dead driver being wrongly held responsible and can deprive the next of kin of an insurance payout [2]. Conversely, genuine drinking can be underestimated if the sample deteriorated before analysis [1, 2]. Olds and Jones also point out that statutory driving limits were framed for blood drawn under sterile conditions from the living, and that an autopsy result should not be set against them without care [2]. The difficulty is that a result can be perfectly accurate analytically and still be hard to interpret in medico-legal terms. The laboratory determines precisely how much ethanol is present in the sample it receives. That does not answer another question, one that matters more to the investigation, namely where that ethanol came from.

When a post-mortem blood alcohol concentration is central to a case, judges, prosecutors and investigators therefore do well to look beyond the number. They need to know where the blood was drawn, what condition the body was in, whether vitreous humour or urine was analysed, how the tubes were stored, and whether EtG, EtS or putrefaction alcohols were looked for. For decomposed bodies, Olds and Jones suggest that the analytical report should itself state that some or all of the measured ethanol may have formed after death [2].

Références :

[1] da Paz A.P.S., Nunez L.B.B. Jr, de Mello V.J. Biological matrices for determining the origin of post-mortem ethanol, recommendations for collection and storage. International Journal of Legal Medicine, 2026, publié en ligne le 1er septembre 2026. DOI 10.1007/s00414-026-03994-9
[2] Olds M.L., Jones A.W. Preanalytical factors influencing the results of ethanol analysis in postmortem specimens. Journal of Analytical Toxicology, 2024, 48(1), 9-26. DOI 10.1093/jat/bkad078
[3] Lin Z., Wang H., Jones A.W., Wang F., Zhang Y., Rao Y. Evaluation and review of ways to differentiate sources of ethanol in postmortem blood. International Journal of Legal Medicine, 2020, 134(6), 2081-2093.
[4] Thierauf A., Kempf J., Perdekamp M.G., Auwärter V., Gnann H., Wohlfarth A., Weinmann W. Ethyl sulphate and ethyl glucuronide in vitreous humor as postmortem evidence marker for ethanol consumption prior to death. Forensic Science International, 2011, 210(1-3), 63-68. DOI 10.1016/j.forsciint.2011.01.036
[5] Vezzoli S., Bernini M., De Ferrari F. Ethyl glucuronide in vitreous humor and blood postmortem specimens, analysis by liquid chromatography-electrospray tandem mass spectrometry and interpreting results of neo-formation of ethanol. Annali dell’Istituto Superiore di Sanità, 2015, 51(1), 19-27. DOI 10.4415/ANN_15_01_05
[6] Direction de l’information légale et administrative. Alcool au volant. Service-Public.fr, fiche vérifiée le 11 juillet 2025.

Crime scene investigation : is prioritizing the “right” decisions the key to identifying relevant and useful evidence?

Despite the ever-accelerating development of new technologies for the detection, collection, and analysis of traces, the forensic process remains fundamentally characterized by human decision-making. The priorities established and the various forms of triage that define each stage of the process, from the incident scene to the laboratory, have the potential either to facilitate the resolution of an investigation or, conversely, to hinder it. Understanding the factors that influence the choices made by crime scene unit personnel therefore appears essential for the continuous optimization of forensic investigation practices.

The importance of decision-making in crime scene unit work

The quality of crime scene unit work at an incident scene is critical for the remainder of the forensic process. Decisions concerning scene delimitation, documentation of the scene condition, identification of relevant areas to be examined, and the search for and collection of traces constitute some of the earliest decisions within that process. Crime scene examiners therefore represent the principal actors responsible for mobilizing and prioritizing the resources dedicated to scene investigation, with little or no opportunity to revisit the scene once it has been released. In this respect, it is essential to recognize the full significance and relevance, for a crime scene examiner, of making the right – or at least the best possible – decisions at a scene in order to maximize the subsequent use of physical and digital traces. Such decision-making has the potential to influence both the successes and failures of the subsequent stages of the criminal investigation, judicial process, and intelligence development.

First, certain decisions made during scene examination can increase both the effectiveness and efficiency of investigative efforts. Building on Locard’s premise that “time that passes is truth that escapes”, numerous empirical studies have demonstrated the importance of the temporal dimension in the successful resolution of criminal investigations. Generally speaking, the longer an investigation remains unresolved, the lower the likelihood that it will ultimately be solved. A scene investigation that is strategically prioritized rather than unnecessarily laborious is therefore likely to exert a positive influence on case resolution.

Similarly, an investigative approach focused on the search for, detection, and collection of relevant and useful traces – that is, traces that are related to the case under investigation and that provide added value – is likely to have a positive impact on the remainder of the forensic process. Such an approach not only increases the likelihood that collected physical traces will contribute to the identification or exclusion of individuals, the reconstruction of events, and the detection of broader criminal phenomena, but also helps prevent the overburdening of forensic laboratories, which are already operating under considerable demand. Improved prioritization of scene sampling is frequently recommended as a means of reducing delays and backlogs, particularly in forensic biology unit, by avoiding the expenditure of substantial resources on material that essentially constitutes background noise.

Finally, decisions made during incident scene investigation must ensure the systematicity, comprehensiveness, integrity, and impartiality of the approach deployed by crime scene unit personnel. Ultimately, sound decisions in forensic investigation are those that reduce the risk of challenges to forensic practices, minimize criticism in judicial procedures, and decrease the likelihood of wrongful convictions. Conversely, hasty decisions, approaches driven solely by intuition, or investigations guided by unilateral assumptions may have significant adverse consequences for the remainder of the judicial process. For example, an insufficiently comprehensive and systematic search for physical and digital traces may result in relevant evidence being overlooked, while an approach lacking sufficient caution and rigor may increase the risk of contamination or alteration of valuable traces. Such errors are particularly consequential because they are irreversible : once the scene has been released, crime scene unit personnel cannot return to it in its original condition. Traces that escape forensic analysis and interpretation may hinder the resolution of an investigation, prevent the detection of a criminal series, or compromise the successful prosecution of an accused individual, as illustrated by cases such as the 1994 O.J. Simpson case in the United States or the 1986 double homicide at the Bayonne harbourmaster’s office in France, which remains unsolved.

More concerning still, inadequate decisions during the forensic investigation of incident scenes may ultimately contribute to the wrongful conviction of innocent individuals. Several internationally documented cases (including the murder of Meredith Kercher in Italy in 2007 and the John Tomaino case in the United States in 1999) have demonstrated how problematic crime scene processing can contribute to miscarriages of justice. Statistics reported by several researchers estimate that wrongful convictions resulting from forensic science errors – whether attributable to scene examination practices or laboratory procedures – account for approximately 30% to 50% of all miscarriages of justice, depending on the jurisdiction and period under study.

Complexity and systematic noise in the prioritization of forensic efforts

The importance of the decision-making process underlying crime scene investigation supports a proposition that is gaining increasing consensus within the forensic community : the search for, detection, and collection of physical and digital traces is not a purely mechanical task. A series of decision points characterize scientific scene investigation and make the prioritization of forensic efforts particularly complex :

Police officers present at the scene may […] or may not collect physical evidence themselves. They determine whether the attendance of a crime scene examiner is necessary. The examiner may or may not be informed about the offence and about the elements that could most productively be sought. They may arrive sooner or later and collect items or physical traces based on what has or has not been communicated to them. They subsequently decide which items should or should not be analysed. In certain cases, analyses will be performed within a police laboratory, whereas in others they will be submitted to an external laboratory. […] (Roux et al., 2018, p. 3).

In other words, because scene investigation is necessarily constrained (e.g., in terms of available human and material resources), it may vary considerably in its degree of comprehensiveness and systematicity. It may be conducted more or less rapidly and extensively, a wider or narrower range of traces may be sought and detected, and a broader or more limited array of techniques and tools may be deployed. The work of crime scene unit personnel therefore almost always involves some form of triage and prioritization, whereby certain actions are favored over others.

Research conducted over recent decades on forensic practices and performance nevertheless suggests that such prioritization is far from homogeneous. Differences have been observed both at the individual level, between crime scene examiners, and at the collective level, that is, between units or organizations. These divergences in working methods and performance have been documented in interventions involving scenes that were ostensibly similar in terms of offence type and severity, and in some cases virtually identical (see, among others, the work of de Gruijter and colleagues using controlled simulated crime scenes).

According to various studies conducted among crime scene practitioners, scientific scene investigation is characterized by substantial heterogeneity in practice. For example, the average time devoted to burglary scene examinations has been reported as approximately 20 minutes in the United Kingdom, 44 minutes in the United States, and one hour in Switzerland for broadly comparable incidents (see, among others, Baechler, Cartier, Schucany and Guéniat (2015)). A similar phenomenon of variation in scene examination duration has also been documented under experimental conditions where all participants were confronted with the exact same simulated crime scene, further reinforcing the observation of heterogeneous performance.

Similarly, over the past fifteen years, several studies investigating expertise in crime scene examination have identified significant differences in search patterns for physical traces, notably through the use of eye-tracking technologies. Different crime scene examiners do not distribute or focus their attention on the same locations within an incident scene. This observation is consistent with the finding that information-gathering practices are not homogeneous among crime scene unit personnel. Several studies examining the prevalence of confirmation bias in forensic science have demonstrated that some individuals (or groups of individuals) collect substantially more information and intelligence prior to conducting their examinations than others, who instead prefer to avoid or minimize the risk that their decision-making may be influenced by contextual information of varying reliability.

Finally, other research has shown that substantial variation exists in the physical traces collected at incident scenes. Certain categories of incident scenes result in more extensive trace collection than others, and some individuals collect considerably more physical traces than their colleagues. Differences are also observed later in the forensic process, revealing variation in the quality, relevance, and/or utility of collected samples. Indeed, some specialist units reportedly achieve identification rates (e.g., through DNA and fingerprint evidence) that are two to three times higher than those of other units when investigating burglary scenes. At the individual level, traces collected by certain crime scene examiners are submitted for analysis up to twice as quickly as those collected by their colleagues and result in a greater number of positive identifications. According to research conducted in the United States and the United Kingdom, the highest-performing practitioners achieve up to three times as many DNA identifications as their least successful counterparts.

Research in cognitive science and psychology over recent decades has undoubtedly highlighted the scale of bias-related issues in forensic science and crime scene investigation. The studies discussed above further support the need for a better understanding of how such issues affect crime scene unit practice. However, they also point toward a phenomenon of systematic noise that has thus far received comparatively less attention within the scientific community. This concept of noise, notably introduced and developed by Daniel Kahneman, Olivier Sibony, and Cass Sunstein (2021), refers to variability in the judgments of a group of experts (e.g., physicians, judges, economists, recruiters, forensic practitioners) who, when confronted with an identical situation, should theoretically arrive at the same conclusion. In the present context, the concept is particularly useful for understanding decision-making in forensic investigations because it reiterates both the complexity of the task and the broad range of possible configurations within crime scene unit practices. Professor Christophe Champod, of the School of Criminal Justice at the University of Lausanne, notably entitled one of his articles “Research focused mainly on bias will paralyse forensic science” in 2014, thereby highlighting the limitations of an understanding of decision-making based exclusively on bias. Such a perspective encourages the emergence of complementary concepts and approaches, such as systematic noise, for studying the heterogeneity of decisions made by crime scene unit personnel.

The systematic noise documented in the scientific literature is not particularly surprising given that, at present, a substantial proportion of textbooks and manuals intended for crime scene investigation practice focus primarily on cataloguing the procedures and techniques available for detecting, enhancing, collecting, and preserving traces while preventing contamination. In doing so, they address only superficially the question of how priorities should actually be established regarding the search for and collection of traces at a scene.

Factors influencing decision-making in crime scene investigation

Various individual, contextual, and organizational factors are likely to influence the prioritization of efforts and the exercise of judgment by crime scene unit personnel during crime scene investigations. Some of these factors have already been examined empirically, whereas others remain hypotheses that have yet to be formally tested.

For example, several studies in the forensic domain have shown that crime scene unit personnel, like all decision-makers, are susceptible to a range of cognitive biases. A bias, sometimes also referred to as a heuristic, may influence decision-making, appropriately or inappropriately, in a manner that favors a pre-existing hypothesis, belief, or item of knowledge. Restricted access to information may reduce the influence of certain biases, whereas greater integration of information may, conversely, promote a more comprehensive understanding of the context. The timing at which information is communicated to crime scene examiners is likewise recognized as a factor capable of influencing decision-making during crime scene investigations. Indeed, early access to contextual information may orient technical and scientific personnel toward a particular line of inquiry from the outset. Conversely, delayed access to certain information may lead to different interpretations and evaluations of traces and their relevance to the case under investigation. The pace at which information is transmitted, as well as its nature, are therefore acknowledged as factors that influence both initial hypotheses and the prioritization of certain traces during the search process. In the absence of a clear consensus in the literature, the most promising approach appears to lie in achieving a balance between awareness of tunnel vision effects, the communication of information that is genuinely relevant to decision-making (as opposed to information likely to introduce unnecessary bias) and the diligent collection and verification of available intelligence and information in support of rigorous reasoning.

In parallel, the educational background of crime scene unit personnel, particularly whether or not they have completed university-level education, appears to be one of the factors influencing the problem-solving abilities that are essential to crime scene investigation. According to the study conducted by Illes et al. (2019), this may in fact be the factor exerting the greatest influence on the mobilization and application of reasoning skills. By contrast, professional experience – which is frequently used as a recruitment criterion within crime scene unit – appears to play a much more secondary role. Whether personnel are sworn police officers or civilian employees also appears to have only a limited impact on reasoning performance. However, this issue remains insufficiently studied to permit definitive conclusions, particularly because this variable is closely associated with educational attainment, as civilian personnel recruited into police organizations often hold university degrees.

Another factor that warrants consideration is the degree of discretionary power, or latitude, available to crime scene unit personnel when carrying out their duties and making decisions regarding the prioritization of crime scene investigations. On the one hand, in some environments, crime scene unit services are still incorrectly perceived as units whose work consists of little more than mechanical assistance. Under such conditions, crime scene examiners have limited capacity to exercise their professional judgment fully and to adapt decisions to the contingencies of the case under investigation. The direction of the intervention also becomes more dependent upon priorities established by other investigative actors, who generally possess more limited forensic knowledge. On the other hand, in other contexts, crime scene unit personnel are viewed as experts fully integrated into investigations and therefore enjoy substantial discretionary authority regarding both the orientation and nature of scene interventions. This operational autonomy, even when implicit, allows practitioners to base their decisions on the knowledge available to them. However, it also has the potential to generate and/or reinforce heterogeneity in practices and decision-making if individuals prioritize their own personal or idiosyncratic preferences. Consequently, it may create a risk of differential treatment depending on the particular examiner or crime scene unit involved.

Nevertheless, numerous questions remain unanswered with respect to fully understanding the factors that influence decision-making within crime scene unit services and identifying the most effective and appropriate approaches for optimizing crime scene interventions. For example, relatively few empirical studies have directly examined the effects of initial and continuing professional training on the decision-making processes of crime scene examiners. Training format, duration, and the pedagogical tools employed (e.g., virtual reality), to name only a few characteristics, are all factors capable of influencing crime scene investigation practices. Another factor increasingly associated by researchers with decision-making in forensic investigations is the importance of optimally selecting candidates for crime scene unit roles. Several authors have identified and grouped a number of key competencies associated with superior performance including cognitive abilities, communication skills, stress management, life experience, holistic knowledge, professional attitude and a genuine interest in the work itself. However, studies that have rigorously tested these characteristics remain comparatively scarce. Similarly, we still know relatively little about the actual impact of other personnel involved in crime scene interventions on the decision-making processes of crime scene unit personnel. To what extent do patrol officers  or criminal investigators exert a beneficial or detrimental influence on decision-making ? This remains an open question. Finally, it should be emphasized that despite the growing international trend toward implementing standardization and quality assurance mechanisms within crime scene units – one of whose explicit objectives is the harmonization of forensic science practices – there remains limited empirical evidence clearly demonstrating the advantages and/or disadvantages of such managerial strategies in terms of actual performance. This, too, constitutes a factor that may theoretically influence the direction of forensic efforts at a crime scene. However, research has not yet provided definitive answers, notably because governance and quality assurance frameworks specifically developed for forensic science (e.g., ISO 21043) have been implemented only relatively recently.

Conclusion : the importance of synergy between practice and research

Despite the rapid technological evolution of the forensic field, decision-making by crime scene unit personnel remains the foundation upon which the entire process of searching for, detecting, collecting, analysing, and interpreting traces is built. It is therefore imperative to rely on sound decision-making practices rather than placing excessive confidence in the ability of technological tools alone to provide the desired answers. However, gaining a better understanding of how decisions are made during crime scene interventions is not merely a matter of identifying biases and the factors that contribute to the failure of certain investigations. It also represents an opportunity to identify the most effective practices and to propose genuine avenues for optimization aimed at maximizing the contribution of physical and digital traces to both public safety and the administration of justice. Such a refined understanding makes it possible to promote a decision-making process that neither ignores the information available nor accepts it uncritically. Instead, information should be integrated into a rigorous hypothetico-deductive reasoning process, in which working hypotheses are continuously confronted with the intelligence and traces detected throughout the investigation. This perspective also underscores the importance of transparent documentation by crime scene unit personnel, so that it remains possible to understand what actions were undertaken, why those actions were undertaken and upon which information the resulting decisions were based.

In seeking to continuously improve our understanding of this decision-making process, collaboration between researchers and practitioners is essential. For crime scene unit services, such collaboration may take the form of greater openness to exchanges with the research community and improved access to operational data, enabling a better understanding of day-to-day realities and a more accurate assessment of the true impact of various factors on professional practices. Research can therefore contribute meaningfully to reflection on the evolution of professional practices and on the mechanisms required to support their development.

For researchers, this also entails recognizing the value of the expertise developed by crime scene examiners in the investigation of incident scenes and integrating this perspective more effectively into research initiatives. Consequently, beyond studies focused on identifying potential biases and developing theoretical models – which remain essential research objectives – there is growing interest in research specifically designed to address the needs of crime scene unit services and to work collaboratively with practitioners in identifying the factors that positively and negatively influence performance. In this respect, understanding how decisions are made in operational environments is not a secondary exercise but rather the key to ensuring more effective approaches to crime scene investigation. Because the robustness of the forensic process is determined by its weakest link, sound decision-making at the very beginning of that process is essential, and its optimization depends largely upon a continuous dialogue between practice and research.

Références

  • Adderley, R., & Bond, J. W. (2008). The Effects of Deprivation on the Time Spent Examining Crime Scenes and the Recovery of DNA and Fingerprints. Journal of Forensic Sciences, 53(1), 178-182. https://doi.org/10.1111/j.1556-4029.2007.00634.x
  • Baechler, S., Cartier, D., Schucany, P., & Guéniat, O. (2015). Les interventions de la police scientifique suite à des cambriolages: Quelle est la perception des lésés et y a-t-il lieu de s’en soucier? Revue Internationale de Criminologie et de Police Technique et Scientifique, 68(2), 228–247.
  • Baechler, S., Morelato, M., Roux, C., Margot, P., & Ribaux, O. (2020). Un modèle continu, non linéaire et collaboratif de l’enquête. Criminologie, 53(2), 43–76. https://doi.org/10.7202/1074188ar
  • Buzzini, P., Kammrath, B. W., & Forest, P. D. (2019). Trace evidence? The term trace from adjective to noun. Wiley Interdisciplinary Reviews: Forensic Science, 0(0), e1342. https://doi.org/10.1002/wfs2.1342
  • Casey, E., Ribaux, O., & Roux, C. (2019). The Kodak Syndrome: Risks and Opportunities Created by Decentralization of Forensic Capabilities. Journal of Forensic Sciences, 64(1), 127–136. https://doi.org/10.1111/1556-4029.13849
  • Chowdhury, M. (2021b). Beyond bagging and tagging—An empirical investigation into the roles, designations and responsibilities of crime scene practitioners. Science & Justice, 61(3), 271–279. https://doi.org/10.1016/j.scijus.2021.01.002
  • Cooper, G. S.; Meterko, V. (2019). Cognitive Bias Research in Forensic Science: A Systematic Review. Forensic Science International, 297, 35–46. https://doi.org/10.1016/j.forsciint.2019.01.016
  • Delémont, O., Bitzer, S., Jendly, M., & Ribaux, O. (2017). The practice of crime scene examination in an intelligence- based perspective. In Q. Rossy, D. Décary-Hétu, O. Delémont, & M. Mulone (Eds.), The Routledge International Handbook of Forensic Intelligence and Criminology (pp. 86–101). Routledge International Handbooks.
  • Howes, L. M., Julian, R., Oosthuizen, T., Reid, C. A., Cashman, K., & White, R. (2022). Redrawing the Boundaries of Criminology: Increasing Forensic Literacy by Including Forensic Studies in the Curriculum. Journal of Criminal Justice Education, 0(0), 1–19. https://doi.org/10.1080/10511253.2022.2096249
  • Illes, M., Wilson, P., & Bruce, C. (2019). Forensic epistemology: Testing the reasoning skills of crime scene experts. Canadian Society of Forensic Science Journal, 1–23. https://doi.org/10.1080/00085030.2019.1664260
  • Kahneman, Daniel, Olivier Sibony, and Cass R. Sunstein. Noise: A flaw in human judgment. Hachette UK, 2021.
  • Ludwig, A., & Fraser, J. (2014). Effective use of forensic science in volume crime investigations: Identifying recurring themes in the literature. Science & Justice, 54(1), 81–88. https://doi.org/10.1016/j.scijus.2013.09.006
  • Mousseau, V. (2024). L’orientation des stratégies de recherche de traces matérielles par les techniciens en identité judiciaire : Entre savoirs occupationnels et pouvoir(s) discrétionnaire(s) [Thèse de doctorat]. Université de Montréal.
  • Roux, C., Willis, S., & Weyermann, C. (2021). Shifting forensic science focus from means to purpose: A path forward for the discipline? Science & Justice, 61(6), 678–686. https://doi.org/10.1016/j.scijus.2021.08.005
  • Thomson, G. S. (2018). Criminal Dismemberment: Determining and Defining Commonly Available Cutting Implements and the Analysis of Kerf Morphology on Proxy Mediums.
  • Tilstone, W. J., Hastrup, M. L., & Hald, C. (2019). Fisher’s Techniques of Crime Scene Investigation First International Edition. Auerbach Publications. https://doi.org/10.1201/b12825
  • Williams, R. (2004). The management of crime scene examination in relation to the investigation of burglary and vehicle crime. Home Office.