Chapter 6: Research and development
Scientific basis for forensic science
152.There are three main areas where there has been increasing scrutiny of the scientific base for forensic science:
(1)The scientific validity of the approaches used to identify the source of a material or mark, and the challenges in addressing complex mixed provenance samples.178
(2)The need to understand better the activity of materials to aid interpretation of forensic science evidence (i.e. the activities that led to the generation or transfer of those materials), and their implications for reaching conclusions when reconstructing crime events.179
(3)Awareness of the importance of human decision-making in the forensic science process and the challenges of identifying factors which can affect judgments (such as cognitive bias; the appropriate use of statistics and probabilities to convey evidential significance; and how forensic science evidence is presented in court).180
Forensic science analysis methods and ‘source attribution’
153.In response to a critical report in 2009 in the United States by the National Research Council,181 President Obama commissioned a study in 2015 to examine the scientific validity of different forensic science methods, including:
- DNA analysis of single-source and simple-mixture samples,
- DNA analysis of complex-mixture samples,
- bitemarks,
- latent fingerprints,
- firearms identification, and
- footwear analysis.
The resulting report182 found that many of these methods did not meet the scientific standards for foundational validity. This is concerning because the methods are routinely used as evidence in court.
154.There are concerns about methodology in pattern recognition and in the analysis of trace materials for ‘source attribution’.
155.In regard to pattern comparison methods, Professor Niamh Nic Daéid, Director of the Leverhulme Research Centre for Forensic Science at Dundee University, explained that the underpinning science for “DNA analysis, toxicology or the measurement of drugs”183 was considered strong. However, “the comparison of fingerprints, toolmarks, footwear, tire marks and ballistics” were “spot-the-difference” techniques in which “there is little, if any, robust science involved in the analytical or comparative processes used and as a consequence there have been questions raised around the reproducibility, repeatability, accuracy and error rates of such analysis.”184
156.Concerns have been raised about the ability of experts to interpret accurately the results of scientific tests. As the sensitivity and resolution of analytical capabilities have increased, it has become increasingly common to identify multiple components in a single trace or specimen, often described as ‘mixed source’ samples. In the then Government Chief Scientific Adviser’s annual report for 2015–16, Sir Mark Walport explained:
“new capabilities create other challenges for our existing systems; in particular, our ability to analyse may outstrip our ability to interpret. Because we can identify very small traces of a substance, we need greater certainty in understanding their significance and better ways to communicate different levels of confidence.”185
These concerns were echoed in the Forensic Science Regulator’s annual reports in 2015,186 2016,187 2017,188 and 2018189 which identified a lack of data to underpin the evaluative interpretation of traces.
157.The concerns in relation to the scientific standards, scientific reliability, methodology, comparison and interpretation expressed by the US National Research Council and others led to some initiatives being taken in the United Kingdom to address these concerns. These included the work by the Law Commission culminating in its Report on Expert Evidence in Criminal Proceedings in 2011, work by the Royal Society (including the work on primers described at paragraph 131 and an international conference in February 2015), a grant from the Leverhulme Foundation of the Research Centre for Forensic Science at Dundee University and the establishment (after initial work by the Chief Scientific Adviser at the Home Office, Sir Bernard Silverman,) of the Science and Justice Forum by Sir Mark Walport (then Government Chief Scientific Adviser) and Sir Mark Sedwill (then Permanent Secretary at the Home Office). What was lacking was strategic oversight which was essential for the reasons and purposes we have set out in Chapter 2.
Forensic science interpretation: assessing the ‘activity level’
158.We heard that there is a lack of research to help experts assess the ‘activity level’ of materials. This refers to the way in which materials, such as DNA and other traces, may transfer between objects and how long they can be expected to remain there.
159.The Metropolitan Police Service illustrated this:
“Advancement in DNA recovery and analysis techniques has provided greater sensitivity and changed the interpretation of the evidence. A practical example of such a gap is in DNA transfer. Whilst there are many published papers on this, further basic research would inform the ability of the scientist to interpret DNA results in the context of transfer, for example, the likelihood that DNA from a surrounding area can be transferred onto an item. Similarly, in fingerprint comparison, there are some gaps in understanding activity level reporting.”190
160.The implications of being able to address how and when a material was generated or transferred were highlighted by the Science and Justice Research Interest Group at Northumbria University, who stated that “when examining occasions in which forensic science has been implicated in a wrongful conviction, often it is because the scientific evidence was relied upon to answer ‘activity level’ questions, when it was unable to do so”.191
Forensic science interpretation: human judgement and decision making
161.Human judgement and decision making are integral to every stage of the forensic process (see figure 1). Human decisions are vulnerable to bias; the issue of bias in forensic science evidence interpretation was raised by many witnesses192 as well as in a number of reports, such as the 2016 report by the President’s Council of Advisors on Science and Technology.193
162.We were told by Dr Itiel Dror about empirical studies that showed that the conclusion in some areas of forensic science could be unduly subjective and influenced by human factors.194
163.This is an important challenge. The University of Leicester stated that it “requires more research focused on human factors in forensic science, including better understanding of the cognitive process of pattern recognition, the psychological nature of ‘expertise’, and sources, causes, and consequences of cognitive bias.“195
164.Dr Christopher Lawless told us that “the use of statistical methods for evidence interpretation has been developed and discussed within the forensic scientific community for some time.”196 However, Professor Norman Fenton indicated that “the statistical aspects of forensic evidence are often either simply overlooked (because they are considered too difficult) or poorly presented by both lawyers and forensic scientists.”197
165.There are difficulties in interpreting forensic science from the use of statistics and probabilities to assign weight and significance. There have been initiatives to increase awareness of the use of statistics, such as the primer from the Royal Statistical Society and the guide from the Inns of Court College of Advocacy (see para 132).198 However, we heard that the challenges are exacerbated by a lack of research into the likelihood of certain activities resulting in the generation or transfer of forensic materials. In addition, Lord Hughes of Ombersley, a former Justice of the Supreme Court, told us that “there is very little in the way of scientific research that helps with evaluating the significance of a match once you have discovered one. How significant the match is depends in the end on what the chances of a random match are. That means research and databases.”199
Research and development
Funding
166.We heard of the desirability of investing in research and development to address technological developments in and foundational research into forensic science. Witnesses referred to the difficulties in obtaining funding for forensic science research. The UCL Centre for the Forensic Sciences wrote that while there were limited “opportunities for short-term, tender-based, funding to develop technology to be deployed within 12 months to the crime scene”, it was harder to fund “mid- to long-term projects, which can develop foundational research that leads to innovations and deployable solutions in the future (e.g. 10–20 years’ time)”.200
167.One reason that funding for forensic science projects is difficult to obtain is that there is no unit of assessment for forensic science in the Research Excellence Framework. This means that there is, according to Professor Dame Sue Black, “no great drive from the universities in terms of their quality of research to put forensic science at the top of that.”201
168.The formalised way in which academic research funding is allocated, and the focus on innovation, may also disadvantage forensic science. Dr Geoffrey Stewart Morrison told us that “forensic science research often focusses on applying existing science and technology to forensic problems and on empirical validation under casework conditions, foci which the granting agencies’ and journals’ reviewers often do not perceive as innovative.”202 Angus Marshall said that the result was sometimes that researchers “find that the best way they can make progress is to disguise their work, “piggy back” it onto another project, or undertake it as consultancy activity through casework, which may result in an embargo or prohibition on publication of their results.”203
169.Despite recommendations over a number of years by the House of Commons Science and Technology Committee204 that forensic science should be a research priority, Rebecca Endean from UK Research and Innovation (UKRI) told us that over the last 10 years only £56 million had been spent on 150 studies205 relating to forensic science. This accounted for a “relatively small percentage” of their overall expenditure in that time, with the “annual expenditure of UKRI over that 10-year period [being] roughly £6 billion.”206 The percentage is less than 0.1%. The list of projects UKRI referred to in supplementary written evidence included under the category of forensic science many projects which, on analysis, did not address forensic science research questions, had little forensic science content or which referred to forensic science as one of many possible applications of the research.207
Research by commercial providers
170.Given the difficulties in obtaining funding for forensic science research in the academic sector, it has been suggested that more research could be carried out by private providers. However, barriers to research in the private sector are just as high, if not higher, than in the academic sector.
171.As explained in Chapter 3, private providers are struggling to remain profitable; in these conditions research and development is deprioritised. Dr Mark Pearse, Commercial Director of Eurofins Forensic Services, told us that Eurofins Forensic Services regards “no area as being at the right level of profitability to sustain reinvestment in innovation”.208 Professor Carole McCartney thought that in the current market conditions the private sector could not be expected “to provide blue-sky thinking and invest in research and development”.209
172.The result was that where research and development was carried out by private providers “it is often focussed on development and improvement of core services to meet primary customer needs, rather than on transformational innovation.”210 ADS Group explained that without “wider changes to the funding and commissioning of forensic science the government must inevitably step up its own forensic science research through [the Defence Science and Technology Laboratory] or other bodies.”211
Research gaps
173.As well as concerns about the scientific basis for some forensic methods (see paragraphs 152–165), we heard about other research gaps.212 Almost every forensic science sub-discipline has areas that evidence suggests would benefit from further research. It is clear that forensic science needs more sustained and coordinated funding for research and development in both technological developments and foundational research.
174.The lack of coordinated strategic thinking around research and development in forensic science means that there is not an established process for legal practitioners and, to an extent, forensic service practitioners to outline the areas they think need more research.
175.The judges who gave evidence to us were keen to see more research on “evaluating the significance of a match once you have discovered one”,213 “researching transfer and persistence”214 and building better databases, an area which has suffered since the demise of the Forensic Science Service.215
Digital forensics
176.Digital forensics is a rapidly growing field. While it faces many of the same challenges as other forms of forensic science evidence, given the speed of developments and the volume of material that is now routinely produced, it also faces specific challenges. In Chapter 5 we considered the challenges of analysing digital evidence in the timescales demanded in the criminal justice system and stressed the need for further investment in the use of modern technology. There is little research into techniques to sift and analyse data such as artificial intelligence and machine learning. Paul Hackett, Group Managing Director at Key Forensic Services Ltd, told us, “Who is pushing the technology drive in artificial intelligence in digital forensics in the UK? … Nobody.”216
177.Care must be taken with the use of artificial intelligence and machine learning. Dr Jan Collie explained that “human biases might be replicated by some of these machine-learning systems” and that “with artificial intelligence, it is very hard to explain what happened and how the machine came up with a particular answer”.217
178.There is also an issue in relation to disclosure of the methods used. It is generally essential that there is disclosure so that the methods used are open to scrutiny and peer review of its accuracy and reliability. However, claims of commercial confidentiality are sometimes made by some private providers who are unwilling “to disclose information about their own development and testing methods [which] means that the evidence base for the correctness of many digital methods is extremely weak or non-existent. “218 Such claims can rarely provide a proper justification for the withholding of disclosure of the methodology, when the results obtained through the methodology are to be used in a criminal trial. However, there are circumstances where the prosecution contends that the disclosure of the methodology will give rise to risks to national security on the basis that disclosure may enable terrorists or criminals to take measures to avoid detection or may compromise an investigation; it will be for the court to decide whether the withholding of disclosure is justified.219
179.Paul Harris explained there were issues with determining search terms for a machine to use: “In a criminal case, there is often a whole hidden undercurrent of different words to describe drugs, drop-offs, firearms and things like that, which makes conducting these types of searches harder.”220
180.While mindful of these limitations, it is clear that investigators will be unable to keep up with demand for analysis of digital evidence without some technological assistance. Mark Stokes told us that a modern mobile phone:
“could have 1 terabyte of data on it, which is 78 million documents or pages of information on one mobile device, and it is becoming impossible for an investigator to review, disclose, analyse, view and read all that information. Therefore, artificial intelligence and machine learning both have a part to play in this, but we have to be very careful in the application of these technologies. We need academia and science to work with us to do the testing and validation.”221
181.Just as with traditional forensic science, research and development in digital forensics needs coordination and strategic planning. It has been made clear to us that this is lacking.222
A new strategic research body
182.There is a clear case for establishing an overarching body with responsibility for directing research into and funding of forensic science. The Forensic Science Service funded some research before it was disbanded.223 Since then, there has been no national research programme for forensic science.
183.Rebecca Endean agreed that a strategic oversight body for research would be “very useful” and cited the Office for Strategic Coordination of Health Research (OSCHR) as a good example of a similar body. OSCHR “brings together the [National Institute for Health Research], [UK Research and Innovation], the devolved Administrations and practitioners—who are really important in this agenda—to talk. That works quite well. You could see something like this—not controlling the money but deciding what the priorities and the gaps are.”224
184.Lord Hughes of Ombersley recommended that this new body should be a Royal College of Forensic Scientists.225 Keele University suggested it should be an “independent national institute”, which should be “quite separate from any university and would need to be government funded. By working closely with both the police, the judiciary and through links with research groups in universities, the institute would be able to focus on long term priority projects and provide consultancy expertise on immediate problems.”226 The institute could “act as a focal point and be able to coordinate inter-agency research and other activities amongst police forces, universities, private commercial providers and indeed various legal groups and bodies.”227
185.In designing a research body, England and Wales could look at Australia and New Zealand’s National Institute of Forensic Science which, as UCL Centre for the Forensic Sciences wrote, “has the strategic intent to promote and facilitate excellence in forensic science through … promoting, sponsoring and supporting research in forensic science in areas of identified strategic importance, and supporting, coordinating and conducting training programmes … for practitioners228.”229
186.There would need to be a dedicated stream of funding, perhaps through UKRI and industry, to deliver a strategic programme of high-quality research that addresses technological developments and foundational research in the short, medium and long term.
187.Current levels of investment in forensic science research are inadequate and do not appear to reflect value to the criminal justice system. We believe that the Home Office has abdicated its responsibility for research in forensic science. We recommend that UK Research and Innovation urgently and substantially increase the amount of dedicated funding allocated to forensic science for both technological advances and foundational research, with a particular focus on digital forensic science evidence and the opportunities to develop further capabilities in artificial intelligence and machine learning.
188.We recommend the creation of a National Institute for Forensic Science within the UK Research and Innovation family, to set strategic priorities for forensic science research and development, and to coordinate and direct research and funding. This body should work closely with the police, the judiciary, universities, private forensic science providers and the Forensic Science Regulator to fulfil these duties. It should be accountable to UK Research and Innovation who should submit an annual report on the activities of the National Institute for Forensic Science to the Forensic Science Board.
178 Executive Office of the President, President’s Council of Advisors on Science and Technology, Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods (September 2016): https://obamawhitehouse.archives.gov/sites/default/files/microsites/ostp/PCAST/pcast_forensic_science_report_final.pdf [accessed 5 February 2019]
179 Forensic Science Regulator, Annual Report November 2016–November 2017, (19 January 2018): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/674761/FSRAnnual_Report_2017_v1_01.pdf [accessed 21 February 2019]
180 Forensic Science Regulator, Guidance: Cognitive Bias Effects, Relevant to Forensic Science Examinations, (FSR-G-217, Issue 1) (2015): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/510147/217_FSR-G-217_Cognitive_bias_appendix.pdf [accessed 21 February 2019] and Executive Office of the President, President’s Council of Advisors on Science and Technology, Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods
181 Committee on Identifying the Needs of the Forensic Science Community, National Research Council of the National Academies, Strengthening Forensic Science in the United States: A Path Forward
182 Executive Office of the President, President’s Council of Advisors on Science and Technology, Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods
185 Government Office for Science, Forensic Science and Beyond: Authenticity, Provenance and Assurance, Annual Report of the Government Chief Scientific Adviser 2015 (2015), p 6: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/506461/gs-15-37a-forensic-science-beyond-report.pdf [accessed 5 February 2019]
186 Forensic Science Regulator, Annual Report November 2014–November 2015, (4 December 2015): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/482248/2015_FSR_Annual_Report_v1_0_final.pdf [accessed 5 February 2019]
187 Forensic Science Regulator, Annual Report November 2015–November 2016, (6 January 2017): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/581653/FSR_Annual_Report_v1.0.pdf [accessed 5 February 2019]
188 Forensic Science Regulator, Annual Report November 2016–November 2017, (19 January 2018): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/674761/FSRAnnual_Report_2017_v1_01.pdf [accessed 5 February 2019]
189 Forensic Science Regulator, Annual Report November 2017–November 2018 (15th March 2019): https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/786137/FSRAnnual_Report_2018_v1.0.pdf [accessed 27 March 2019]
191 Written evidence from Science and Justice Research Interest Group (RIG) and Northumbria University (FRS0051)
192 Written evidence from Professor Gary Edmond, University of New South Wales (FRS0022), The Chartered Society of Forensic Sciences (FRS0025), UCL Centre for the Forensic Sciences (CFS) (FRS0041), Northumbria University Forensic Science Unit Research Interest Group (FRS0050), Keith Borer Consultants (FRS0061), University of Portsmouth Forensic Innovation Centre (FRS0058), Forensic Access (FRS0066), Dr Geoffrey Stewart Morrison (FRS0074), University of Leicester (FRS0082) and Inns of Court College of Advocacy (ICCA) (FRS0089)
193 Executive Office of the President, President’s Council of Advisors on Science and Technology, Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods
198 ICCA, RSS, Statistics and probability for advocates: Understanding the use of statistical evidence in courts and tribunals (2017)
204 See, Science and Technology Committee, Forensic Science Strategy (Fourth Report, Session 2016–17, HC 501), Science and Technology Committee, Forensic science: Volume I (Second Report, Session 2013–14, HC 610), and Science and Technology Committee, The Forensic Science Service: Volume I (Seventh Report, Session 2010–12, HC 855)
205 This number was revised to £56 million spent on 150 studies in the supplementary written evidence from UK Research and Innovation (UKRI) (FRS0105)
212 See, for instance, written evidence from Institute of Traffic Accident Investigators (ITAI) (FRS0023), Alan Turing Institute (FRS0030) and International Union of Geological Sciences (IUGS) (FRS0033)
219 Royal Courts of Justice, Kelly, R v England and Wales Court of Appeal, (EWCA) Crim 1893 (2018)
227 Ibid.
228 Australia New Zealand Policing Advisory Agency (ANZPAA), The Peak Body for Forensic Science in Australia and New Zealand (March 2019): http://www.anzpaa.org.au/forensic-science/about [accessed 25 February 2019]