Much has been written about how Bayesian interpretation can be applied to different evidence types over the last fifteen years, and lately there have been attempts to adopt the approach right the way through the forensic process, in some cases even starting before items are submitted, and in other cases ending with the scientist making up a defence scenario on behalf of the defendant.
Here in the UK, the Forensic Regulator has recommended that accreditation to 17025 (and further to a new forensic standard to be published in due course) is the only standard that forensic laboratories need. The Forensic Regulator was also of the view that individual practitioners employed by organisations who hold the standard do not need anything further, as merely by virtue of their employers status as an accredited laboratory this ensures their competence.
Part of the 17025 standard requirement is for proficiency testing to occur on a regular basis. If the test is externally administered it is considered to be an excellent means of evaluating the performance of a laboratory and by implication it is a measure of the laboratory systems and the competence of its staff.
There are very few commercial providers of forensic proficiency testing in trace evidence in the world. The oldest and most accepted is CTS in the US. They produce and administer proficiency tests over a wide range of evidence types every year. A report on each test is produced and published, the identities of the laboratories are kept anonymous by means of assigning each lab a unique code.
The CTS fibre proficiency test tends to follow a set format that tests the laboratory's ability to identify and compare fibres from three samples. For this part of the test there are three possible outcomes either the samples match, don't match, or the findings are inconclusive. In reality, for laboratories undergoing the test there are only two outcomes - either you get it right or you get it wrong. Get it right and the lab's systems are deemed in good working order, and its staff are deemed competent (at least those who took the test anyway), get it wrong and the accredited lab has to investigate and take action according to their process. Getting it wrong is as serious an issue as it gets for an accredited lab and its staff.
The second part of the test is of general interest but does not count as a proficiency in real terms because there is no expected outcome. The participant lists the techniques used and their conclusions. There is no right or wrong answer for these and the participants are under no obligation to enter any data at all here.
This year's report on the fibre proficiency test 10-539 has just been published. As all accredited laboratories in the UK are likely to have performed this test, and how all are likely to have used the Bayesian approach, the publication of the report provides an opportunity for an assessment of how labs interpret fibre evidence in this way. Each lab appears to have got the same (correct) scientific result with respect to the matching fibres and all were given precisely the same information.
The scenario given was straight forward, a sample from the inner lining of a jacket, and two clumps of fibres, one found on the inner door latch on the passenger side of a truck, the other on the driver's seat of the truck. After testing, only the clump on the inner door latch was found to be indistinguishable from the lining of the jacket.
Here's the conclusions that had Bayesian elements from this year's test, they can't all be right, you can decide which approach you like best:
Lab 1
Lab 2
Lab 3
Lab 4
Lab 5
Lab 6
Lab 7
Lab 8
Lab 9
Lab 10
Lab 11
But what is clear is that given the same scenario and the same results, that the approach to Bayesian interpretation of fibre evidence varies widely from one laboratory to another, and by implication, one scientist to another. These issues are not unique to fibre evidence. Using Bayesian interpretation as a standard approach in this way does provide a common language for us as forensic scientists to communicate our findings. We just don't seem to be very good at it. It would appear that the common language does more to separate us than unite us.
Is it a matter of competence of the individual, the system of peer review, or a cultural approach by the laboratory? Either way, current quality systems do not address it. But perhaps this is where the regulatory line should be drawn, and that the courts, not the regulatory bodies, are better placed to test matters of individual opinion, after all it is the individual scientist, not the organisation that employs him that is on the stand.
Here in the UK, the Forensic Regulator has recommended that accreditation to 17025 (and further to a new forensic standard to be published in due course) is the only standard that forensic laboratories need. The Forensic Regulator was also of the view that individual practitioners employed by organisations who hold the standard do not need anything further, as merely by virtue of their employers status as an accredited laboratory this ensures their competence.
Part of the 17025 standard requirement is for proficiency testing to occur on a regular basis. If the test is externally administered it is considered to be an excellent means of evaluating the performance of a laboratory and by implication it is a measure of the laboratory systems and the competence of its staff.
There are very few commercial providers of forensic proficiency testing in trace evidence in the world. The oldest and most accepted is CTS in the US. They produce and administer proficiency tests over a wide range of evidence types every year. A report on each test is produced and published, the identities of the laboratories are kept anonymous by means of assigning each lab a unique code.
The CTS fibre proficiency test tends to follow a set format that tests the laboratory's ability to identify and compare fibres from three samples. For this part of the test there are three possible outcomes either the samples match, don't match, or the findings are inconclusive. In reality, for laboratories undergoing the test there are only two outcomes - either you get it right or you get it wrong. Get it right and the lab's systems are deemed in good working order, and its staff are deemed competent (at least those who took the test anyway), get it wrong and the accredited lab has to investigate and take action according to their process. Getting it wrong is as serious an issue as it gets for an accredited lab and its staff.
The second part of the test is of general interest but does not count as a proficiency in real terms because there is no expected outcome. The participant lists the techniques used and their conclusions. There is no right or wrong answer for these and the participants are under no obligation to enter any data at all here.
This year's report on the fibre proficiency test 10-539 has just been published. As all accredited laboratories in the UK are likely to have performed this test, and how all are likely to have used the Bayesian approach, the publication of the report provides an opportunity for an assessment of how labs interpret fibre evidence in this way. Each lab appears to have got the same (correct) scientific result with respect to the matching fibres and all were given precisely the same information.
The scenario given was straight forward, a sample from the inner lining of a jacket, and two clumps of fibres, one found on the inner door latch on the passenger side of a truck, the other on the driver's seat of the truck. After testing, only the clump on the inner door latch was found to be indistinguishable from the lining of the jacket.
Here's the conclusions that had Bayesian elements from this year's test, they can't all be right, you can decide which approach you like best:
Lab 1
Door latch on the passenger side of the suspect's truck: Orange polyester fibres (item 2) of two types, indistinguishable from the two types of polyester fibres used in construction of the jacket lining (item 1),were found on the inner door latch on the passenger side of the suspect's truck. In interpreting the scientific findings I have considered two alternative propositions: The jacket lining (item 1) had been in contact with the inner door latch on the passenger side of the suspect's truck. The jacket lining (item 1) had not been in contact with the inner door latch on the passenger side of the suspect's truck and the polyester fibres originated from another item(s) constructed of the same two types of indistinguishable orange polyester fibres. In my opinion, the scientific findings provide strong scientific support for the former proposition rather than the latter. In expressing the evidential significance of my findings, I have used the following scale: no scientific support, limited, moderate, moderately strong, strong, very strong and extremely strong scientific support. Driver's seat of the suspect's truck: In my opinion, the fibres (item 3) recovered from the driver's seat of the suspect's truck did not originate from the lining of the jacket (item 1).
Lab 2
I have considered the following propositions for the finding of the orange polyester fibres of two types (item 2) indistinguishable from the two types of polyester fibres that comprised the lining of the victim's jacket (item 1) on the inner door latch of the passenger side of the suspect's truck: The victim had been inside the vehicle and fibres were transferred from the victim's jacket to the door latch. The victim had not been inside the vehicle and the fibres found originated from another item(s) composed of the same two types of indistinguishable fibres. In my opinion, the findings provide strong scientific support for the former proposition rather than the latter. In my opinion, the fibres recovered from the driver's seat (item3) did not originate from the lining of the victim's jacket.
Lab 3
These findings provide very strong support for the suggestion that the fibres from the inner door latch on the passenger's side of the suspect's truck came from the inner lining of the victim's jacket. These findings show that the fibres from the driver's seat of the suspect's truck did not come from the inner lining of the victim's jacket. I have chosen the above phrases from the following scale; weak support, moderate support, moderately strong support, strong support, very strong support and extremely strong support. The conclusion was addressed at source level due to the limited samples received. In a case situation the jacket and further samples from the truck would have been available, in which case the activity level
could have been addressed.
Lab 4
The orange polyester fibres of Item 2 were indistinguishable in terms of their microscopical appearance and colour from the orange polyester fibres used in the construction of Item 1. In assessing the laboratory findings and drawing my conclusions, I have considered the following two alternative propositions: The fibres recovered from the inner door latch of the passenger side of the truck, Item 2, originated from the inner lining of the jacket, Item 1. The fibres recovered from the inner door latch of the passenger side of the truck, Item 2, did not originate from the inner lining of the jacket, Item 1. In my opinion, the laboratory findings are more likely if the first proposition were true rather than the second alternative. In my opinion, the laboratory findings provide very strong scientific support for the proposition that the fibres recovered from the inner door latch of the passenger side of the truck, Item 2, originated from the inner lining of the jacket, Item 1.
Lab 5
Sixteen fibres from the driver's seat (Item 3) was recovered for examination by comparison microscopy. No fibres matched the component fibres of the jacket inner lining (Item 1). Sixteen fibres from the inner door latch (Item 2) were recovered for examination by comparison microscopy. A sample of these fibres were further analysed by microspectrophotometry and infrared spectroscopy. All were indistinguishable from the component fibres of the jacket inner lining. These findings would support the proposition that there had been direct contact between the inner door latch and the lining of the jacket.
Lab 6
In our opinion, orange polyester fibres from a yarn recovered from the passenger side inner door latch of the suspect's truck (Item 2), correspond to the orange polyester fibres of the victim's inner jacket lining. In addition, the presence of the orange polyester fibres in the form of a yarn, in our opinion, means that it must have been shed from a damaged garment, such as the victim's jacket. In assessing the evidential significance of these findings, we have considered two possible explanations for the matching fibres: (1) The yarn of orange polyester fibres from the passenger side inner door latch of the suspect's truck, is from the inner lining of the victim's jacket; Or, (2) The yarn of orange polyester fibres from the passenger side door latch of the suspect's truck is NOT from the inner lining of the victim's jacket, but originates from another damaged garment that sheds yarns that happen to match the orange polyester fibres of the jacket lining by coincidence. In our opinion, the orange polyester fibres of two types found on the passenger side door latch of the suspect's truck are relatively uncommon. The presence of these two fibre types matching the inner lining of the victim's jacket in our opinion provides greater evidence of a link between the victim's jacket and the passenger side inner door latch of the suspect's truck, than would have been the case with an apparent transfer of a single fibre type alone. Therefore in our view this makes the first explanation for the findings far more likely than the second. In our opinion, the fibre transfer evidence therefore provides strong scientific evidence of a link between the victim's jacket and the inner door latch on the passenger side of the suspect's car. No fibres matching the component fibres of the victim's jacket lining were found within Item 3.
Lab 7
The findings in this case, and the information provided, have been used to address the following propositions: a) The damaged lining of the victim's jacket has been in contact with the inner door latch of the suspect's vehicle. b) Some other item left the fibres on the inner door latch of the suspect's vehicle and any matching fibres recovered are due to chance. Taking the results of the examinations in this case together with published studies and casework experience, in my opinion, the likelihood of finding two distinct fibre types of an orange colour randomly in the population is fairly low. In addition, given that these fibres would shed only due to the lining being damaged, that there are two distinctly different fibres comprising the warp and weft of the fabric, and that both of these fibre types were recovered from the inner door latch and are found to be indistinguishable in terms of colour, dye and chemical composition, to the polyester fibres of the warp and weft of the jacket lining then, in my opinion, there is very strong support for the view that the damaged lining of the victim's jacket has been in contact with the inner door latch of the suspect's vehicle rather than the fibres transferred originating from some other item by chance. The term "very strong support" is selected from a scale of standard terms used to express the relative level of scientific support for a proposition over its alternative, as discussed above. These terms are: Limited, Moderate, Moderately Strong, Strong, Very Strong, Extremely strong Additionally, in some instances, a proposition may be conclusively supported, if the findings are such that the alternative can be dismissed. If the findings provide no greater support for one proposition over the other, then the findings are described as inconclusive. My interpretations and conclusions are based on the background information available to me at present. If any of this information changes, or is found to be incorrect, then I may need to reassess my findings.
Lab 8
Item 1 comprised a square of orange fabric with one glossy side and one matte side. The fabric was composed of orange, delustered, polyester warp fibres with hexagonal cross section and orange, non-delustered, trilobal polyester weft fibres woven together to form a 2:1 twill. Item 2 comprised orange, delustered, polyester fibres with hexagonal cross section. These fibres corresponded in appearance, composition and colour with fibres from the warp threads in Item 1. The results support the proposition that the fibres recovered from the inner door latch on the passenger side of the suspect's truck shared a common origin with the known section of fabric from the inner lining of the victim's jacket (Item 1). Item 3 comprised orange, multilobal acetate fibres. These fibres differed in appearance, composition and colour with the fibres from Item 1. The results do not support the proposition that the fibres recovered from the driver's seat of the suspect's truck shared a common origin with the known section of fabric from the inner lining of the victim's jacket (Item 1)
Lab 9
The sample from the victim's jacket (item 1) contained two types of orange polyester fibres. The sample from the inner door latch of the suspect's truck (item 2) contained two types of orange polyester fibres. These fibres could not be distinguished from the fibres from the victim's jacket (item 1) by microscopy, colour comparison and infrared spectroscopy. The sample from the driver's seat of the suspect's truck (item 3) contained orange acetate fibres. These fibres were different to the fibres of the victim's jacket (item 1) and therefore are not from the victim's jacket. In my opinion the scientific evidence strongly supports the proposition that the fibres from the inner door latch of the suspect's truck (item 2) are from the victim's jacket (item 1). I have chosen the term "strongly supports" used in describing the scientific evidence above from the following scale of terms: inconclusive, slightly supports, supports, strongly supports, very strongly supports and conclusive.
Lab 10
The forensic significance of fibres is partly dependant on the number of fibres present, and how common that fibre type and/or colour is in textile manufacturing. The fabric in Item #1 was constructed from two different types of orange polyester fibre. The orange polyester fibres from Item #2 were indistinguishable from one type of the fibres in Item #1. The number, colour and length of the fibres present strongly supports the hypothesis that the fibres in Item #2 originated from the same source as Item #1. The fibres in Item #3 were different to those from Item #1 and, as such, could not have originated from the same source.
Lab 11
Based on visual, chemical and colour analysis the two types of orange polyester fibres recovered from the inner door latch on the passenger's side of the suspect's truck (item 2) could have come from the inner lining of the victim's jacket (as represented by item 1), or from a textile containing fibres with the same visual, chemical and colour features. Based on visual, chemical and colour analysis the fibres recovered from the driver's seat of the suspect's truck (item 3) could not have come from the inner lining of the victim's jacket (as represented by item 1). In my opinion, this fibre evidence supports the proposition that the inner lining of the victim's jacket (item 1) came into contact, either directly or indirectly, with the door latch on the passenger's side of the suspect's truck. Note that as fibres are mass produced another source of fibres with the same visual, chemical and colour is possible.Some use two propositions, some using one, some using scales of support some not. Some attach a very strong level, others a strong level, others no level at all. Some place the victim inside the vehicle, others include the possibility of indirect transfer. Some used different scales. Some believe that the scientific findings can be used to show support for the fibres coming from the victim's jacket rather than any other item made of fibres indistinguishable to it, others believe that the scientific findings can be used to show support for the fibres coming from the victim's jacket rather than via contact with any other item by chance. These are not the same propositions.
But what is clear is that given the same scenario and the same results, that the approach to Bayesian interpretation of fibre evidence varies widely from one laboratory to another, and by implication, one scientist to another. These issues are not unique to fibre evidence. Using Bayesian interpretation as a standard approach in this way does provide a common language for us as forensic scientists to communicate our findings. We just don't seem to be very good at it. It would appear that the common language does more to separate us than unite us.
Is it a matter of competence of the individual, the system of peer review, or a cultural approach by the laboratory? Either way, current quality systems do not address it. But perhaps this is where the regulatory line should be drawn, and that the courts, not the regulatory bodies, are better placed to test matters of individual opinion, after all it is the individual scientist, not the organisation that employs him that is on the stand.
