The characterization of Monte Carlo errors for the quantification of the value of forensic evidence

A multivariate extension to the standard 4σ criterion for comparison of forensic glass evidence

This manuscript presents a more accurate and easy to implement multivariate generalisation of the international standard 4σ forensic glass comparison technique. Many crimes result in glass breaking, and the broken glass found at a crime scene can be important forensic evidence. The chemical composition of this glass can be measured to establish whether it can be distinguished from glass fragments found on a suspect's clothing. The chemical composition can be measured using laser ablation-inductively coupled plasma mass spectrometry. A commonly used method to compare fragments of glass is the 4σ interval criterion. This method, however, compares each element individually and does not take advantage of the multivariate nature of this data. We introduce a multivariate extension to this method, which makes use of the correlation structure between the elements. We demonstrate that this method results in an improvement in the false positive rate, with only a small compromise in the false negative rate. The improvement in false positive rate is desirable as false positives translate to misleading evidence against a potentially innocent defendant. The multivariate generalisation improves accuracy while retaining a similar interpretation, and so is suitable to present in court.

Source‐anchored, trace‐anchored, and general match score‐based likelihood ratios for camera device identification

Forensic camera device identification addresses the scenario, where an investigator has two pieces of evidence: a digital image from an unknown camera involved in a crime, such as child pornography, and a person of interest’s (POI’s) camera. The investigator wants to determine whether the image was taken by the POI’s camera. Small manufacturing imperfections in the photodiode cause slight variations among pixels in the camera sensor array. These spatial variations, called photo‐response non‐uniformity (PRNU), provide an identifying characteristic, or fingerprint, of the camera. Most work in camera device identification leverages the PRNU of the questioned image and the POI’s camera to make a yes‐or‐no decision. As in other areas of forensics, there is a need to introduce statistical and probabilistic methods that quantify the strength of evidence in favor of the decision. Score‐based likelihood ratios (SLRs) have been proposed in the forensics community to do just that. Several types of SLRs have been studied individually for camera device identification. We introduce a framework for calculating and comparing the performance of three types of SLRs — source‐anchored, trace‐anchored, and general match. We employ PRNU estimates as camera fingerprints and use correlation distance as a similarity score. Three types of SLRs are calculated for 48 camera devices from four image databases: ALASKA; BOSSbase; Dresden; and StegoAppDB. Experiments show that the trace‐anchored SLRs perform the best of these three SLR types on the dataset and the general match SLRs perform the worst.

Measuring calibration of likelihood-ratio systems: A comparison of four metrics, including a new metric devPAV

Numerical likelihood-ratio (LR) systems aim to calculate evidential strength for forensic evidence evaluation. Calibration of such LR-systems is essential: one does not want to over- or understate the strength of the evidence. Metrics that measure calibration differ in sensitivity to errors in calibration of such systems. In this paper we compare four calibration metrics by a simulation study based on Gaussian Log LR-distributions. Three calibration metrics are taken from the literature (Good, 1985; Royall, 1997; Ramos and Gonzalez-Rodriguez, 2013) [1-3], and a fourth metric is proposed by us. We evaluated these metrics by two performance criteria: differentiation (between well- and ill-calibrated LR-systems) and stability (of the value of the metric for a variety of well-calibrated LR-systems). Two metrics from the literature (the expected values of LR and of 1/LR, and the rate of misleading evidence stronger than 2) do not behave as desired in many simulated conditions. The third one (C_llr^cal) performs better, but our newly proposed method (which we coin devPAV) is shown to behave equally well to clearly better under almost all simulated conditions. On the basis of this work, we recommend to use both devPAV and C_llr^cal to measure calibration of LR-systems, where the current results indicate that devPAV is the preferred metric. In the future external validity of this comparison study can be extended by simulating non-Gaussian LR-distributions.

Evaluation of glass evidence at activity level: A new distribution for the background population

For evidence evaluation of the physicochemical properties of glass at activity level a well-known formula introduced by Evett & Buckleton [1,2] is commonly used. Parameters in this formula are, amongst others, the probability in a background population to find on somebody's clothing the observed number of glass sources and the probability in a background population to find on somebody's clothing a group of fragments with the same size as the observed matching group. Recently, for efficiency reasons, the Netherlands Forensic Institute changed its methodology to measure not all the glass fragments but a subset of glass fragments found on clothing. Due to the measurement of subsets, it is difficult to get accurate estimates for these parameters in this formula. We offer a solution to this problem. The heart of the solution consists of relaxing the assumption of conditional independence of group sizes of background fragments, and modelling the probability of an allocation of background fragments into groups given a total number of background fragments by a two-parameter Chinese restaurant process (CRP) [3]. Under the assumption of random sampling of fragments to be measured from recovered fragments in the laboratory, parameter values for the Chinese restaurant process may be estimated from a relatively small dataset of glass in other relevant cases. We demonstrate this for a dataset of glass fragments collected from upper garments in casework, show model fit and provide a prototypical calculation of an LR at activity level accompanied with a parameter sensitivity analysis for reasonable ranges of the CRP parameter values. Considering that other laboratories may want to measure subsets as well, we believe this is an important alternative approach to the evaluation of numerical LRs for glass analyses at activity level.