I have supervised one student, Geva Maimon who has done work in the area of

forensic statistics at both at the Master’s and Ph.D. level (both co-supervised by

James Curran from the University of Auckland, who is an expert in applications of

statistics to forensic statistics). Glass evidence is amongst the most common types

of trace evidence used in forensics. Geva’s Master’s thesis work resulted in a paper

which presented a method for determining the extent of non-spatially structured

heterogeneity in plate glass refractive index, which will allow forensic scientists to

properly account for such heterogeneity when attempting to match glass evidence

from a crime scene to that found on a potential suspect. For Geva’s Ph.D. thesis,

we are solving problems encountered in the analysis of forensic DNA evidence. In

particular, we first developed a sequential Monte Carlo algorithm to assist in the cal-

ibration of PCR (polymerase chain reaction) techniques that allow for more accurate

genotyping of human tissue and blood samples by taking into account stutter peaks,

which result from imperfect amplification of the true sample DNA. The second part

of Geva’s thesis focuses on improving analysis of the raw electropherogram data (the

pre-processed data from the PCR process) by summarizing genotype profiles with

functional estimates rather than the current practice of summarizing profiles merely

by peak height. We hope to show that the method will be particularly useful in the

case of DNA mixtures, where one or more suspects DNA material is mixed in the

evidence with that of the victim (or victims).