Share Email Print

Proceedings Paper

Statistical significance of trace evidence matches using independent physicochemical measurements
Author(s): Jose R. Almirall; Michael Cole; Kenneth G. Furton; George Gettinby
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

A statistical approach to the significance of glass evidence is proposed using independent physicochemical measurements and chemometrics. Traditional interpretation of the significance of trace evidence matches or exclusions relies on qualitative descriptors such as 'indistinguishable from,' 'consistent with,' 'similar to' etc. By performing physical and chemical measurements with are independent of one another, the significance of object exclusions or matches can be evaluated statistically. One of the problems with this approach is that the human brain is excellent at recognizing and classifying patterns and shapes but performs less well when that object is represented by a numerical list of attributes. Chemometrics can be employed to group similar objects using clustering algorithms and provide statistical significance in a quantitative manner. This approach is enhanced when population databases exist or can be created and the data in question can be evaluated given these databases. Since the selection of the variables used and their pre-processing can greatly influence the outcome, several different methods could be employed in order to obtain a more complete picture of the information contained in the data. Presently, we report on the analysis of glass samples using refractive index measurements and the quantitative analysis of the concentrations of the metals: Mg, Al, Ca, Fe, Mn, Ba, Sr, Ti and Zr. The extension of this general approach to fiber and paint comparisons also is discussed. This statistical approach should not replace the current interpretative approaches to trace evidence matches or exclusions but rather yields an additional quantitative measure. The lack of sufficient general population databases containing the needed physicochemical measurements and the potential for confusion arising from statistical analysis currently hamper this approach and ways of overcoming these obstacles are presented.

Paper Details

Date Published: 10 February 1997
PDF: 17 pages
Proc. SPIE 2941, Forensic Evidence Analysis and Crime Scene Investigation, (10 February 1997); doi: 10.1117/12.266309
Show Author Affiliations
Jose R. Almirall, Metro-Dade Police Dept. (United States)
Michael Cole, Univ. of Strathclyde (United Kingdom)
Kenneth G. Furton, Florida International Univ. (United States)
George Gettinby, Univ. of Strathclyde (United Kingdom)

Published in SPIE Proceedings Vol. 2941:
Forensic Evidence Analysis and Crime Scene Investigation
John Hicks; Peter R. De Forest; Vivian M. Baylor, Editor(s)

© SPIE. Terms of Use
Back to Top