Development and evaluation of a goodness-of-fit test for comparing within-source distributions of scores suitable for small sample sizes
Law, Probability and Risk · 05.02.2026 · 25 (1)
DOI: 10.1093/lpr/mgag008 ↗Kurz gefasst
Fachartikel in der Zeitschrift Law, Probability and Risk, online erschienen am 03.07.2026. Verfasst von Janean L Hanka und 4 weiteren Autor:innen. Die Arbeit nennt 23 Literaturangaben. Gefördert von: FBI Laboratory Division, the National Institute of Justice, Office of Justice Programs, U.S. Department of Justice.
Abstract (Originalsprache)
In the forensic source identification problem, two objects (typically from known and questioned sources) are compared; statistical methods using a score function applied to all pairs of objects may be used for this evaluation. Due to the limited number of samples per source typical of forensic data, a usual assumption is that all within-source scores follow the same distribution of scores and, therefore, can be pooled together to estimate a common within-source distribution of scores. However, this may not be a reasonable assumption, and published tests do not work well for the small sample sizes (per source) commonly encountered in forensic science. Here, we propose and develop a test for deciding if two within-source distributions of scores should not be considered statistically similar, which is applicable to small sample sizes. To complement the developed test, we present an ad-hoc strategy for pooling sources that accounts for significant discrepancies between the within-source distributions of scores. We then evaluate the Type I error and power of this test on several univariate and multivariate distributions. Finally, we apply the test to a forensic research data set consisting of micromorphometric measurements of aluminum powders.
Förderung
- FBI Laboratory Division
- the National Institute of Justice, Office of Justice Programs
- U.S. Department of Justice
- National Science Foundation
- National Research Traineeship
- Cyber-Physical-Social Systems for Understanding and Thwarting the Illicit Economy
- South Dakota Board of Regents and Governor’s Office of Economic Development
- Governor’s Research Center Program and the Center for Understanding and Disrupting the Illicit Economy
Zitieren
APA
Hanka, J. L., Ommen, D. M., Hietpas, J., Buscaglia, J., & Saunders, C. P. (2026). Development and evaluation of a goodness-of-fit test for comparing within-source distributions of scores suitable for small sample sizes. Law, Probability and Risk, 25(1). https://doi.org/10.1093/lpr/mgag008
Vancouver
Hanka JL, Ommen DM, Hietpas J, Buscaglia J, Saunders CP. Development and evaluation of a goodness-of-fit test for comparing within-source distributions of scores suitable for small sample sizes. Law, Probability and Risk. 2026;25(1). doi:10.1093/lpr/mgag008
BibTeX
@article{Hanka2026,
author = {Hanka, Janean L and Ommen, Danica M and Hietpas, Jack and Buscaglia, JoAnn and Saunders, Christopher P},
title = {Development and evaluation of a goodness-of-fit test for comparing within-source distributions of scores suitable for small sample sizes},
journal = {Law, Probability and Risk},
year = {2026},
volume = {25},
number = {1},
doi = {10.1093/lpr/mgag008}
}
EUVIMED