Pattern Evidence

Developing Statistical Foundations for Pattern Evidence

Most forms of pattern evidence are not built on an agreed-upon underlying quantitative framework. Forensic scientists typically reduce crime scene pattern evidence to a set of features to examine and analyze, which leaves room for inconsistencies and varying certainties — and lessens the strength of the forensic evidence. DNA evidence, with an underlying biological science and probabilistic model, provides a roadmap for how forensic science, data and statistical methods can enable forensic scientists to best draw conclusions from pattern evidence.

CSAFE researchers work to develop explicit probability models for the measurements obtained from forensic pattern evidence and to use these models to draw conclusions about the probative value of pattern evidence. Research is underway to:

  • Develop a taxonomy of acceptable probabilistic/statistical reporting of forensic evidence.
  • Develop a statistical modeling framework for different forms of pattern evidence (fingerprints, blood splatter and tool marks).
  • Develop approaches to measure quality of pattern evidence, including design of proficiency tests.
  • Integrate physical science components in plausible probabilistic models for different types of pattern evidence.
  • Design a sampling plan to assemble a database of tool marks, beginning with firearms, and assemble the database of tool marks.

What constitutes pattern evidence?

Pattern evidence includes any markings produced when one object comes into contact with another object:

  • Bloodstains
  • Firearms and toolmarks
  • Latent prints/Tenprints
  • Questioned documents
  • Shoeprints
  • Tire treads

Pattern evidence also involves the evaluation of handwriting, typewriting and writing instruments.


Vital to developing plausible probabilistic models for the different kinds of pattern evidence, we collaborate with:

  • Quantitative researchers, including statisticians, computer scientists and engineers
  • Forensic practitioners
  • Partners at universities, medical examiner labs and crime labs
  • National groups including the Innocence Project
  • Materials scientists
  • Chemists studying surface properties of materials