Quantifying the strength of palmprint comparisons: Majority identifications with surprisingly low value

Area of Science:

• Forensic Science
• Biometrics
• Pattern Evidence Analysis

Background:

• Friction ridge examiners report palm impression comparisons similarly to fingerprints, but with key differences.
• Palm print analysis involves extensive search, unique minutiae rarity, and orientation challenges not typical in fingerprint comparisons.
• Current US laboratory reporting scales (Identification, Exclusion, Inconclusive) lack calibration against the actual strength of palmprint evidence.

Purpose of the Study:

• To measure the evidential strength of palmprint impressions.
• To calibrate the language used to articulate conclusions in palm print comparisons.
• To explore alternatives to traditional categorical reporting scales.

Main Methods:

• Construction of likelihood ratios using an ordered probit model.
• Utilizing distributions of examiner responses from an error rate study.
• Quantifying the strength of evidence for palmprint comparisons.

Main Results:

• Calculated likelihood ratios indicate that current articulation scales may overestimate evidence support.
• Potential overestimation can be as high as five orders of magnitude for same-source propositions.
• Modest likelihood ratios were frequently calculated, suggesting a need for revised interpretation standards.

Conclusions:

• Likelihood ratios provide a method to calibrate the articulation of evidence in palm print comparisons.
• The findings suggest current reporting scales may not accurately reflect the strength of evidence.
• Likelihood ratios may offer a more objective alternative to categorical reporting scales in forensic casework.