Hidden multiple comparisons increase forensic error rates

Area of Science:

• Forensic Science
• Criminalistics
• Evidence Analysis

Background:

• Toolmark analysis, including striations on cut wires, is used in forensics to link evidence to its source.
• Existing methods for toolmark comparison, like those for bullets, are well-established.
• Wire comparisons present unique challenges not fully addressed by current forensic practices.

Purpose of the Study:

• To identify and address the complexities inherent in comparing cut wire surfaces.
• To highlight the 'multiple comparison problem' in forensic wire examination.
• To relate this issue to similar problems in other forensic disciplines, such as database searches.

Main Methods:

• Analysis of the process of comparing striation marks on two distinct wire cut surfaces.
• Theoretical examination of the statistical implications of multiple comparisons in this context.
• Comparison with established forensic methodologies for toolmark analysis.

Main Results:

• Comparing two wire cut surfaces inherently involves multiple distinct comparisons.
• This multiplicity of comparisons significantly increases the expected false discovery rate.
• The issue is analogous to the multiple comparison problem encountered in forensic database searches.

Conclusions:

• The practice of comparing cut wire surfaces introduces statistical complexities that can inflate error rates.
• The 'multiple comparison problem' is a critical consideration for the reliability of wire striation analysis.
• Further research and methodological adjustments are needed to mitigate these risks in forensic wire examination.