The 89th Annual Meeting of the American Association of Physical Anthropologists (2020)


Correlation Between Saw Blade Width and Kerf Width

MELISSA D. MENSCHEL and JAMES T. POKINES.

MS Forensic Anthropology, Boston University

April 16, 2020 , Platinum Ballroom Add to calendar

Previous studies of saw mark analysis primarily focused on using qualitative methods to determine saw class and addressed their potential for identifying suspect saws utilized in cases of dismemberment. The present study examined the extent to which metric analysis can correlate saw blades with saw marks based on measurements of kerf width. A sample of 56 partially defleshed white-tailed deer (Odocoileus virginianus) long bones were utilized as proxy for human remains. The specimens were cut using a variety of new and used commercially available saws, including 11 hand-powered and 5 mechanical-powered saws. A total of 500 false start kerfs were measured using digital calipers to ascertain whether their minimum kerf width measurements could be used to determine the type of saw used to produce those cuts. Two experiments were performed, with the first test examining the kerf widths of false starts produced on specimens restrained using clamps, while the second test analyzed the kerf widths of false starts produced on minimally restrained specimens.

Statistical analysis using Hierarchical Linear Modeling found that the average kerf width (mm) was statistically significantly larger for comparatively thicker blades (p < 0.001), with average kerf width 29% larger than blade width. Additionally, variables other than saw blade thickness influence kerf width; for example, the average kerf width was statistically significantly larger for used blades compared to new blades (p = 0.007). However, a comparison of hand to mechanical saws (p = 0.50) and unrestrained cuts (p = 0.45) were not statistically significantly different.

Funding for this project was provided by the MS Forensic Anthropology program at Boston University School of Medicine. 


Slides/Poster (pdf)