Bitemarks 357
evidence. This study, called the ABFO Bitemark Workshop No. 4, underscores
the gap between odontologists’ opinion-based testimony and their ability to
correctly identify a defendant as the creator of a bite mark. The median overall
error rate is 12.5%—out of a maximum possible error rate of 27%. Thus, they
were wrong nearly half the time they tried to identify the source of a bitemark.
More specifically, it is their false positive error rate—the tendency to conclude
that an innocent person’s dentition matches the bitemark—that accounts for
the bulk of that overall error rate. On average, 63.5% of the examiners com-
mitted false positive errors across the test cases. If this reflects their perfor-
mance in actual cases, then inculpatory opinions by forensic dentists are more
likely to be wrong than right.” (Affidavit of Dr. Bowers in Brewer v. State^79 )Considering the ABFO position paper, and Arheart and Pretty’s cautions
about the study design, Bitemark Workshop 4 does not appear to have been,
as Dr. Bowers claimed, a “study regarding the reliability of bite mark identi-
fication evidence,” nor did it “produce data on the accuracy of results in bite
mark identification forensic casework.” The allegations that “they were wrong
nearly half of the time” and “63.5% of the examiners committed false positive
errors across the test cases” seem to directly misinterpret and contradict the
information reported by Arheart and Pretty in the published results. In fact,
in their paper Arheart and Pretty reported, “The ROC area calculated by the
non-parametric trapezoidal method is 0.86, a fairly high accuracy, indicating
that the examiners are able to correctly identify the dentition belonging to
a particular bitemark. Bootstrap 95% confidence intervals are 0.82–0.91.”^77
They also wrote, “This survey indicates that analysis of bitemark evidence
is a relatively accurate procedure among experienced forensic odontologists
when the results are examined in combination.”^77 They concluded, “The
results of the present survey indicate that bitemark examination is an accu-
rate forensic technique, at least with cases such as used in this study.”^77 These
are very different conclusions from those reported by Dr. Bowers and those
critics of bitemark analysis who later quoted him.
That the results of the ABFO Bitemark Workshop 4 have not been accu-
r ately repor ted by some i s u nden iable , a nd t he e ffec t of t hos e i nac c u r ate repor t s
has been decidedly negative. The authors of the Forensic Science International
paper correctly stated in closing, “This study, despite its limitations, has opened
the debate into evidence-based forensic dentistry. Forensic odontologists must
ensure that the techniques they employ are backed by sound scientific evidence
and that the decisions they present in Court serve to promote justice and to
strengthen the discipline. Committed to these high professional standards,
the ABFO is proactive in the continuing education of odontologists.”^77
Ultimately, the publication of this information may have a more positive
effect. Because the number of cases in the study was too few, the reasons can
be neit her properly a na ly zed nor clea rly u nderstood f rom t hese data. However,