Biology Reference
In-Depth Information
The first phase of morphoscopic trait analysis is the selection of the character states that
best match the configuration exhibited by the unknown specimen (see example below).
This is completed for each observable trait. Following this stage, appropriate classification
statistics (e.g., ordinal regression, logistic regression, canonical analysis of principal coordi-
nates (CAP), k-nearest neighbor (k-nn), discriminant function analysis) and suitable refer-
ence populations are selected. Once the statistical analysis has placed the unknown
specimen into a population, the probability of group membership for the unknown spec-
imen, and the overall error rate (misclassification rate or classification accuracy) of the model
are reported along with the assigned group membership. This approach is likely familiar, as
it is the same reporting strategy used in metric analyses.
The computer programMacromorphoscopics, designed specifically for the collection of mor-
phoscopic trait data, is available from
Hefner and Ousley (2005)
and is also incorporated in
the computer program Osteoware that is available at no charge from the Smithsonian Insti-
tution (
https://osteoware.si.edu/
). These programs facilitate data management and also
standardize trait descriptions. To put the entire process in perspective the following example
cranium is presented (
Figure 5.1
).
The morphoscopic traits for this cranium were scored following
Hefner's (2009)
illustra-
tions and definitions. The following trait scores are noted: (1) the anterior nasal spine is
well-developed and markedly protrudes from the face (ANS
¼
3); (2) the inferior nasal aper-
ture is consistent with the straight morphology (INA
¼
3); (3) interorbital breadth is interme-
diate (INA
2); (5) the nasal bones
exhibit steep lateral walls, with an accompanying broad surface plateau (NBC
¼
2); (4) nasal aperture width is intermediate (NAW
¼
¼
2); (6) nasal
overgrowth is pronounced (NO
¼
1); and (7) no post-bregmatic depression is observed
(PBD
0). These visual observations alone are enough for the experienced forensic anthro-
pologist to make an educated guess of ancestry.
However, recall that stopping at this point is unempirical and therefore not scientific. We
must go a step further and apply a statistic to assess the overall classification. In this example,
originally conducted for a medical examiner's office, I (JTH) used a novel classification
¼
FIGURE 5.1
Cranium for case study: assessing ancestry for an unknown.
