Biology Reference
In-Depth Information
America. All of the individuals were born during the twentieth century and died in 2005 or
2006. The sample in this study consisted of 126 individuals (44 females, 81 males) with
a mean age of 47 years. The statistical methods used included univariate discriminant func-
tion analysis with cross-validation and univariate analysis of variance (ANOVA) using the
statistical software program SPSS
. In cross-validation, each case is classified by the func-
tions derived from all cases other than that case. Only the bones of the left side of the
body were included in the analysis in order to lessen the effect of handedness and potential
occupational markers, especially in the upper limb.
The results of this study indicated the same general pattern of classification effectiveness
as seen in the North American sample by
Spradley and Jantz (2011)
using the univariate
ANOVA and cross-validated univariate discriminant function analysis. All of the variables,
except for the humeral epicondylar breadth, showed statistically significant sexual dimor-
phism. The cross-validated univariate discriminant function analysis demonstrated that
the best accuracy rates were for the scapular height and humeral head at 86% each. The scap-
ular breadth came in third with an accuracy rate of 84.6%. Surprisingly, the accuracy rate for
the femoral head (80.8%) was lower than that from the
Spradley and Jantz (2011)
study
(86
88%), probably reflecting population specific behavior (
Moore et al., 2012
).
The results of this study could be extremely useful for identification, as the humerus is
more resistant to taphonomic properties than many of the bones (
Moore et al., 2012
). Further-
more, the ability to achieve such a high degree of success from a single bone is more efficient
and thus preferable for the fast-paced human identification laboratories in Colombia that see
hundreds of cases each year. This research has the potential to play an important role in the
development of population standards for Colombia and Latin America and may help
provide a better standard of practice for forensic anthropology that can pass scrutiny by
the Colombian judicial system and international criminal courts.
While our results were somewhat preliminary because of the relatively small sample size,
the Colombian Modern Skeletal Collection has now more than tripled, enabling a reevalua-
tion of this study. This case study illustrates that important research need not be those that
are novel and unique. We used the
Spradley and Jantz (2011)
paper as a model and followed
their methods to set up methods for our own study. This enabled us to generate the first
population-specific metric sexing standards for a population. We recommend that the reader
consider a similar option when contemplating a research project, be it in sex estimation or
any other subject.
e
CONCLUSION
One way to develop a research project in sex estimation (or in any subject) is to test the
results of a previously published research project. This is how science advances. This also
makes developing the methodology easier, because you simply follow the same format in
the original research project. This is extremely valuable when trying to develop population
standards and serves to simplify (and demystify) the process for a young researcher. There
is no need to reinvent the wheel. In order to begin your analysis of sex estimation, first deter-
mine whether there is a significant difference between males and females in the trait for
which you are estimating. Seriation can help you better visualize the variation in the trait
