Metric Sex

I've never been particularly good at math, in fact the reason I had to take the GRE so many times was because I couldn't get my math score up to a competitive level. Embarrassing, but true. So color me surprised when I find myself preferring and appreciating metric analysis to determine sex and ancestry. If only metrics could give a better idea about aging, but I think there we are stuck with morphological characteristics for awhile.

Morpho traits have been the main determinant in sex and ancestry estimation. Some of these can be incredibly easy to discern especially when they deal with the pelvis; the presence or lack of a subpubic angle, the depth and breadth of the greater sciatic notch, the shape of the ventral face of the pubic bone.


(Image borrowed from http://scienceblogs.com/afarensis/)

But people are variable and geographic groups differ in their amount of sexual dimorphism. Even within the most studied groups in America (European and African descent groups), there can be individuals who are extreme outliers. My favorite story concerning this issue comes from my undergrad forensics professor who attended a conference where one presenter laid out several skulls and asked participants to attribute sex and ancestry. One skull was consistently scored as European-American female based on morphological traits. Then the known determinations were revealed. The European-American female turned out to be an African-American male. This is a pretty big discrepancy to have missed on both counts.

This is why in the forensics class I took in my final semester of grad school, I appreciated the encouragement to try out new ways of determining the biological profile. Around mid-semester someone found a JFS (Journal of Forensic Sciences) article using a scapula metric calculation for sex (Dabbs et al. 55(1) 2010).

This article has become my new favorite method of double-checking my morpho assessment. There are two different methods described in the article. The 5-measurement method uses non-standard measurements from a 1928 article that I couldn't get my hands on, which was a shame because there are no pictures of the exact measurement method and the descriptions could have been more detailed. I was left approximating what I thought was probably the right way to do them. So basically, guessing. The 2-measurement method was much better, standard measurements (breadth and length) that are part of the usual measurements you would do for a case anyway.


This method correctly determined sex for African- and European-American cases, and most importantly, for Hispanic individuals, which have had less skeletal assessment for modern populations and are becoming a bigger presence in forensic cases. The only down side was that the cases were all male, so I can't vouch for the female side of things, though Dabbs et al. had an overall accuracy of 91% for the 2-measurement method. As always, more research is needed.