- This event has passed.
“Identification of Deceased Border Crossers: Investigating Spatial and Skeletal Attributes of Migrant Deaths” by Caitlin C.M. Vogelsberg
August 27, 2018 @ 1:00 pm - 3:00 pm
Join us as Caitlin C.M. Vogelsberg defends her dissertation “Identification of Deceased Border Crossers: Investigating Spatial and Skeletal Attributes of Migrant Deaths”
Monday, August 27, 2018
A-438 East Fee Hall
International migration research has primarily focused on cultural, sociological, and economic components. Understanding geographic mortality patterns and the skeletal attributes among the deceased using a mixed-model framework can provide insight into route selection for irregular entrants via the Mexican border. This research investigates the spatial and skeletal properties of deceased undocumented border crossers (UBCs) recovered in the southern Arizona desert and the relationship between recovery location and country of origin.
Previous research investigating spatial patterns in the distribution of identified UBCs recovered in the jurisdiction of the Pima County Office of the Medical Examiner (PCOME) in Tucson, Arizona demonstrated positive spatial autocorrelation between individualizing attributes (such as biological sex and country of origin) and their recovery location (Vogelsberg 2018). Those results indicated several influencing factors, such as country of origin, on final recovery location. This research project combines spatial patterns and cranial skeletal indicators of geographic origin to improve country of origin prediction during the identification process.
An optimized global linear model developed using craniometric and macromorphoscopic factors for a sample (n = 25) of identified Mexican and Guatemalan individuals analyzed at the PCOME was incorporated into several geographically weighted regression (GWR) platforms to predict country of origin. The best performing GWR analysis accounted for just over half of the variation in the data (R2= 0.540). This is an increase from the global model (R2 = 0.432) which did not incorporate recovery location and attributes of other individuals found nearby. Other indicators of model goodness-of-fit show more accurate country of origin predictions using the GWR method.
Model testing on individuals with presumptive Mexican identifications (n = 8) resulted in the correct allocation for country of origin for two individuals and provided promising results for future application. Although sample sizes were small, the potential for applying mixed-model methods is clearly demonstrated. As more individuals are identified and added to the model reference sample, the utility of this predictive method will improve. Furthermore, the application of these techniques to situations in which the physical location of an individual might correlate with their personal attributes is demonstrated.
This research provides the forensic and humanitarian community with supplemental information to aid in the investigation of undocumented border crossers recovered from the southern Arizona desert. Enhancements to the identification process by better directing missing persons searches may increase the number of identified individuals and the return of their remains to their families.