Reference population bias

What is reference population bias

• When identification methods are built on one groups skeletal data but applied to individuals from underrepresented populations

Spradley et al (2008)

• Most identification criteria on US forensic databases are based on black and white Americans

• Hispanics now comprise 1 in 8 Americas

‘ Hispanic’ as a ancestry category

• Problematic - it is a linguistic definition

• Group has highly diverse genetic admixture so is difficult to build reference population

Sex estimation in Hispanic populations

• Methods calibrated on white populations correctly identify Hispanic females but not males

• Thus cross-group application produces sex estimation errors

Stature problems in Hispanic populations

• Black and White american’s are on average taller than Hispanics and variation exists within Hispanic groups

• Population specific formulae are needed

Poor skeletal databases

• Many populations are poorly represented as historical collection practices favoured certain groups

• Thus ‘newer’ groups lack comparable reference data

Cunha & Ubelaker (2019) - Value of ancestry

• Despite database limitations ancestry remains useful for linking to law enforcement databases

• Practitioners must recognise many populations are underrepresented

Ross et al (2004) →Cuban study

• Used to show even within Hispanic groups there is great variation

• Cubans make up Florida’s largest Hispanic community and that the unique population history of Cubans makes them more similar to American Blacks