Modeling, simulating, and comparing biased archaeological mortuary assemblages

Abstract

This paper uses a novel approach to compensate for inherent sampling biases and to compare the age profiles of two ancient Maya sacrificial assemblages to expectations from a model life table for traditional horticultural populations. It seeks to statistically rule out the possibility that either site is accumulated due to a standard mortality process experienced in horticulturalist populations. This analysis utilizes data from Midnight Terror Cave (MTC), Belize and Chichén Itzá (CI), Mexico to compare the observed versus expected death counts by age. Monte-Carlo based estimates of preservation bias were modeled assuming a normal distribution with mean and variance based on expert opinion. This model was used to up-adjust age-specific death counts for both sites to make more robust sample sizes, which were compared to those expected from a model life table at the 5th, 50th, and 95th percentiles of the resampled distribution of preservation bias. At low levels of estimated bias (5th percentile), neither MTC nor CI assemblages could be distinguished from the null-mortality model. At average to higher levels of estimated bias (50th and 95th percentiles), both populations could be statistically distinguished from the null mortality model either across all age intervals or within specific age ranges. After accounting for preservation bias, the findings suggest that both MTC and CI assemblages were unlikely to have accumulated due to a normal mortality pattern experienced within traditional horticulturalist populations, further supporting the ethnographic and archaeological evidence that indicates that the sites are accumulated due to cultural practices related to human sacrifice.