Page 94 - IJPS-7-2

International Journal of
Population Studies Modeling archaeological mortuary assemblages

If we relax the power level cut off for statistical It is interesting that the cohort specific differences
significance to P = 0.05 or even P = 0.1 for the z-score test documented for MTC do not correspond perfectly with
results at the level of age-groups, the number of age intervals the statistically significant differences for results from CI’s
that are statistically significant increases. This increase is demographic data. This indicates that preferred ages of
more noticeable among the MTC cohorts, which would sacrificial individuals may not be consistent throughout the
then have significant differences for more than half the Maya area. The weak difference noted between the MTC
age intervals in the 5 , 50 , and 95 percentiles. Given the models could also be due to a variety of cultural choices
th
th
th
small sample sizes involved and the investigative nature of relating to sacrifice at the site. For instance, if sacrificial
the analysis reported here, that relaxation may be justified. preference changed over time or a more randomized
selection process for sacrificial individuals, then these
4. Discussion choices may be contributing to why the models are not
By examining which age cohorts have statistically different, but a number of differences are noted at specific
significant differences, it becomes possible to rule out age intervals.
particular causes and further verify that these assemblages Before the development of the types of statistical
were most likely the work of sacrifice. Modern examples of analysis employed in this research, bioarchaeologists could
warfare demography are seen in the cases of mass deaths in look at overall patterns of mortality curves and determine
Palestine and Srebrenica (Bosnia), where civilian mortality if mortality for a site was normal or not. Data that reflected
was the greatest for individuals between the ages of 15 and a J-shaped or “bath-tub shaped” curve demonstrated a
25, with the highest numbers being in 20s (Brunborg et relatively normal distribution with high infant mortality
al., 2003; Radlauer, 2002). However, this does not explain that decreases before rising again for older individuals
the differences between children and adults. The elevated (Weiss, 1973). This pattern is seen in Mesoamerican
number of children is not typical for pre-industrial societies mortuary assemblages interpreted as burials at Caves
(Chamberlain, 2006, p64: Figure 2), nor is it expected for Branch Rockshelter, Belize (Glassman & Bonor Villarejo,
periods of famine (Chamberlain, 2006, p72: Figure 3). 2005) and Teotihuacan, Mexico (Storey, 1992). However,
For both the MTC and CI assemblages, at varying this is not the type of pattern demonstrated by MTC or
percentiles of the modeled distribution, we have noted CI. The analysis and modeling introduced by this article
statistical differences among the very young children, older provide a visual and statistical means to determine if an
children, young adults, and older adults. These age cohorts assemblage is normal for a horticultural society. The
tend to be those ethnohistorically chosen for sacrifice specific age intervals where statistical differences were
(older children and young adults) (Fuentes & Guzmán, noted also allowed for affirmation that sacrifice could have
1932; Roys, 1943; Scholes & Roys, 1968; Tozzer, 1941), contributed to the formation of the assemblage.
and those not typically targeted for sacrifice (older adults). Because the two different assemblages did not
These cohort specific statistical differences indicate that demonstrate the same statistical differences, it suggest that
their presence or absence in the assemblage is observable. there may not always be an obvious way to statistically
However, these differences do not necessarily show levels demonstrate differences between mortuary assemblages
of significance at high powers. This is especially when the accumulated through natural processes and those that
models for the site of MTC are compared to each other. accumulate due to human sacrifice. Demographic
The cohort-specific differences for MTC and CI also relationships are complex, and not every site or assemblage
aid in ruling out the possibility that the bimodal age is bound to be created or influenced by processes in
distribution of the assemblages is due to some sort of the same manner. Overall, the number of statistical
“accident hump” within the once living population that differences demonstrated across the percentiles of the
attributed to these assemblages. Gage and Mode (1993) modeled distributions at the age interval level as well as
note that when “accident humps” are noticeable they the comparison of models (Tables 3-5) leads us to reject
should only have miniscule mortality increases. As the the null hypothesis that either the MTC or CI assemblages
simple Z-score test results indicate (Tables 3-5), there accumulated as the result of a standard mortality
are statistically significant differences among specific process experienced within a traditional, horticulturalist
cohorts for both MTC and CI compared to what a normal population. Thus, the demographic analysis can be used
horticultural society should display. These significant to further support the archaeological and ethnohistoric
differences rule out the possibility that these are miniscule evidence from these sites that indicate sacrifice as a
mortality increases and thus dismisses the likelihood of an contributing cause of death for the people represented by
accident hump. these assemblages.

Volume 7 Issue 2 (2021) 88 https://doi.org/10.36922/ijps.v7i2.300