Biology Reference
In-Depth Information
The evidentiary process also should proceed with the osteologist blind to any prior infor-
mation when they do their basic data collection, but once again the contextual information
becomes vitally important in completing the analysis. The evidentiary problem should
always proceed by calculating what is referred to as a likelihood ratio. A likelihood ratio
in this setting is just the probability of getting the observed data if a putative identification
is correct divided by the probability of getting the observed data if the putative identification
is incorrect. How one calculates the probabilities contained in the likelihood ratio is beyond
the scope of this chapter, but Steadman and colleagues (2006) provide a worked example.
Their example is for what could be referred to as an “open” population setting, in that if
the identification was incorrect then they must consider the “population at large” to find
the probability in the denominator. Defining the “population at large” is indeed problematic.
In a “closed” population setting (see Hackman (2009) for a good description of “open” versus
“closed” populations), as for example in a plane crash where the flight manifest enumerates
everyone, the denominator probability is much less problematic. The first author of this
chapter is currently producing a worked example of an application to such a “closed” pop-
ulation setting.
CONCLUSION: WRAPPING IT UP
The paleodemographic world is certainly much more complicated than when the first
author of this chapter published his first journal article ( Konigsberg, 1985 ) over half a lifetime
ago. That paper on the paleodemography of a classic Ohio Hopewell site violated just about
every proscription we have written here and even (shudder) included the use of life tables.
The paper also included a “Table 1” that gave individual age estimates “ahead” of doing any
demographic analysis, precisely what we have argued against doing in this chapter. But with
the benefit of increasing age, and the possible accumulation of wisdom, how would we
approach a paleodemographic analysis now and what would we recommend to others
who undertake this endeavor?
The simple answer to this question, unfortunately, is not simple. Life tables can be conve-
niently calculated within computer spreadsheets and even hazard models can be fit to life
table type data within such spreadsheets 4 . However, the vast majority of the methods we
have described in this chapter cannot be easily handled either within spreadsheets or using
standard statistical analysis packages. Dr. Darryl Holman at the University of Washington-
Seattle has long provided his program “mle,” which is a workhorse for fitting models to
demographic data 5 . Dr. Jesper Boldsen's software for transition analysis (ADBOU), an
advanced technique we have not described here, has recently been made available for down-
loading from the web from Dr. George Milner's website. 6 One of the common opinions
voiced at a number of workshops on paleodemographic methods at the Max Planck Institute
for Demographic Research in Rostock, Germany is that any methods under development
4 See https://netfiles.uiuc.edu/lylek/www/LibbenHaz.xlsx for an example of both.
5 The software and manuals are freely available from http://faculty.washington.edu/djholman/mle/index.html .
6 http://www.anthro.psu.edu/projects_labs/bioarch/bioarch_lab.shtml .
Search WWH ::




Custom Search