| Abstract: |
The authors contend that there are several limitations in attempting to predict demographics of the past populations. While the title suggests that anthropologists should just cease doing paleodemographic works, they really just suggest that there needs to be a better way of doing it. We should not rely upon reference populations with age classes of about the same size. Further there is a need to come up with techniques which better assess the true age of adults at death. Not being able to assess adults as well as subadults is due to 2 reasons: 1) The error in age determination due to the amount of variability of the human skeleton in the aging process. 2) the range of age groups going into the structure of deaths to be estimated. The wider the class range, the lower the number of misclassified skeletons. The standard error is a measure of error in prediction. The STE is not STD of an individual observation (Xi) but rather of the sample mean. The STE is a measure of the variability of sample means obtained from repeated random samples of size \in\i drawn from the same population. STE is directly proportional to both 1/srt \in\i and to the population standard deviation of an individual observation. It justifies the concern with sample size in assessing the accuracy our mean estimates of the unknown population mean. The reason it is preferable to estimate population mean from a sample size of 400 rather than from one of 100 since the first sample will be one-half as large as in the second sample. Thus, the larger sample would provide more accurate estimate of the population. Notice that the accuracy of our estimate is also affected by the underlying methodological accuracy producing the individual observations from the population, a factors which unrelated to sample size. Reference population - the set of skeletons of a known age to which we compare our graveyard populations so as to calculate their ages at death. A major problem in assessing the age structure of a population from a reference population is that 1) the mean in the reference population is not wholly biological in nature and 2) the mean age depends to a large extent on the age structure of the reference population.
|
