Read Edit History




Obtaining a part of a larger whole, a population, for analyses. There are many ways to sample a population, random and selective, where the usual goal is to optimize the representation.

Full text

An important aspect in archaeology and statistics is that of sampling. Usually, and especially in archaeology, you are not dealing with a complete dataset (it is unlikely you’ll have access to all tombs from the cemetery you are studying, all surface material of a city you are surveying, all altitudes of a valley that you are mapping, etc.) but with a smaller subset of items. Basically almost all material studied by archaeologists are samples of a population (the population here means the whole group or set of items). E.g. if you’d study a set of excavated burials, they are a sample of a complete cemetery, which would be the population, about which you probably want to say something.

The main issue with sampling, which makes it so important for archaeology, is representation. To be able to say something about the population based on items in a sample, you should optimize the chance that you have a sample that shows a comparable degree of diversity to the items of the population. E.g. if you’d study the social structure of a given settlement population by an examination of burial ritual, you should try to include the burial rituals of all biological and social groups in the settlement. Of course, since, apart from the biological groups, part of these social groups will only be known after your study, it is really hard to know for sure whether you included all relevant groups or not. Sampling is all about reducing the possibility of such exclusion.

Therefore, in archaeological studies it is important to be conscious of this principles. When you are taking samples of a site or region (whether by surveying, test trenches or prospective techniques like georadar) you should devise a sound sampling strategy. If you are dealing with data that has been collected by someone else, which is often the case, you should estimate the representation of the sample in relation to the population which you study to understand its potential.

Research topics: Software & Technology