Sampling in Archaeology

Dmanisi Excavations, 2007
Dmanisi Excavations, 2007. Georgian National Museum

Sampling is the practical, ethical method of dealing with large amounts of data to be investigated. In archaeology, it is rarely prudent or possible to excavate all of a particular site, survey all of a particular area, or extensively analyze all of the soil samples or potsherds you collect. So, how do you decide where to expend your resources?

Key Takeaways: Sampling in Archaeology

Sampling is a strategy that an archaeologist uses to investigate a region, site, or set of artifacts. 

A proper strategy allows her to gain a critical understanding of her data while preserving a subset for future research. 

Sampling strategies need to incorporate both random and representative techniques. 

Excavations, Survey, and Analytical Sampling

Excavating a site is expensive and labor-intensive and it is a rare archaeological budget that allows for the complete excavation of an entire site. And, under most circumstances, it is considered ethical to leave a portion of a site or deposit unexcavated, assuming that improved research techniques will be invented in the future. In those cases, the archaeologist must design an excavation sampling strategy that will obtain enough information to allow reasonable interpretations of a site or area, while avoiding complete excavation.

An archaeological surface survey, where researchers walk across the surface of a site or region in search of sites, should also be conducted in a thoughtful manner. Although it may seem that you should plot and collect every artifact you identify, depending on your purpose it may be best to only use Global Positioning Systems (GPS) to plot selected artifacts and collect a sample of the others.

In the laboratory, you will be faced with mountains of data, and all will require further investigation to some degree. You may wish to limit the number of soil samples you send for analysis, preserving some for future work; you may want to select a sample of plain potsherds to be drawn, digitized, and/or curated, depending on your current budget, current purposes, and potential for future investigation. You may need to decide how many samples are send for radiocarbon dating, based on your budget and how many are needed to make sense of your site.

Types of Sampling

Scientific sampling needs to be carefully constructed. Consider how to obtain a thorough, objective sample that will represent the entire site or area. To do that, you need your sample to be both representative and random.

Representative sampling requires that you first assemble a description of all the pieces of the puzzle that you expect to examine, and then select a subset of each of those pieces to study. For example, if you plan to survey a particular valley, you might first plot out all the kinds of physical locations that occur in the valley (floodplain, upland, terrace, etc.) and then plan to survey the same acreage in each location type or the same percentage of area in each location type.

Random sampling is also an important component: you need to understand all parts of a site or deposit, not just the ones where you might find the most intact or the most artifact-rich areas. You could make a grid over the top of an archaeological site and then use a random number generator to decide which additional excavation units need to be added to remove some bias.

The Art and Science of Sampling

Sampling is arguably both an art and a science. You need to think through what you expect to find before you begin, and at the same time not let your expectations blind to what you haven't considered possible yet. Before, during, and after the sampling process, you need to constantly rethink and reconsider what your data are showing you, and test and retest to identify whether your return is valid and reliable. 

Selected Sources