By Mike Smith, with comments from the peanut gallery by Angela Huster
After a rather involved set of phone conversations and emails among various project members trying to establish the frequency of green obsidian at the site during each phase, we realized that half our problem was the we were working with different samples. We had defined the Domestic Context Sample of lots strongly associated with dated houses several years ago, but any analyses that wanted to work with a larger sample were a free-for-all.
Based on Mike's previous projects in Morelos, we defined five samples based on their value for the analysis of domestic artifacts and conditions. These run from the Domestic Context Sample, (now called Domestic Sample 1, or DS-1), which consists only of well-dated midden deposits, to DS-5, which is the entire sample of all excavated contexts at the site. In addition there are other samples of lots that make sense for particular analyses or materials. Most of these samples are nested; e.g., all other samples include DS-1, and DS-5 includes all other samples. Samples DS-3 and DS-4 intersect in a non-nesting fashion, however.
DS-1 (the Domestic Context Sample). 178 lots.
This sample consists of well-dated midden deposits associated with houses. It can be subdivided into domestic components; that is, deposits from a single phase in a single unit. It was designed to provide a robust sample of materials from contexts with abundant artifacts for optimal quantification. This sample is used for:
Household comparisons of ceramic type frequencies.
· Ceramic type frequencies for the Aguas Celestiales chapter
It is also the source of sub-samples for particular technical analyses, including:
Ceramic attribute recording
· Obsidian source samples
· Ceramic petrographic samples
· Angela’s NAA samples
DS-2 (the Extended Domestic Context Sample). 340 lots.
This sample extends the domestic context sample to include other lots dating to the same phases at individual units. Units without representation in the domestic context sample are not included in DS-2. The advantage of a larger sample is offset by the inclusion of contexts such as fill and colluvial overburden whose association with the occupation of a house is less secure. This is used for:
· Interhousehold comparisons of rarer items in Angela’s dissertation analyses – ground stone, copper, jewelry, whorls, and molds
DS-3 (the all-Phased sample) 1,146 lots.
This sample consists of all lots phased to a particular period. It includes the transitional or uncertain phases (3, 5, and 44*) and phase 1 (pre-Postclassic). This is a much larger sample than DS-1 or DS-2, and its value lies in the fact that it includes the maximal number of lots that can be phased. Many of the lots have not had their ceramics classified; they are phased through stratigraphic position or associations with lots dated from their classified ceramics and/or radiocarbon dates. Its disdvantage is the inclusion of many lots whose direct association with the house occupation of each unit is more tenuous (fill, overburden, etc.). Its primary use is for comparing frequencies of rare artifacts. It is used for:
· Rare artifact types by phase for the general project
DS-4 (the Classified Sample) 664 lots.
This is the sample of lots whose ceramics have been classified. It has two related disadvantages compared to sample DS-2: many of the lots have not been phased; and many of the lots have only a few sherds. It is not used in any analyses.
DS-5 (all Excavated Lots). 1,668 lots.
This is the total number of lots that were excavated. It includes many tiny lots from Alex’s soil sampling, lots from architectural excavations, and other small lots whose value for artifact analysis is minimal. It’s primary use is to generate inventories of all excavated artifacts of a given type, irrespective of phasing or quality of context. This is useful for descriptive purposes (i.e., we want to describe all of the figurines, not just the ones that fall into a more restricted sample), but not for making comparisons among units or parts of the site.
The following diagram shows the relationship among these samples.
If anyone needs to know which samples their data fall into, please contact Angela.