Regional Comparison of Household-level Inequality

Jeffery Clark and Joshua Watts

Visualize the spatial distribution of household 'wealth' across the Southwest, from 750 to 1350 CE. Hovering over a bubble will give you more details about that time and place. Site locations are approximate, and the datasets shown were pulled from cyberSW in December 2025.

1. Bubbles show ranked, scaled values for assemblage richness indices in 100-year intervals. Small black dots show all the households in the database at the time this was posted, but the blue circles show the data for the current plot.

2. Plots show ranked, scaled values for assemblage richness in 100-year intervals. Values are counts of artifact types from features assigned (by the authors) to "households" that date to the interval divided by the count of undecorated ceramics from those features.

How was this map made?

This map shows household assemblage richness in 100-year intervals from several projects across AZ, NM, and CO. The focus is on sites in cyberSW where we have high-resolution intrasite data with detailed information on features and artifacts. For the current analysis 311 households were included. Several important assumptions from the authors need to be clarified before interpreting these results.

First, we assigned date ranges to each of the primary features at all of the sites (sometimes this info was provided by the original researchers). This was necessary to isolate features and artifact assemblages from relatively narrow temporal intervals. This process itself involved many assumptions, but in most cases relied on extending Ortman's Uniform Probability Density Analysis method to work on features and strata within sites (rather than for entire site ceramic assemblages, where that method is usually deployed). A detailed explanation of that workflow is available upon request. Features that fall within the temporal intervals---or more precisely where half or more of a feature's date range falls within that interval---shown for the map were included in the household assemblages described below.

Second, how should we think about the concept of a "household" in concrete terms when dealing with archaeological contexts? For the purposes of this analysis we relied on pre-contact architectural features. Small clusters of pit structures (ie., courtyard groups) or small compounds of masonry/adobe rooms were designated as households. Large complex pueblo villages are mostly not considered for this analysis unless sampling strategies led us to define features from different blocks as unique households.All of the primary features (domestic architecture, trash features, pits, etc.) that could be reasonably associated with a given household were assigned to that household. The map shows our best attempt at imposing the abstract concept of a household on settlements.

Third, household "richness" as defined here is not an index of wealth, precisely. Instead the concept was drawn from the ecological concept of species richness, usually simply defined as the count of species in a defined area. The data shown were queried from the cyberSW database, representing the total count of ceramic and material types found (aggregated at the household level). In this context, each ceramic type, each species of animal bone or marine shell, and each material of chipped stone all count equally with no bias towards goods that might better represent prestige or wealth items. These values were then scaled by the count of undecorated ceramics found at the household during the shown temporal interval. This was done to control, somewhat, for assemblage sample size differences. Other common indices can be toggled above (they are described below).

Ranked plots of those scores are show below for each of the six intervals. Interpreting these ranked plots can be tricky but the pattern from relatively linear ranking towards a more-concave shape suggests that a small number of households in the later intervals had greater access to more stuff. (see Watts and Ossa 2017, etc etc for thoughts on these sorts of patterns).

Scaled Richness: Values are counts of artifact types from features assigned (by the authors) to "households" that date to the interval divided by the count of undecorated ceramics from those features.

Dec Ceramics: Values are counts of decorated ceramics from features assigned (by the authors) to "households" that date to the interval divided by the count of undecorated ceramics from those features.

Shell: Values are counts of shell artifacts from features assigned (by the authors) to "households" that date to the interval divided by the count of undecorated ceramics from those features.

Artio Index: Values are a simple Artiodactyl Index (counts of artiodactyl bone divided by lagomorph bone) from features assigned (by the authors) to "households" that date to the interval.

Mineral: Values are counts of mineral artifacts (e.g., turquoise) from features assigned (by the authors) to "households" that date to the interval divided by the count of undecorated ceramics from those features.

Tonto Basin: Sites from the RPMS, RCD, and TCAP projects described in the Tonto Basin Example.

Shields Pueblo: Crow Canyon project in CO.

EMAP Cluster: Sites investigated by the Eastern Mimbres Archaeological project in NM.

Honey Bee Village: Desert Archaeology, Inc., project near Tucson, AZ.

SCARP Cluster: Sites investigated by the University of Arizona field school near Show Low, AZ.

San Pedro: Sites investigated by Archaeology Southwest in the Lower San Pedro River Valley.