Settlement Population Estimator

What Is the Settlement Population Estimator?

The Settlement Population Estimator works out an approximate population for an archaeological site using either a floor area coefficient or a house count multiplied by average household size. Archaeologists use these estimates to compare settlement sizes across a region, model demographic change over time, and support site significance assessments. According to Yale's eHRAF cross-cultural hypothesis archive on floor area and population, Raoul Naroll's 1962 cross-cultural study of dwelling floor area first proposed that population size could be predicted directly from built floor area, a relationship that became one of the most widely cited formulas in settlement archaeology.

Figure out which input data your site actually has recorded, total roofed floor area or a count of contemporaneous structures, before choosing a mode, since the two approaches draw on genuinely different field evidence. Neither method produces an exact headcount; both are working estimates built on ethnographic analogy, and treating them as anything more precise than that risks overstating how much the underlying numbers can actually support.

Floor Area Coefficients: Naroll's Constant and Its Restudy

Naroll's original 1962 formula set the coefficient at 10 square metres of floor area per person, based on ethnographic data from 18 societies. This became known as "Naroll's Constant" and has been applied across a huge range of archaeological contexts since. Barton McCaul Brown's 1987 restudy, published in American Antiquity, revisited the relationship with a larger dataset and proposed a considerably lower working coefficient of 6 square metres per person.

That is a 67% difference in the population estimate for the exact same floor area, driven entirely by which published coefficient was applied. Work out both figures side by side rather than reporting a single number without stating which coefficient produced it, since a reader has no way to judge the estimate's reliability without that context.

House Count and Household Size: An Alternative Approach

Where contemporaneous structures or hearths can be reliably counted, multiplying that count by an average household size offers a different route to the same estimate. Look into regional ethnographic or ethnohistoric analogy to set a defensible household size figure rather than defaulting to a round number, since average household size varies considerably across cultures, periods, and settlement types, and using hearth counts as a household proxy carries its own assumptions about one hearth per household that may not hold everywhere. Carry out this cross-check whenever your site's floor area figures are uncertain or incomplete, since a reliable structure count can stand in for the area-based method entirely in those cases.

Come back to both methods when they are both available for the same site, since agreement between an area-based estimate and a household-based estimate provides real corroboration, while a large disagreement between them is a signal worth investigating rather than ignoring. This same cross-checking logic underlies related settlement-pattern research; published work connecting artifact density to population density in settlement studies uses multiple independent proxies for the same reason, since no single measure of past population size is reliable enough to stand entirely on its own.

Why No Single Coefficient Is Universally Correct

Set out the reasoning behind whichever coefficient you choose, since the underlying ethnographic samples both Naroll and Brown worked from cover a wide range of societies with genuinely different housing customs, climates, and household structures. Brown's own restudy findings point to settlement size and density, rather than climate or the timing of contact with outside societies, as the factor that best explains why the floor-area-to-population relationship varies so much between different ethnographic samples. Some researchers have proposed settlement-specific or region-specific adjustments rather than a single universal figure, and allometric models modified for particular settlement types, such as hunter-gatherer camps, exist specifically because a single fixed coefficient does not transfer well across every kind of site.

Given that coefficient choice alone can shift a population estimate by well over half, treat any single reported figure with appropriate caution until you know which method and coefficient produced it. Once you have a working population estimate, the Artifact Density Calculator can help relate artifact recovery rates back to that estimated population, and the Age-Depth Model Calculator can anchor a population estimate to a specific occupation phase if the site has multiple stratigraphic levels.

Accuracy and Limitations

The arithmetic in this calculator, dividing area by a coefficient or multiplying house count by household size, is exact given accurate inputs. That said, both methods rest on ethnographic analogy rather than direct measurement of the actual past population, and neither this calculator nor the underlying published coefficients from the Naroll and Brown studies can verify whether a specific site's culture matches the assumptions built into either sample. Floor area figures themselves also depend on accurately distinguishing genuinely contemporaneous structures from those built and abandoned at different times within a site's occupation, a judgment call this calculator cannot make on your behalf. Turn out a range rather than a single figure wherever your site's occupation history is not fully resolved, since including structures that were never actually occupied at the same time inflates any population estimate regardless of which coefficient is applied.

The Most Common Settlement Population Mistake

The mistake I see most often is reporting a single population figure as though it were a precise measurement, when in practice it depends entirely on which coefficient or household size assumption was applied. With that in mind, always state the specific coefficient or household size figure behind any population estimate you report, and consider presenting a range rather than a single number whenever more than one defensible coefficient exists for your context. On top of that, treat close agreement between an area-based and a household-based estimate as a meaningful check worth seeking out, since two independent methods converging on a similar figure is considerably more persuasive than either method reported alone, in line with how the original Naroll-Brown restudy literature itself was built by testing one formula's assumptions against an expanded independent dataset rather than accepting the first published coefficient uncritically.