ABSTRACT
Settlement pattern research often examines relationships between archaeological distributions and natural and social environments at the regional level. Environmental settings (soils, access to water) that influence site selection and settlement are examined here. Early investigations simply tabulated archaeological distributions by classes of environmental types, where high class frequencies implied their importance. Statistical analyses that examined these distributions against environmental classes through goodness-of-fit tests soon followed. Later, two-sample approaches compared environmental characteristics exhibited at sample sites against random samples of locations from the background environment. Significant differences implied locational tendencies for specific environmental properties. Related analyses compared environmental properties between two archaeological site types. With the advent of GIS, one-sample testing forms became possible because encoded background “populations” no longer required sampling. Randomization tests could also be performed. Archaeological location modelling frequently built on environmental knowledge gained from such analyses. Some multivariate statistical approaches yielded linear functions that gave insights into relationships between archaeological distributions and environment through the signs and sizes of associated coefficients. Moreover, minor principal components derived from a suite of environmental variables have defined the most restrictive dimensions of variation in settlement locations by zeroing-in on more constant relationships with environment.
