An integrative strategy for the definition of behaviorally meaningful archaeological units

Abstract

(pp. 30-31)

In the sections above, we have presented the theory and reviewed the methods presently available for the assessment of homogeneity and spatial patterning within the constraints of gridded data. While our approach has been largely critical, we have attempted to present an integrated strategy consisting of the best of the various alternatives. In terms of a comparative assessment of the indices of clustering, Morisita's index vs. dimensional analysis of variance, the examples included in this paper would indicate an agreement with Brase and Scarry (1976) in that both indices are equally powerful in the approximation and recognition of spatial patterns. However, in the larger work from which these examples were drawn (Newell and Wiersum 1977), the use of Morisita's index led to the recognition of significant clustering at an earlier stage in the block size incrementation than did dimensional analysis of variance in a number of instances. This apparent greater sensitivity clearly lends support to Price's (n.d.a) claim that Morisita's index is the better of the two technigues. Nevertheless, inherent problems in the application of this and similar inductive strategies still remain. Many of these have been recognized and corrected by other authors (Pielou 1969; Whallon l973a; Price n.d.a; Schiffer 1974; Riley 1974; Clay 1975), and in some cases we have been able to indicate corrections and/or suggest alternatives. Other problems, such as the accommodation of features and the relation of artifacts to features, still remain unsolved.