Broken K Revisited: A Short Discussion of Factor Analysis

Publisher Website
Google Scholar
Abstract
The use of factor analysis in archaeology has increased considerably in recent years as archaeologists have recognized the advantages of using a statistical technique that can elicit otherwise unrecognizable patterns in a set of data. But there are several questions relating to scaling transformations and reliability assessments that must be answered within the context of specific research designs before reliable results can be obtained. A re-analysis of pottery type distributions at Broken K Pueblo is used to illustrate these questions. The factor solutions produced by the re-analysis exhibit relatively low reliability levels, indicating that the ratio of specific and error variance to common variance in the data is somewhat high.