Comparing Corpora with Wordsmith Keywords

Abstract

In this article, I review some of the features available for language analysis in the KeyWords tool of WordSmith (Scott, 1996). In general, KeyWords has proved a reliable tool for comparing language samples, be them texts or corpora. Its results have helped identify not only differences across texts but also within texts. I present some arguments related to the use of chi-square in comparing word frequencies, and propose two techniques for extracting a representative subset of key words for analysis.