Using fresh data from the British National Corpus , this book criticizes and expands Biber's seminal study (1988) of variation in spoken and written English. The author assesses the validity, stability and meaningfulness of Biber's results, critically re-evaluating the statistical methodology. In the process, many issues of importance in corpus-based research are touched upon, research design, corpus construction, representativeness, search algorithms, operationalization of linguistics concepts, and interpretation of statistical results.