‘At Risk’ Constructs as Research Object: Research Design and Methods

The chapter outlines the methods and design of the research underpinning the book. The analysis uses The Times archive at the Corpus Approaches to Social Sciences (CASS) research centre at Lancaster University. The chapter discusses data quality and Optical Character Recognition (OCR) issues when studying historical newspapers, the composition of the corpus is divided into 23 subcorpora to keep the large corpus manageable by CQPweb, decisions about collocation windows and measures such as log likelihood (LL) and log ratio (LR). Finally, it outlines the affinity of ‘at’ and ‘risk’ and justifies examining the constructs at the risk, at risk and at-risk separately.