Regression analysis in translation studies

1.1 Types of regressions and variables One of the most remarkable current trends in theoretical and applied linguistics is the evolution of the field towards more empirically rigorous and quantitative methods. In theoretical linguistics, after a long reign of generative approaches to grammar and their largely intuitive grammaticality judgments, there is now a lot of interest in, and work on, probabilistic theories of language acquisition, representation, and processing, and such approaches rely more and more on experimental and observational data that are analyzed with statistical tools. In a similar vein, many areas of applied linguistics such as second language acquisition also have turned to quantitative tools, and translation studies are no exception. Given that this is only a relatively recent development and that different kinds of data are only slowly becoming available (e.g., corpora on lesser-studied languages and/or parallel and aligned corpora), the move towards more quantitative methods is still in progress. Practitioners are constantly learning about, and developing, new methods and areas of application for existing methods. One particularly flexible and widespread family of methods is that of regression analysis. This method involves analyzing the degree to which a dependent variable is correlated to one or more predictors, where we use predictors as a cover term for both individual independent variables and their n-way interactions. The dependent and independent variables in a regression can be of various levels of measurement: