Meta-analyzing Corpus Linguistic Research

Research synthesis and meta-analysis (RS/MA) comprise a set of procedures for conducting comprehensive and empirically-grounded reviews of previous research. In contrast to traditional literature reviews, RS/MA provide enhanced objectivity and systematicity. It is not surprising, therefore, that researchers have turned in recent years to RS/MA as a means to bring together and examine previous research in a number of individual subdomains. Corpus researchers, however, have applied RS/MA only sparsely and in a very small number of subdomains. Considering the broad range of questions addressed in corpus linguistics, we believe there is massive potential for meta-analysis as a means to systematically synthesize findings in the field of corpus linguistics. In order to enable such efforts, the chapter provides both a conceptual and procedural introduction to RS/MA, addressing all major steps including (a) defining the domain and searching for primary literature, (b) developing and implementing a coding scheme, (c) calculating and aggregating effect sizes (using the langtest.jp web app and the metafor package in R), and (d) interpreting results. The discussion considers each step with respect to the challenges unique to meta-analyzing research in corpus linguistics.We also propose a novel use of corpus linguistic techniques as a means to enhance the efficiency and perhaps accuracy of extracting synthetic data from primary studies.