Quantitative Designs and Statistical Techniques

Abstract

This chapter will provide an overview of quantitative designs in corpus linguistics. Section 3.1 introduces the kinds of research design questions nearly every quantitative corpus linguistic study must involve at the planning stage: (1) what corpus linguistic statistic(s) to use and (2) how to evaluate them to inform the conclusions of a study. Section 3.2 is devoted to statistical methods that are, in a sense, ‘specific’ to corpus linguistic applications beginning with different kinds of frequencies, entropies, and keyness values, before turning to co-occurrence phenomena and its association measures as well as dispersion measures. Section 3.3 is concerned with ‘general’ statistical methods. It begins with a short mention of monofactorial statistics (e.g., chi-squared tests or correlation coefficients) before turning to multifactorial statistics, in particular fixed- and mixed-effects regression models, their extensions and combination, and increasingly prominent tools such as structural equation modeling or tree-based methods. This is followed by a brief discussion of exploratory methods such as multidimensional analysis (MDA) and approaches like cluster or correspondence analysis. I conclude with a few words of caution and desiderata regarding what practitioners need to bear in mind as they gravitate to the more complex methods our field often requires.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 81 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.