Choose your own adventure: Quantitative methods in Signed Language Corpus Linguistics

Data analysis and cognitive sense-making are similar processes. Both can be bottom-up in that we create a hypothesis about the world on the basis of observations or top-down in that we evaluate observations about the world to a set of beliefs (Winter, 2022). A data analyst must consider whether their focus is on uncovering new patterns in their data, with minimal influence from pre-existing beliefs, or if they are using their data to validate or challenge their existing hypotheses. This talk outlines the basics of the former – confirmatory statistics, with a focus on linear models and takes a deep dive into the latter – exploratory statistics, looking at Factor Analysis, Clustering Analysis, and Classification and Regression Trees (CART). Examples of each type of analysis are detailed from research done on Nordic signed languages (Ferrara et al., 2022, 2023, Puupponen et al, forthcoming), highlighting the ground breaking work being done with these corpora.