A Guide to Build (ING) Generalised Linear Mixed Model Trees in Canadian Maritime English: Part 1, Social Factors

ABSTRACT This two‐part guide/research report introduces the Generalised Linear Mixed Model (GLMM) tree analysis technique to variationist sociolinguistics using (ING) variation in Canadian Maritime English (CME) as a test case. GLMM tree analysis combines the advantages of tree‐based recursive partitioning with the ability to include random effects in statistical modelling. In this, Part 1, the GLMM tree technique reveals a more nuanced pattern for age, gender, and education effects on variation between [ɪn] and standard [ɪŋ] for (ING) than simple mixed‐effect regression modelling alone. Linguistic constraints on (ING) variation in this data, as well as the GLMM tree analysis's merits for testing (multi‐)collinear predictors are explored in Part 2.