Abstract As a usage-based approach to the study language, cognitive linguistics is theoretically well poised to apply quantitative methods to the analysis of corpus and experimental data. In this article, I review the historical circumstances that led to the quantitative turn in cognitive linguistics and give an overview of statistical models used by cognitive linguists, including chi-square test, Fisher test, Binomial test, t-test, ANOVA, correlation, regression, classification and regression trees, naïve discriminative learning, cluster analysis, multi-dimensional scaling, and correspondence analysis. I stress the essential role of introspection in the design and interpretation of linguistic studies, and assess the pros and cons of the quantitative turn. I also make a case for open access science and appropriate archiving of linguistic data.