Skip to main content


Examining the statistical methods of socio-linguistics

Research Achievements

Examining the statistical methods of socio-linguistics

IGERT Trainee Kyle Gorman (Linguistics) published a paper [Gorman, 2010] arguing for improvements in the statistical methods used in socio-linguistics. He writes: "A class of generalized linear models known as hierarchical (or random-effects) models are, I argue, a way to deal with the well-known fact that while speech communities share similar internal and external constraints on variation...subject variation may be sufficient to swamp those trends. Addressing subject-level variation in a principled fashion allows researchers to assert that the results of a study are likely to generalize to another sample from the same community...I demonstrate that hierarchical models should be used when samples include multiple observations from the same subject. I [also] examine the sociolinguistic stratification of linguistic negation in Philadelphia, demonstrating that the use of hierarchical models has positive empirical consequences."