ABSTRACT
This chapter discusses how to summarise, graph and test data for significance when focusing on one independent variable at a time. This is the simplest way to begin people data analysis, but most sociolinguistic studies are of course going to be examining the effects of more than one factor on the dependent variable. Multivariate analysis is a statistical test that can tease apart independent variables when they are tied up with one another. Multivariate analysis is also known as regression: specifically, logistic regression in the case of a categorical dependent variable and linear regression in the case of a continuous one. The chapter focuses on estimated probabilities that are based on multivariate analysis: they represent the rate, at which the different age groups under study would use, body pronouns with or without postnominal modifiers, all other linguistic and social factors being equal. Sociolinguists have traditionally performed multivariate analysis using software called Goldvarb and Varbrul.
