Goldvarb has the advantage of having been designed for use with linguistic data and there are books that provide a step-by-step account of how to conduct a GV analysis of a sociolinguistic variable (see Tagliamonte’s Analysing Sociolinguistic Variation). One drawback is that it can only handle binary or categorical variables and not continuous features (e.g. formant measures). Goldvarb also cannot produce Mixed Effects Models, this can mean that it sometimes overestimates the effect size of social factors like gender and age as these properties are nested within individual speakers.

For a link to the Goldvarb website, please click HERE