This study examines the effects of social and political changes that were occurring during the eighteenth century in Scotland on the use of written Scots, focussing in particular upon authors who were known to have been for or against the Union of the Parliaments in 1707. In order to capture a holistic representation of the levels of Scots in writing, I explore the proportion of Scots lexemes, compared with their corresponding English lexemes, in a purpose-built corpus containing a range of eighteenth-century texts. This corpus contains both texts that were produced by a general cross-section of Scottish society, and a number of politically-active individuals. I take a quantitative sociolinguistic approach to historical data by utilising statistical techniques that examine linguistic variation in a data-driven manner. This enables a more detailed and empirical exploration of Scots in the eighteenth century, which until now has been largely examined on a descriptive basis only. Using a number of statistical tools that are well suited to historical analyses, such as Variability-based Neighbour Clustering (Gries & Hilpert, 2008), conditional inference trees (Hothorn et al., 2006) and random forests (Breiman, 2001), I have been able to reconstruct both the general patterning of the Scots language over time and the extralinguistic factors encouraging or suppressing its presence in writing. In particular, I compare the use of Scots between the general literate population and political individuals active during this time period. I also explore the effect of the latter’s political sympathies on their language choices, and uncover several new and interesting effects conditioning the levels of Scots in their writings. I tie these results to the underlying political change and discontent characterising Scotland during this time, as well as the general linguistic changes taking place across the eighteenth century as a result of broader processes of change over time.