A Snowball Sampling Approach for Studying Digital Minority Languages

This thesis discusses the methods used to assess the linguistic diversity of the internet. I critique the current literature on internet language diversity, arguing that existing methods-which aggregate textual data from many languages, to the exclusion of video and audio data-are unsuited to the study of minority languages. To address these shortcomings, I propose a snowball sampling approach for studying an individual language's online use. I provide a series of case studies, in which I apply this method to several so-called "low-density" languages to demonstrate its potential. Finally, I conclude by describing how future studies could be improved through partial automation of the data collection process.