Using unfamiliar research methods requires confidence and open-mindedness. I am not a confident mathematician and could previously have been called numerophobic. But a research question as to why the British Government was losing more immigration appeal cases in court forced me to engage with quantitative methods. I sought to explain the fact that 25 of 78 immigration appeal cases were lost by the crown from 1970 to 1994, but 90 of 174 cases were lost in just 17 succeeding years. I suspected the explanation lay in the increased indeterminacy of parliamentary legislation. My question, and theory, demanded I use computer-assisted natural language processing of legal documents, and that I then analyze the results with multivariate logit regression analyses, including models for random effects. These methods would, in the past, have left me bewildered. But my journey of methodological discovery was, in the end, rewarding and demonstrated the centrality of confidence to teaching and learning.