Intelligibility testing in Vanuatu in 2015 investigated how well speakers of three closely related varieties could understand each other. But the research also generated questions about (1) anomalous results; (2) which linguistic factors influenced comprehension; and (3) the relative intelligibility of varieties. This paper interrogates these questions and finds that, first, while most anomalous results are difficult to account for, others are easily explained; the insights gained will help to refine the design of future intelligibility tests. Second, some variables appear to be more important than others in terms of the degree to which they impede intelligibility. Third, test participants' higher comprehension of a relatively distantly related variety over a more closely related variety leads to explanations that draw on both linguistic and social factors. The insights gained in this study contribute to existing research on the same topic in European languages and establish a starting point for similar research on Pacific languages.