This article discusses some methodological issues that arose when analysing data collected in a pilot study of the SOFRA project. We aimed at piloting a semi-structured interview protocol designed to collect qualitative data with nine Syrian asylum seekers and refugees studying French at university, using an interview schedule that targeted, among other things, information about learners’ interaction opportunities and attitudes about their new environment and learning experiences. Analysing the manners in which the interviewer asked the questions and coped with comprehension difficulties, as well as the way the interviewees responded to the questions, led to the identification of a number of problems that are partly related to question type and wording. The article concludes with a reflection on how to elicit relevant answers during a semi-structured interview with migrant learners.