Conducting a survey with refugee populations involves particular challenges. Among these challenges, sampling, implementation and survey translation are crucial. This becomes even more obvious when trying to access hard-to-survey subgroups of a larger refugee population. This paper demonstrates possible ways of addressing sampling, implementation and translation problems and some of the associated pitfalls based on the example of a quantitative survey intended for refugee parents in the state of Lower Saxony/Germany. First, we argue that based on the need to collect representative data, adjusting the target population may be one way to respond to the lack of a sampling frame for the original target population. Second, we reason that under unforeseen circumstances, such as an ongoing pandemic, ad-hoc changes may be needed in implementation strategies, and we shed light on some disadvantages of self-administered web-based surveys in refugee studies. Third, we claim that surveys involving linguistically and culturally diverse refugee populations benefit to large extent from using a modified variant of the team translation approach (TRAPD). In our conclusion, we critically reflect on adaptable strategies for ensuring well-defined samples, a reasoned implementation and translation practice that meet the challenges of representing a distinct refugee population’s diversity.