Pedagogical translanguaging research often shows positive effects on student engagement and learning (e.g., García & Kleyn, 2016; Paulsrud, Rosén, Straszer & Wedin, 2018; Paulsrud, Tian & Toth, 2021). Given that each classroom is a unique sociocultural context (Byrnes, 2020), often exhibiting complex multi-causality (Baker & Wright, 2021), most of this research is qualitative in nature. In this paper, we present mixed-method research (Baker & Wright, 2021) from a quasi-experimental, longitudinal intervention study in a multilingual school in Sweden. Quantitative data were used to trace student learning of English vocabulary in a year-9 class (16 students) over twelve weeks: a pre-test, immediate post-test and delayed post-test. With the purpose of gaining an in-depth understanding of student learning, qualitative data (questionnaires, interviews and vocabulary learning materials used in class) were collected from three of the students. All three were successively trilingual users of a migrant L1 and Swedish and English as additional languages (L2s). Their first exposure to Swedish was on their arrival in Sweden at the age of 6, 7 and 8 respectively. Results on the pre-tests, post-tests and delayed post-tests show that students (N=16) in the class as a whole learnt significantly more new English vocabulary when they were provided with translation equivalents in Swedish than when provided with definitions or explanations in English. This presentation focuses on the mixed-method data from the three focal students, addressing the following questions: a) how many new English words did the three focal students learn when provided with English explanations/definitions, translation equivalents in Swedish, and translation equivalents in Swedish and their L1s (Turkish, Pashto, Danish and Arabic), b) to what extent did they use their different languages in class, and c) what were their attitudes, orientations and beliefs towards the use of these languages when learning English in the classroom?