Th is educational design research study contribute s to understanding s about the impact the complex ity in multilingual upper s e condary mathematics classroom s has on attempts to improv e student - to - student interaction s . The research study was carried out in a cyclic process in three cycles as an intervention to give all students the opportunit y to develop their reasoning and communication abilities in mathematics. In the intervention, students in a first year class of the social scien ce program in a n upper secondary school were followed during one semester while they worked in small groups with various problem - solving tasks in mathematics . The intervention included three sets of open problem - solving tasks and three support means in the form of question lists, problem - solving support lists , and various communicative roles. The student s’ interactions were recorded , and interviews were conducted in each cycle . Both the students’ interactions and perceptions about working with mathematics w ere then analy s ed . The research questions were: 1. How did the students’ interactions and perceptions change over time as a result of an intervention to improv e student - to - student interactions? 2. What contribute d to the changes in the students’ interactions an d perceptions when an intervention was provided to improv e student - to - student interactions? 3. What did working with an intervention on improving student - to - student interactions indicate about the complexity in the multilingual upper secondary mathematics classroom? The results showed possibilities for chang ing student - to - student interactions through improving how students listened to each other, built onto each other’s interactions, worked with mathematical questions , and determined what they focused on in gr oupwork. The intervention was developed over the cycles by adapting the tasks and support means according to the results from the previous cycles. Tasks were formulated more openly, groups were arranged so that everyone had the opportunity to speak, and ma thematical questions became an important part of student - to - student interaction s . In the study , two theoretical frameworks were used — Alrø and Skovsmose’s (2004) IC - model and Fuentes’ (2009) framework for analysing student communication. The frameworks we re coordinated and then used to justify choices in the intervention and to analy s e interaction s . Interviews were used to learn mo re about students’ perceptions. Complexity in multilingual upper secondary mathematics classrooms was connected to three perspe ctives : teaching and learning, research , and school development. In the study, complexity from the teaching and learning perspective was discussed concerning how students were given equitable learning opportunities, how language affected attempts to improv e the student - to - student interaction s , and how the students’ interactions and perceptions affected groupwork in mathematics . It proved to be important not to see multilingual students as a homogenous group with special needs simply because they are multili ngual, but rather , to be flexibl e in meet ing the needs of the students. From a r esearch perspective, complexity was discussed in relation to how educational design research and the two theoretical frameworks were used to promote improved student - to - student interaction s in mathematics . Finally, from a school development perspective, complexity concerning generalizations was discussed. This study point ed out the difficulty of making generalizations about promoting improved interaction s. Since many factors aff ected interaction s between students , instead it was considered important to be flexible and study the interaction s and students’ perceptions side - by - side to make changes.