Design
The pilot trial is designed as an individual randomised study in three sites, conducted from August 2021 to June 2022. Participants randomised to treatment condition will receive language training with Multilingual Support. Participants randomised to control condition will receive language training without Multilingual Support. The number of assistants in the classroom will be equal in treatment and control condition, however, only the assistants in treatment condition will master the participants' primary language and follow the guidelines set out in Multilingual Support manual.
Participants
To be eligible for participation, persons must be at early stages of their language learning (i.e. below A2 levels), and eight or less years of education in primary and lower secondary school. In order to limit the number of languages provided by the Multilingual assistants, and hence reduce the complexity of the study, we pilot the intervention and trial within the one or two largest language groups at each site (i.e. Arabic and Kinyabwisha/Swahili in pilot 1; Coastal town; Arabic and Dari in pilot 2; Middle-town; and Arabic in pilot 3; Suburb-town. Persons who are illiterate are excluded from the study.
Recruitment
Staff at the local Adult Vocational centre identify eligible participants for language training with Multilingual Support from June to August 2021. When an eligible person has been identified, the staff provides the person with information about the study. This information will be provided in writing in Norwegian and in the person's primary language. In addition, the information will be given orally, in eligible participants primary language, making it possible for eligible participants to ask questions about the project. If the person agrees to participate in the study, the staff deliver the participants email address and information about their primary language to the researchers. Only participants that provide written content will be included in the study.
Randomisation procedures
When receiving the participant's email address and information on primary language, the researchers send a request to the participant, inviting him/her to respond to a baseline survey. When responding to the baseline survey, participants are not aware of their allocation to treatment or control condition. The survey client (Nettskjema.no) includes a function to randomise participants. In this study we randomise participants into two groups, with 2/3 probability of being allocated to the treatment condition. Randomisation is conducted as independent randomisation (no stratification). After the participant has submitted the baseline survey, the researchers can obtain information on treatment allocation, and inform the staff at the local Adult Vocational centre about the randomisation result, whereafter the Centre invites the participant to the Multilingual Support or the control condition.
