Abstract Background The cluster-randomized controlled trial DRINK was implemented in 2021 to evaluate the effect of interventions aimed at reducing sweetened beverages and increasing tap water consumption among children in French-speaking primary schools in Belgium. Multiple participation issues related to the implementation in the school setting were encountered. Our aim was to describe the individual characteristics of pupils at baseline and their participation in the data collection. Methods In 2021, out of a sample of 168 schools stratified by region and socioeconomic index (SEI), 46 were included. Children in grades 3 to 5 were asked to complete a 4-day diary to record all beverages consumed. Questionnaires were completed by children and parents, from which individual characteristics were summarised, such as the Family Affluence Scale (FAS; in three groups). Results School participation (overall: 29.0%) was higher in Wallonia (33.0%) than in Brussels (22.0%), and in schools with an intermediate SEI (33.0%). At baseline, a total of 3,631 children participated in the data collection, either through the 4-day diary, the child questionnaire, or the parent questionnaire; 2,013 completed all three instruments. Participating children were more likely to be girls, to have an intermediate FAS, to be born in Belgium, and to have two working parents. The proportion of children attending schools in the Brussels region was 27.9% of the total participating children and dropped to 11.5% for those who returned three valid questionnaires. The proportions of low FAS (14.4% to 11.8%), birth outside Belgium (11.8% to 8.7%), and low parents’ schooling (35.3% to 33.8%) were lower among children who completed the three instruments. Conclusions The low participation of Brussels schools, possibly exacerbated by the Covid-19 impact, may explain why disadvantaged children were less represented in the baseline sample. Such knowledge is valuable for interpreting the findings of such a trial. Key messages • Participation issues may have affected the representativeness of less advantaged children. • Documenting such disparities is key to understanding the context in which interventions were implemented.