When supervisors pick the research problem