I'm very new to data analysis. I'm trying to find the causal effect of seating row and laptop use on grades at a specific university. I have data from 15 introductory economics lecture sessions delivered by 5 different lecturers (who had three weekly sessions each). Each student was randomly assigned at the beginning of the course to one of these lecturers. The only assessment to gauge student performance was carried out at the end of the course. In total, the dataset has information on 1,223 students. Some students attended all three sessions, and some did not, leading to 2,647 individual (student-session) observations.
So I have multiple observations per student (up to 3) the grades for each observation are obviously the same, but seating row and laptop use may differ per recorded observation.
My question is whether I should take an average seating row and whether or not the student used a laptop at all (y/n) to run a regression and just have 1 observation per student, or use the entire data set for the regression.