Correcting sampling bias (looking for data science or statistical approach)?
Lets say I am looking at a sample of students and their grades achieved across 5 subjects. 90% of my sample is biased towards worst performing students and 10% of my sample is random.