The problem with finding this sort of data is that it is self unselecting; i.e. it suffers from survivour bias. So you would need to asses base level corruption

The problem with finding this sort of data is that it is self unselecting; i.e. it suffers from survivour bias. So you would need to asses base level corruption in organisations over their histories collectively and severally. Look for the quiet spots and compare the relative successfulness of under capitalised students, i.e. do students from poorer backgrounds do better during years where there is less scandal? You might also want to compare to low level scandal to form a relative comparison; how many research papers are non reproducible, had to be withdrawn, etc. Then you need to figure if there is a causal link; this is left as an exercise for the reader.