I have a load of data, showing me the predicted grades then actual gcse grades of students at school. I have these grades for english lit, english lang, maths and dt. I have marked down the number of people with each of the grades (a*-u), predicted and actual grades. I have put these onto a graph. That bit is all fine.
Now my problem. On one graph, I would like to put the differences between actual and predicted grades for each of the subjects, but on one graph. so i counted the differences for each grade (A*-U), for each separate subject. this is fine BUT because english and maths have 83 students and DT only has 12 students, then the differences on my graph looks funny as obviously there are a lot less dt students. so I thought, what if I somehow made the number of differences into percent, then my graph would be more accurate? trouble is, i dont know how to do this and need some help mats isnt a strong point of mine and im not afraind to admit it, but im trying and would really appreciate anyones help-and hope i make sense! thank you in advance