I need some help with formulating a methodology for my current dissertation. I am investigating the impact of ability setting on students. I am focussing on a single year group of approximately 70 students for this, and it is my aim to investigate how moving up or down in terms of their ability group may affect subsequent scoring on tests. I have lots of data for this group of students - it consists of a number of unit test scores throughout the course of the school year. When a student has been found to perform consistently well or poorly they are usually moved up or down a set and I ultimately aim to explore the impact of this on their levels of motivation and subsequent attainment.

The students are in 5 sets. The top 3 study the higher scheme of work and the bottom 2 study lower level work. Each stream has its own separate test at the end of each unit of work.

My initial inclination with this is to plot individual scores for consecutive unit tests on a line graph and for each class also plot a regression line. Those who demonstrate a a sharp rise or fall in their academic achievement following set movement will be targeted.

I am eager to hear a statisticians viewpoint on the presentation of such data and subsequent ananlysis. What would you do in my position? Would you simply use Excel to plot individual performance in subsequent tests or otherwise?

Any ideas or feedback would be gratefully received,