The graph is pretty interesting.
Looking at the plots for math, I see that the non-GT program kids cluster along the line, but for the kids in GT-program, quite a few have developed a .1 to .2 SD separation from the line and the other kids, which suggests that for them - the GT program has had quite an effect.
Another explanation for the continuity between the two groups and which explains the high-achievers is that some kids are already being pushed to their limit and this sets the curve in both - and these kids will not accelerate their learning because they are already saturated. For kids who are not at the edge of their ability to learn, they will "bump up" after being in a tougher program - hence the outliers.
It should be easy to set up a "gate" to cut out the kids who are saturated to make the one who benefit stand out. You could then do a regression test on some other aspects to then come up with a way to find those who will benefit the most.
Last edited by Austin; 08/01/11 01:01 PM.
