Many of our studies make use of comparisons where the implementation of a program has already begun or where the research sponsor has no control over planning for conditions such as a phased roll-out. Here we employ standard econometric methods such as difference-in-differences and quasi-experimental (QE) designs from the education tradition. Where appropriate, these studies estimate moderator effects. Our studies may use student, teacher, and/or school-level data. For an example see our report of a QE in Atlanta, Georgia.
