The project reported here is conducting repeated randomized control trials of an approach to high school geometry that utilizes Dynamic Geometry (DG) software to supplement ordinary instructional practices. It compares effects of that intervention with standard instruction that does not make use of computer drawing/exploration tools. The basic hypothesis of the study is that use of DG software to engage students in constructing mathematical ideas through active investigations results in better geometry learning for most students. The study tests that hypothesis by assessing student learning in classrooms randomly assigned to treatment and control groups. The project is currently in its second year, and has just completed its first implementation of the DG approach, related data collection, and some initial data analysis. HLM models showed that the treatment group significantly outperformed the control group in geometry achievement. While the effect of the DG treatment is of moderate size for all participating students the largest effect size occurs with students in Regular Geometry classes.