Our results show that teacher participation in AMSP is beginning to have independent, positive effects on student learning. After controlling for other variables that are known from the literature to affect student scores, our measures of AMSP participation suggest that more hours of training with the AMSP partnership program is leading to middle and high school math and science score gains at the .001 level of significance. At this point, the gains for elementary schools are at lower levels of significance. When estimating the AMSP participation by the share of teachers participating, we again find gains in middle and high school.