Exhibit ES-5. Estimated Impacts on Key Outcomes for Early and Late Award Sites, by Grade
The complete RF study sample includes 248 schools from 18 sites (17 school districts and 1 state) located in 13 states. 125 schools are Reading First schools and 123 are non-Reading First schools. There are 8 late award sites, with 137 schools, and 10 early award sites, with 111 schools. Impact estimates are statistically adjusted (e.g., take each school's rating, site-specific funding cut-point, and other covariates into account) to reflect the regression discontinuity design of the study.
A two-tailed test of significance was used, and where applicable, statistically significant findings at the p..05 level are indicated by *.
EXHIBIT READS: For grade one, the impact of Reading First on the percent of students reading at or above grade level was 7.55 percentage points for late award sites, which was statistically significant (p..05). The corresponding impact for grade one in early award sites was .2.34 percentage points, which was not statistically significant.
Sources: RFIS SAT 10 administration in the spring of 2005 and 2006, as well as from state/district education agencies in those sites that already use the SAT 10 for their standardized testing (i.e., FL, KS, MD, OR); and RFIS Instructional Practice in Reading Inventory, spring 2005, fall 2005, and spring 2006.