Twenty-eight school districts plus one state Reading First program that met the preceding criteria were identified. Sixteen districts plus the state program were chosen from this pool to participate in the regression discontinuity design; the final selection reflected wide variation in district characteristics and provided enough schools to meet the study's sample size requirements. One other school district agreed to randomly assign some of its eligible schools to Reading First or a control group. The 17 school districts and one state Reading First program are referred to as study sites. The regression discontinuity sites provide 238 schools for the analysis and the randomized experimental site provides 10 schools. Half of these schools at each site are Reading First schools and half are non-Reading First schools; the study schools comprise some, not all, of the RF schools in study sites.
Exhibit ES.1 compares background characteristics of Reading First schools in the study sample to those of all Reading First schools in the 18 study sites, all Reading First schools in the 13 study states, and all Reading First schools in the nation. Visual inspection of the data displayed in this exhibit suggests that, overall, the present sample is similar to the other three groups of Reading First schools. Almost all are eligible for Title I support, they enroll high percentages of students eligible for free or reduced price lunch, and their past third grade reading scores are near their state averages for Reading First schools. The RFIS sample, on average, has proportionally lower percentages of Hispanic students and higher percentages of Black students than Reading First schools in the study states or in the nation; at the same time, RFIS sample schools, on average, have a lower percentage of Black students and a higher percentage of White students than Reading First schools in study districts. A greater proportion of Reading First schools in the study sample are in large or mid-size cities, and not other locales, than are Reading First schools in the study states or in the nation. Also, the sizes of Reading First schools in the study sample, on average, are somewhat smaller than those in the three other groups. Further, these data cannot provide conclusive evidence that the study sample fully represents the experience of the entire national Reading First program, as the study sample might differ from the Reading First population in other ways that were not observed.