The Study Sample

The study sample was constructed, and the study was conducted, over two years. We began in late 2003 by identifying as many potentially eligible AC programs as possible. Among those states not known to have selective admissions criteria for their AC programs (12 total)⁴ we compiled a list of 165 programs, from which we drew a random sample of 63, stratified to ensure diversity in terms of geography (state) and types of programs within states. For the 2004–2005 school year, we recruited schools that had hired teachers from a purposive subsample of the 63 sampled programs.⁵ For the 2005–2006 school year, we sought more teachers from the same programs and also directly approached new districts in some of the same states that hired large numbers of AC teachers (for example, because they operated their own program). Schools could be included in the study only if they had at least one eligible AC and one eligible TC teacher in the same grade, in kindergarten through grade 5. To be eligible, teachers (1) had to be relative novices (three or fewer years of teaching experience prior to 2004–2005, five or fewer years prior to 2005–2006); (2) had to teach in regular classrooms (for example, not in special education classrooms); and (3) had to deliver both reading and math instruction to all their own students. The final study sample included 87 AC teachers and 87 TC teachers (some of whom participated in the study both years) from 63 schools in 20 districts and 7 states, as shown in Exhibit 1. Fourteen of the 20 districts were in urban areas, and 4 were on the fringe of one. Although we identified and sampled from a large number of less selective AC programs operating in 2003–2004, the programs and teachers that were included in the study sample were not necessarily representative of all AC programs operating at the time.