Telephone Sample Design and
Telephone interviews were conducted among parents with at least one child in a Washington, D.C. charter or public school. For the first wave, interviews were conducted between September 12 and December 11, 2001. All interviews were conducted by the Center for Survey Research at the State University of New York at Stony Brook. As a quality-control measure, up to fifteen callbacks were made per number and an attempt was made to convert all initial refusals. Almost 52 percent of all interviews were validated on a subsequent call after the interview had been completed.
Parents were drawn from two distinct samples: a random sample of parents with children in charter and public schools, created through “random digit dialing” (RDD) and a sample of parents randomly chosen from a list of charter-school parents provided by D.C. charter schools. Charter parents were thus oversampled by design due to their relatively low incidence in the population (about 50 percent of the parents in our sample have children in charter schools. But by chance we would expect to find only about 20 percent, which we thought was too small a proportion for the key variable of interest in this analysis). All analyses reported in the book adjust for this design through poststratification weighting.
A list-assisted method of RDD was used to obtain phone numbers in the main state sample. Numbers were purchased from Genesys. Under the list-assisted sampling method, random samples of telephone numbers are selected from blocks of one hundred telephone numbers that are known to contain at least one listed residential telephone number. These blocks with at least one residential telephone number are referred to as “1-plus” working blocks. According to Survey Sampling Inc., roughly 40 percent of telephone numbers in 1-plus working blocks are residences, although