Several limitations in the current study warrant mention. Data for the current study were gathered on one school from each program type. As such, school differences might reflect idiosyncratic building level differences, rather than the effect of a particular
program orientation. In addition, students in the study had been in existing programs and program implementation and fidelity were not experimentally controlled. Although random assignment was not possible, matching procedures at the building level, as well as statistical controls (i.e., covariance), were used to minimize potential student differences between schools that might have accounted for performance variability. In addition, no subject data was available on duration of enrollment in a specific program. Although Glenn (1996) found that number of years in a Montessori program was not associated with the demonstration of “Montessori qualities,” the potential impact of duration in any of the programs may be related to the efficacy of the program.