One of the difficulties in making cross-program comparisons
is that programs can vary on many dimensions that are hard to
control and may be related to program quality (Frede, 1998).
Many characteristics influence a program’s capacity to provide
children with frequent optimal interactions with adults, other
children, and the physical environment. If experimental studies
were conducted on a large scale with large numbers of classrooms,
variations in the other program characteristics could be
expected to average out. But most curriculum comparison studies
have been quite small.