No significant differences in child cognitive outcomes were
consistently found among the different models or between the
model programs and typical Head Start. There was considerable
variation within different classrooms using the same model as
well as between classrooms using different models. However,
the comparisons were fraught with possible confounds. Fidelity
of implementation was highly variable; the same supervisors
were used for model and regular Head Start classrooms, making
treatment diffusion likely; and, as stated above, outcome measures
appropriate for one curriculum are unlikely to be appropriate
for another.