As part of the national Head Start Planned Variation Study, this study used a relatively small sample in an intensive evaluation of program implementation in one field community using the Tucson Early Education Model (TEEM). A modified Solomon four-group research design formed the organization framework. Evaluation of six TEEM classrooms and two locally implemented Comparison classrooms included a child data battery, classroom observation, situational tasks, Director's ratings, and demographic information. The child data battery, which emphasized both the cognitive and socio-affective domains, was given in the fall of 1971 to approximately half the TEEM children and to one of the Comparison classes. Spring testing included all groups. Analysis indicated that Comparison groups were not optimally matched to TEEM children; Comparison children were slightly older and performed better on pre-test. The difference in levels of inappropriate behavior during teacher absence between TEEM and Comparison classrooms was significant. Analysis of classroom observation indicated that TEEM classrooms had more child-initiated learning sequences combined with lower teacher initiated learning sequences. Due to limited sample size in this study, wide-ranging generalizations are not justified. But the utility of classroom observation techniques for assessing process goals within the open classroom framework was indicated. (Author/KM)