A systematic study of the composition and dispersement of Title I projects assigned to elementary schools in Philadelphia was conducted. Categorical variables were identified from four major derived variables (program density code, school aggregate fund, pupil service component, achievement-growth differential score) and four major demographic variables (school staff, average daily attendance, per pupil and per teacher expenditures). Content analysis of the four program densities revealed (a) that four distinct implementation patterns (models) existed and (b) that the thrust of each model was programmatically different. The four implementation models were identified as (1) Educational/Cultural Enrichment Experiences; (2) General Instructional and Supervisory Support System; (3) Intensive Instructional and Supervisory Support System; and (4) Remediation Programs. Assessment of Pupil outputs provided by each model indicated that although significant differences were present between the grades, the anticipated levels of output were not achieved. However, six rational procedural alternatives for increasing the capability of the schools to increase their achievement levels within each model are provided. Findings suggest that traditional evaluation techniques are not suitable for assessing the programmatic effect of, or for developing operational management information about federal programs. (Author/AG)