CONTRARY TO THE OPINION OF MANY PEOPLE, PROJECT HEADSTART (HS) IS NOT A STABLE AND UNIFORM PROGRAM WHICH DEALS WITH AN EASILY DEFINABLE POPULATION. THERE ARE, THEREFORE, SEVERAL PROBLEMS WHICH EXIST IN CONNECTION WITH EVALUATIVE RESEARCH CONCERNED WITH HS. IN ORDER TO PROVIDE GUIDANCE IN PROGRAM PLANNING, THIS RESEARCH SEEKS TO DESCRIBE POTENTIAL RECIPIENTS OF HS ATTENTION AND POTENTIALLY USEFUL PROGRAMS, TO ESTABLISH SPECIFIC RELATIONSHIPS BETWEEN PROGRAM ELEMENTS AND POPULATION CHARACTERISTICS, AND TO EVALUATE SPECIFIC HYPOTHESES IN TERMS OF USEFULNESS. DUE TO (1) THE COMPREHENSIVE MULTI-DIMENSIONAL NATURE OF HS, (2) THE SIMULTANEOUS PURSUIT OF BOTH IMMEDIATE AND ULTIMATE IMPACT, AND (3) THE PAUCITY OF INFORMATION ABOUT THE DISADVANTAGED POPULATION AND ABOUT PRESCHOOL EDUCATION PROGRAM ELEMENTS, THE GREATEST INITIAL PROBLEM CONCERNED WITH HS EVALUATIVE RESEARCH IS A CONCEPTUAL ONE, THE FORMULATION OF QUESTIONS WHICH ARE PROPERLY "RESEARCHABLE." THE SECOND PROBLEM IS THAT OF METHODOLOGY, HOW TO SAMPLE AND TO DEVELOP MEASUREMENT INSTRUMENTS. SAMPLING PROBLEMS ARE ENCOUNTERED BECAUSE OF THE NON-RANDOM VARIATIONS IN HS POPULATIONS AND THE INACCESSIBILITY OF SUITABLE CONTROL GROUPS. THE THIRD PROBLEM IS THAT OF LOGISTIC DIFFICULTIES. IT IS NECESSARY FOR EVALUATIVE PROCEDURES TO BE UNOBTRUSIVE. TYPICAL CIRCUMSTANCES OF THE DISADVANTAGED HOME, LOW LITERACY LEVELS, AND THE PROBLEM OF RAPPORT BETWEEN DISADVANTAGED ADULTS AND MIDDLE-CLASS SCIENTISTS CONTRIBUTE TO THE LOGISTIC DIFFICULTIES ENCOUNTERED IN GATHERING RESEARCH DATA. THE FOURTH PROBLEM IS THE INTERPRETATION OF DATA IN HS EVALUATIVE RESEARCH. THIS PAPER WAS PRESENTED IN A SYMPOSIUM AT THE AMERICAN EDUCATIONAL RESEARCH ASSOCIATION MEETINGS, CHICAGO, ILLINOIS, FEBRUARY 10, 1968. (JS)