A survey of administrators, teachers, and parents in 15 Elementary Secondary Education Act Title I school districts indicated that evaluations are used locally to meet federal and state requirements, to inform parents and staff, and to confirm positive attitudes toward a program. Standardized achievement tests, the backbone of Title I evaluations, are viewed as inadequate for judging or improving programs. Respondents felt tests were biased and preferred other measures: skill-specific tests, observation, self-concept or attitude measures. Evaluations were not used to improve programs for several reasons: program stability; funds; politics; unavailability of results in time for decision making; differing information needs of federal, local, and state agencies. Respondents disliked evaluation and ignored negative results if they believed in a program. Until these underlying attitudes change, evaluation results, even if technically sound, will not be used. Of the two current federal strategies for Title I evaluation, an independent national study and a local-to-state-to-federal reporting scheme, the latter is more likely to be used. The federal government should be committed to increasing communication between program staff and evaluation staff, and to promoting evaluation among local staff. (CP)