In past Head Start evaluations, cognitive measures have been weighed heavily. This has not accurately reflected the relative unimportance of cognitive program goals; child performance gains are not an objective with high priority for most Head Start programs. Evaluation planners need to weigh previously encountered measurement problems carefully and decide to adopt either a reliability-based strategy placing emphasis on careful test administration or a validity-based strategy assuming that what is needed is a fundamental reconceptualization of the measurement of cognitive effects, developing new measures. As priorities for cognitive measurement, this study argues that the new evaluation should stress readiness, cognitive process, and social competency and if it is decided to adopt a validity-based strategy, lists of clearly defined behavioral objectives must be drawn up in those realms of stress and then to create or adopt instruments to measure these objectives. What is needed is a battery of face-valid, empirically based, criterion-referenced instruments intended to measure short-term effects. Choice of measures is integrally related to choice of evaluation design. The new evaluation might consider some departure from pre- and post-testing, instead testing three times during the year or only once at the end. (RC)