THE EFFECTS OF DATE OF TESTING AND METHOD OF INTERPOLATION ON THE USE OF STANDARDIZED TEST SCORES IN THE EVALUATION OF LARGE‐SCALE EDUCATIONAL PROGRAMS

摘要: In evaluating the impact of educational programs, most common criterion employed is academic growth. many nonexperimental studies, standardized achievement tests are administered before and after program implementation gain scores computed. This approach widely used (e.g., see literature concerning Title I evaluation models: Tallmadge & Wood, 1976; Horst, 1976), in some cases decisions on funding, expansion, termination local projects based results from this simple straightforward procedure California State Department Education, 1976, 1977). view these uses scores, it seems appropriate to examine critically actual procedures calculate represent The specific considered here those typical large-scale studies where neither test nor testing dates can be controlled. A widespread analytical practice such transform fall spring raw standard using publishers' norms as difference between scores. While a number questions raised about David Pelavin, 1977; Devito Long, Linn Slinde, 1977), present paper focuses two issues computation use gains: effects 1) estimated 2) discrepant by publishers.