Pitfalls In Evaluation

Abstract

Dangers in evaluation are of two types: failures of perception and failures in timing. Avoidance of failures of perception requires care on design of the evaluative experiment to assure internal validity of inference without loss of generalizability. It also requires careful definition of objectives of the innovation and expression of objectives in a comprehensive set of measures of effectiveness and cost, plus an awareness of and accounting for human factors involved. Avoidance of failures in timing require the integration of evaluation into the stages of research, development and diffusion.