In decision-theoretic psychometrics, to determine the worth of individualizing instruction, equations are developed for expressing the cost and gain for applying an instructional sequence. This report is divided into three parts: first, a logical analysis of guessing; second, the effect of guessing on the quality of personnel and counselling decisions; third, the worth of individualizing instruction. The expected return from assigning instruction on the basis of admissible probability measurement, admissible choice testing, conventional choice testing, prior information only and matching the average student is computed for each of seven distributions of state of knowledge. The performance of admissible probability measurement is outstanding, matching the average student is surprisingly good, while the others are disappointing. (BB)