The adequacy of using a combination of criterion-scaled non-school variables to predict achievement score means of schools is discussed. Non-school data gathered on eighth-grade students were criterion-scaled, using the total score of the Metropolitan Achievement Test. Eight non-school category scores and actual achievement scores of selected schools were intercorrelated. A stepwise regression analysis yielded a multiple correlation of .949 between predicted and actual scores. Further steps, including a quasi cross-validation study, confirmed the feasibility of the method. Results suggested that the technique could lead to a satisfactory accountability model, could direct educators to focus efforts on non-school variables, and could justify concentration on neglected goals. (For related document, see TM 002 508.) (Author)