As the first step in a program to develop quantitative techniques for prescribing the design and use of training systems, the present study attempted: to compile an initial set of quantitative indices, to determine whether these indices could be used to describe a sample of trainee tasks and differentiate among them, to develop a predictive methodology based upon the indices, and to assess that methodology. The compilation of indices includes the Display-Evaluative Index, a set of panel lay-out indices, and a set of task-rating scales. Application of the indices proved feasible; differentiation among three training devices and within four trainee subtasks was possible. The predictive method which was then generated is an adaptation of the standard multiple regression model. Mean task scores replace the usual individual criterion scores, and quantitative task index values are used as predictor scores. The adaptation was tested using data from published studies in tracking. Significant multiple correlations using task indices were found for criterion data obtained during the early stages of practice. A combination of task and training indices did predict later performance. This result supports the contention that a prescriptive method must include both training and task indices in order to account for advanced levels of proficiency. (Author/JY)