In an effort to validate a predictive (as opposed to a traditional formative or summative) cost-effectiveness model, a study was conducted to evaluate Kearsley and Compton's (1981) Benefits Model. Costs were input into the model as they applied to the design, development, and dissemination of two training programs on the topic of teaching individuals how to detect the level of drowsiness of their colleagues during team operations. The benefits of the training programs were identified, classified, and quantified as they applied to two media: lecture and computer-based multimedia.

The experimenter identified the training system parameters, training benefits, and operational benefits. Then, for the predictive approach, the relationships between training system parameters, training benefits, and operational benefits were classified based on expert opinion. Quantification concerned the assignment of values (-1 or +1) based on expert opinion. The costs to design, develop, and disseminate the training programs were determined based on the parameters of the project. Finally, based on all information present, experts determined which of two training programs would be the most cost-effective to disseminate.

To determine the accuracy of the Benefits Model as a predictive assessment tool, the same identified training system parameters, training benefits, and operational benefits were evaluated from a traditional formative evaluation approach. An empirical evaluation was conducted for the two training programs and a determination of the most cost-effective training medium was made. The data collected in the traditional formative evaluation approach was then compared to the experts' ratings and choice of training programs.

For both the predictive and formative evaluation approach to determining cost-effectiveness, the computer-based multimedia was chosen as the most cost-effective training medium. However, for the predictive approach, the experts’ choice was based either solely or heavily on dollar amounts associated with design, development, and dissemination, while the data obtained through the validation process were given little or no weight. All experts stated that it would not have been possible to use the information gathered through application of the Benefits Model to determine cost-effectiveness with any confidence.