A simple multiattribute utility procedure for evaluation

Multiattribute utility models have been proposed as a way to determine the worth or utility of objects/alternatives in a wide variety of situations, e.g., land management, water quality evaluation, etc. A simple multiattribute utility procedure, smaup, is proposed here under the following conditions: (1) The attributes can be scaled to be monotonically increasing in utility and (2) the attributes are conditionally monotone as well. In SMAUP, experts are not needed to determine utility functions or importance weights. Therefore, the procedure is very easy to use. An example is presented using the evaluation of player performance in the National Basketball Association, NBA, for the 1973-1974 and 1974-1975 seasons. Eight attributes of player performance were used as input to the model. The model yielded a rank order of all forwards, guards, and centers for each season. This ordering was then compared with the NBA all-star team. The results showed that the model predicted the all-star team well. However, the basic issue of whether clinical judgment or multiattribute utility procedures yields better results is discussed. Furthermore, various methodological issues raised by the procedure are examined.