Abstract
The product (or services) evaluation is a necessity in many steps of the development and release of such producs (services). The product evolution and the market share definition are two steps that evolves some kind of evaluation. Although, the evaluation occur by assigning grades not related with the perception of the product (service) quality –assigning values at random. This work presents 4 new models, 2 of them based on the CUB (Combined Uniform and shifted Binomial) model, for evaluation of items where there are two or more levels of preference by the evaluators and may have evaluators assigning grades at random.