<p>Intracriteria evaluation plays a crucial role in Multicriteria Decision Making (MCDM), especially for methods based on additive models within the scope of Multiattribute Value Theory (MAVT). The main goal of the intracriteria evaluation step in additive models is to build marginal value functions that represent the decision-maker’s (DM’s) values and interests when evaluating increments of values within the scale of consequences of each criterion of the problem. However, a major challenge arises when it comes to eliciting a DM’s preferences to build such functions, since the process is generally highly cognitively demanding and time-consuming. In order to avoid these issues, it is a common practice to assume linearity in the form of the marginal value functions. In such cases, no information gathered from the DM’s preferences and the value functions is obtained using a simple normalization process. This simplification, however, may lead to distortions in the final results obtained by the MCDM model, since the DM’s preferences are not accurately reflected in it. Therefore, this paper aims to address the challenge of eliciting preferences in the intracriteria evaluation step, by proposing a structured partial information-based elicitation procedure for doing so, in order to save the DM effort and time. The proposed approach allows nonlinearities inherently present in a DMs’ preference structure to be reflected by the marginal value functions, while offering a less cognitively demanding elicitation process based on partial information. The proposed approach was incorporated into the structure of the FITradeoff multicriteria method and its Decision Support System (DSS). A practical application illustrates the proposed approach in a supplier selection problem, accompanied by a discussion based on the results obtained.</p>