Managers, in decision making process, are usually facing with choice problems. Choosing an alternative between the number of alternatives, selecting a number of factors among a set of criteria (screening phase) in order to better analyzing a problem, and choosing a strategy to improve public welfare service among a number of strategies are the examples of this situation. In this paper, selection of factors among a set of different criteria is discussed. This activity can be called "screening phase of criteria". There are different approaches to screen the criteria suchas filtering in a group with experts, voting, 80-20 rule, etc. The problem with these methods is the existence of multiple stakeholders with diversity of viewpoints. Each expert has his/her own view about a specific criterion and, for this reason, the experts put emphasis on their selected criteria. Therefore, each individual expert evaluates his/her chosen criteria more than the others. Usually, those criteria are chosen, which are common among experts in order to cover the views points of stakeholders. In other words, those criteria finally should be considered to have minimum disagreement. This paper uses goal programming approach for maximizing the consensus and minimizing the disagreements between the experts with ordinal scale preferences. The goal programming models are discussed in different aspects. In order to validate the models, a real case problem has been used.