Purpose: This paper questions the robustness of social life cycle analysis (LCA), based on four social LCA case studies. To improve robustness of social LCA, it is a necessity to fight against its weaknesses. The paper addresses three questions: (1) what are its weaknesses? (2) How can they be combated? There are solutions suggested by the Conventions theory. The Conventions theory asserts that people are capable of adopting conventions (agreements between members of a group) to define what is fair and what is not, depending on the problem. The suggested solution consists in creating a new group (which has been called “extended community of peers”), which will define a new convention adapted to each new situation. The third question is, therefore, (3) do we need to resort to an extended community of peers to combat the social LCA weaknesses? Methods: To contribute to these debates, we discuss the classification of weaknesses defined by the Roy's decision-making assistance methods: (1) not dealing with the lack of knowledge, (2) attributing undue preferential meaning to certain data, (3) implementing misleading models, and (4) using meaningless technical parameters. We discuss the literature about creating new conventions thanks to peer involvement. To determine whether the creation of an extended community of peers influences the robustness, we will analyse four case studies (social LCA) which we conducted in 2011, 2012 and 2013. The first ones were conducted in Southern territories, relating to various agricultural products (banana, meat, orange). Another case study comes from a northern region, with the objective of comparing direct local supply systems and large-scale supply chains of various agricultural products. Results and discussion: About weaknesses in LCA, we highlight that environmental LCA authors have identified in their own works the same weakness points as Roy had done for other decision-making tools. We display that these weaknesses are present also