What’s up and where to go in social life cycle impact assessment cause-effect model development?

Purpose The paper contributes to the S-LCA impact assessment by understanding its maturity level to identify each characterization model’s good practices and limitations. Method Our research analyzes the S-LCIA type II (impact pathway) models’ current state of the art through a systematic review. To this end, we identified publications through international databases without period restriction up to 2020. Gray literature, theoretical propositions, and papers covering solely case studies of former developed models were not included. Among the 102 papers identified, only 14 papers presented type II modeling. These models were evaluated considering the following criteria and sub-criteria: (i) scope coverage (contextualization, time lag, and data level), (ii) scientific robustness (presentation of cause-and-effect chain, transparency, and accessibility), and (iii) characterization factor (CF) (existence, feasibility of regionalized CFs, and presence of uncertainty or sensitivity analysis). Each criterion is composed of sub-criteria related to desirable characteristics. We classified models on scales comprising two or three levels for each sub-criterion. Results and discussion We noticed some models need to provide clear cause-and-effect chains, relationships with social sciences, and measurement models are not always explicit. Most of them do not present ready-to-use characterization factors, and only a few models consider the time lag between cause and effect. Finally, several models are not always connected to company data. Based on the results, we provide recommendations for impact modeling in S-LCIA to support life cycle social and sustainability assessment. Conclusions Despite advances in S-LCIA impact pathway modeling, there are necessary improvements to support S-LCA. We identified good practices, but no model presents all characteristics. Even if the criteria proposed to evaluate the type II models are only partially comprehensive, this framework supports S-LCIA model development by listing desirable characteristics and recommended good practices. We advise (i) including the model’s usage limitations, (ii) using time series analysis to include the time lag between cause and effect or to inform if the effect is immediate, (iii) and using company or sector data to develop the characterization factors. The characterization models with links to social sciences, the cause-effect chain, and the reproducibility of the measurement model can improve the models’