This project proposal is a joint project between the universities in Stavanger, Norway, Lund, Sweden and SINTEF, Trondheim. The project proposal emphasises specific issues and judgements to be made in accident investigation, and how learning processes fro m the accident investigations are dependent of their specific societal contexts. A comparison between Norwegian and Swedish (also Danish) accident investigations and follow up activities is sought in order to explain the significance of contributing facto rs in the learning processes of the two countries. The project will test the main hypothesis that accident investigations play an insignificant role in the learning at individual, organizational and cross sectoral levels in society. This hypothesis presup poses 5 sub-hypotheses:The learning perspective of accident investigation is poorly founded. The learning effects are minor for actors not directly involved in the accident. Accident investigation reports only infer changes and learning when governments instruct changes to regulations. Accident investigation results are often fronted in political populist cases, which usually fall to the ground without long term results. The learning within transport organizations are "arbitrary", culturally and direct e xperience driven, rather than systematic experience transfer. The presence and extent of learning across societal sectors are highly dependant on the structure (and position in society) of the accident investigation committees.The project consist of four workpackages, which mirrors the primary aims: a deduction of a analytic framework for assessing learning effects from accident investigation (WP1), a comparison of Norwegian and Swedish approaches and related effects on learning based on cases in each co untry (WP2), a study of Norwegian emergency management organisations? learning processes (WP3), and a case study comparing multilevel learning from accidents and incidents within transport organisations (WP4).