Hi Sam,
It depends on which region dataset you’re looking at. Our Australasian dataset is based on the AusLCI dataset which may also use EcoInvent as a shadow database. Other regions are based on EcoInvent. You’re also right that the supply chain is very complex for electronics therefore the data will vary depending on which sources you reference and their assumptions. However there will be a range where the impacts typically fall within.
Best practice is to generally err on the side of ‘safety’ (i.e. what’s best for the planet) such as being conservative and assuming worse case scenarios. Ultimately the purpose of LCA is to be ‘generally right but not precisely wrong’ so as to steer outcomes that are better for the planet.
