It's very well possible to have the right conclusion ('climate change is man made') while working with imprecise data and models. High precision is not required for consensus about cause and effect to be established, merely that within the boundary of error no other explanation is viable.