Disentangling blended feedback

Having just been working on a project where workforce and audience acceptance of a mix of new hardware, software and content needed to be assessed & tested, I'm keen to find out the best way of gaining a blended system acceptance view without confusing the feedback for the separate elements. I know this should focus on asking precise questions, and directing users to complete specific targeted tasks, but users often seem unable to disentangle feedback in one dimension from another... (usually blaming hardware or software for eveerything). Does anyone have any helpful advice? I must confess that I'm not used to adding the hardware angle.