Just a suggestion. The assessment of a societal impact of innovation is theoretically impossible when the innovation causes a structural change. So, you could change the scale and considering case studies, you could analyse and produce only an evaluation of the local impact. Considering the case study as an experimentation, you could (or not) generalize your results.

Thanks Maurice, I am aware that this is an ambigious intent. I was just trying to work the other way around by using results. Combining them while telling apart other possible causes would be the approach this time.