This essay provides some useful concepts for the discussions we've been having:

http://distill.pub/2017/research-debt/

In particular, it takes up the concept of "interpretive labor" from the anthropologist David Graeber and proposes that we think of a lack of this kind of labor in the context of scholarly work as "research debt." 

As I understand it, Distill is a new publication that tries to catch up on some much needed interpretive labor in machine learning, but I can see applications for the term in other areas, such as computational social science, as well.

For instance, we might ask questions like: Where do we see research debt building up in our field? Is programming a "high interest credit card" of research debt in the social sciences, to borrow a formulation from Sculley et al. (https://research.google.com/pubs/pub43146.html)?

I hesitate to single out one particular field or area of research as being particularly research debt-laden, but I thought this might be a stimulating idea to share with the list.

I look forward to your thoughts,
John