garretchristensen
commented
9 years ago similar to Leamer, but in sociology: http://web.stanford.edu/~cy10/public/mrobust/Model_Robustness.pdf
Open garretchristensen opened 9 years ago
garretchristensen
commented
9 years ago similar to Leamer, but in sociology: http://web.stanford.edu/~cy10/public/mrobust/Model_Robustness.pdf
The section on human rights abuses seemed a bit unnecessary for example (it's not really nec to go into the background there in order to mention IRBs)
One missing topic: you mention but didn't really go into the topic of confidential data, IRBs, and informed consent. So - where data-sharing can't be public b/c of risk of identifying participants e.g fine-grained geospatial data + other variables, what are the best practices and what resources are out there? Other missing topics: you seem to be covering a lot of ground on reproducible research and requirements, but didn't go into these topics: Data citation principles Metadata - what should be included? Funder data access policies and data management plans (along with resources to help with this) Relation between what's shared (data/code) and how much data/code should be shared to allow for fuller "reliability checks" etc.
One bigger picture thought. Some of the guide briefly goes into practical "how to" recs such as on code/data management... that feels a bit out of keeping with the rest of the topics, which are background/narratives... I wonder if you might consider segmenting off the practical tips even more than you do. For example, having a Part I: background, and Part II: tips for reproducible research. That way, people could more easily skip the first part if they just want your "hands-on" advice. Otherwise, I'm not sure those people would keep reading long enough to get to the concrete recommendations.