to do notes #1
Description
We'll probably have a different repo, but to formalize our notes from the meeting:
1- Changes to the underlying codebase:
- Make citation networks respect triadic closure
- Make sure that network properties are ensured post publication bias
- Vary sample size on a study by study base
- speed up model estimation using cmdstanr and update (instead of recompiling)
2- Extract from metalab
- Citation networks (to identify distributions of relevant network parameters)
- meta-analytic effect size + sample size by study
3- Overview of the story:
- Information flow and the impact it has on posterior passing
- linear vs. citation network
- increasing publication bias
- running meta-analysis (for practical implications)
4- Additional references
- Kossinet & Watts for estimating triadic closure (in citation networks?) & Mikkel Werling for code implementing the generation of networks of that kind (if useful)
- https://www.bmj.com/content/339/bmj.b2680 (on extreme citation bias)
- https://academic.oup.com/sf/article/93/4/1687/2332185?login=true (on information flows, more to check in their bibliography)