Bayesian Causal Inference in Political Science

An essay for harmony between Bayesian analysis and causal inference in political science

About this repo

It probably doesn't work to build it yourself from scratch yet. See 'technical' section in to-do list

To do

Legwork

• WIPs (Hall)
  • Decide which interventions (unif, logit)
  • Priors and PPCs (?)
  • Estimate
  • Diagnoses (appendix)
  • Posterior figs
  • Posterior R inline
  • Replace "intro" code with brm (or clean it up, whatever)
• Pooling / Regularization (Reeves et al)
  • Which priors
  • PPCs into paper somewhere (ugh)
  • Estimate
  • Diagnosis (appendix)
  • Posteriors
• Regularized conjoint (cut)
  • Decide if this is gonna make it

Writing

• Feedback solicitation
  • What's the audience
  • What's the purpose (argument vs. explanation vs. tactical demonstration)
  • What needs more careful explanation or (yuck) "hand-holding"
• Intro
  • onboarding in the right place?
• Shared goals
• Norm clash
• Examples
• Going forward
  • Nonparametrics?
  • Bayesian ML tactics?
  • Sensitivity testing
  • Marginalizing over choices (mediation sensitivity parameter, matching, assumptions)

Reverse outline

• build
• fix

Technical

• Github
• Some kind of build file
• bib isn't self-contained
• font notes
• notes files