1.4 Motivation: The causal inference ‘revolution’

  • Revolution of identification (K. Imai 2011; Keele 2015a, 102)1
  • Revolution of potential outcomes (e.g. Rubin 1974)2
  • New Style of writing/conducting data analysis
  • New rationales for evaluating research (Causal empiricism)
  • Impact on research questions3
  • Journal of Causal Inference4
  • Exciting ongoing debates (e.g., on RCTs)5
  • Methodological socialization across generations

Notes

….

References

Imai, Kosuke. 2011. “Introduction to the Virtual Issue: Past and Future Research Agenda on Causal Inference.” Political Analysis.

Keele, Luke. 2015a. “The Discipline of Identification.” PS Polit. Sci. Polit. 48 (01): 102–6.

Rubin, Donald B. 1974. “Estimating Causal Effects of Treatments in Randomized and Nonrandomized Studies.” J. Educ. Psychol. 66 (5): 688–701.

  1. Focuses on what part of an association (e.g. a regression coefficient or a correlation) can really be attributed to the causal effect of D on Y. And, what assumptions we need to make to reduce an association to its causal component.↩

  2. Classically used methods (e.g. OLS regression) are reinterpreted departing from the framework of potential outcomes.↩

  3. Focus on causal inference has triggered certain trends, e.g. searching for exogenous variation, e.g. weather, natural disasters, historical factors etc. Once a particular design becomes popular, researchers tend to copy and hunt for use-cases of that design (e.g. regression discontinuity design).↩

  4. “Existing discipline-specific journals tend to bury causal analysis in the language and methods of traditional statistical methodologies […] JCI highlights both the uniqueness and interdisciplinary nature of causal research” (The editors)↩

  5. See for instance the debate on the value of randomized control trials (e.g. see Cuesta and Imai 2016; Deaton 2010; Imbens 2010). See videos 1, 2 and 3.↩