Research indexNon-randomized trial
因果推論與觀察研究:「反事實模型」之思考
Huang
1 Apr 2008
NobleIDark:/48914/w1/20260522/4C18E0F8DOI10.30401/rss.200804_2(1).0001
因果推論與觀察研究:「反事實模型」之思考
Huang · 2008 · Subscription
NobleIDark:/48914/w1/20260522/4C18E0F8DOI10.30401/rss.200804_2(1).0001
The full text is at the publisher
You can read the abstract here, or open it at the publisher.
Abstract
The core of the counterfactual model of causality (CMC) is simple. To argue that D is the cause of Y, we must ask ”What would Y have been if D were not the case?” In other words, we should not rely solely on the observed regularities to infer causality. Instead, researchers need to compare the realized outcome (i.e. factual) with its potential outcome (i.e. counterfactual). This potential outcome model forces us to explicitly state and make operational the counterfactual with a clear implication of what should be controlled or compared. This has been developed into a unified framework for causal inference based on randomized experiments, quasi-experiments, natural experiments, as well as non-experimental observational studies. This recent trend is indeed exciting for social science research targeted to address cause-and-effect questions and yet impossible or difficult to conduct lab experiments. CMC stimulates a new wave of reexamination of more traditional concepts and methods of causal inference in social science research.