Rigorous causal research is essential for producing evidence that can inform better policies, programs, and decision-making. A new article, published in Observational Studies and authored by Westat researchers,introduces a Total Causal Error (TCE) framework for systematically examining threats to causal conclusions throughout the research process.
Building on the Total Survey Error (TSE) framework and established causal inference methods, TCE organizes potential sources of error along 3 parallel paths: causal identification, representation, and measurement. The framework helps researchers consider how decisions made during study design, data collection, processing, analysis, and interpretation can shape the causal conclusions that a study can support.
As the authors note, “causal conclusions are more fatally threatened by upstream design choices than by downstream analytic adjustments.” TCE therefore encourages researchers to consider potential threats early, examine trade-offs among different sources of error, and make the assumptions underlying causal conclusions more explicit.
By providing a shared language across research domains, TCE can help researchers examine why studies addressing similar questions may produce different findings and strengthen the cumulative evidence used to address complex challenges.
Learn More
Elizabeth Willow Eisenhauer, Daifeng Han, Brad Edwards, and J. Michael Brick