Methods and research design
Statistical method comparison
An outcome model, an identification argument and a covariance choice answer different questions. Specify your setup to examine the distinctions, relevant Stata syntax and the assumptions still requiring evidence.
Your comparison
Change setup ↑Choose an outcome, observation structure, purpose and dependence assumption.
Three distinct decisions
- Outcome model
- What relationship or distribution does the specification describe?
- Research design
- What evidence would allow a causal interpretation?
- Inference
- Which observations can be treated as independent?
“Causal design unresolved” and “Not established yet” are valid selections. They leave the unresolved assumptions visible.
What to report
State the estimand, units, sample restrictions, covariance, identifying assumptions and material limitations. Distinguish a conditional association from a causal claim. None of these selections validates a model.