Confounding Variable Identification Chain

validated scientific chain v1.0.0 cc-by-sa

Block ID: 7b1f97ea-184e-45af-9d9f-2a1f6dc411d7

Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.

Template

Identify and address confounders for the planned comparison in {comparison_context}. (1) State the exposure (or treatment) and outcome precisely, then map the causal question you actually want answered. (2) Brainstorm candidate confounders systematically by category — demographic, behavioral, environmental, temporal, and selection-related — a confounder must plausibly influence both the exposure and the outcome, so test each candidate against that dual requirement rather than listing every correlated variable. (3) Distinguish true confounders from mediators (which lie on the causal path and must not be adjusted for) and from colliders (adjusting for which introduces bias) — this distinction is where careful analysts and careless ones diverge. (4) For each confirmed confounder, choose a control strategy: design-stage (randomization, restriction, matching) is preferable to analysis-stage (stratification, regression adjustment) because it also handles unmeasured correlates. (5) Flag confounders that cannot be measured or controlled and state honestly how they limit the conclusions. (6) Specify the residual-confounding sensitivity analysis: how strong would an unmeasured confounder need to be to overturn the finding?

Identify and address confounders for the planned comparison in {comparison_context}. (1) State the exposure (or treatment) and outcome precisely, then map the causal question you actually want answered. (2) Brainstorm candidate confounders systematically by category — demographic, behavioral, environmental, temporal, and selection-related — a confounder must plausibly influence both the exposure and the outcome, so test each candidate against that dual requirement rather than listing every correlated variable. (3) Distinguish true confounders from mediators (which lie on the causal path and must not be adjusted for) and from colliders (adjusting for which introduces bias) — this distinction is where careful analysts and careless ones diverge. (4) For each confirmed confounder, choose a control strategy: design-stage (randomization, restriction, matching) is preferable to analysis-stage (stratification, regression adjustment) because it also handles unmeasured correlates. (5) Flag confounders that cannot be measured or controlled and state honestly how they limit the conclusions. (6) Specify the residual-confounding sensitivity analysis: how strong would an unmeasured confounder need to be to overturn the finding?

Variables

Name	Type	Required	Trust level
comparison_context		yes

confoundingexperimental-controlsstudy-designchain-of-thought

Ratings

0.0

Overall (0)

0.0

Accuracy

0.0

Consistency

0.0

Clarity

0.0

Efficiency

Benchmarks

Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.

Submitted by James P FounderMod via mcp · 2026-07-18