Tool

Causal Hypothesis Builder

An interactive, saveable worksheet that helps practitioners convert observations into clear causal questions, map causal pathways (including confounders and mediators), choose feasible identification strategies (DiD, IV, RDD, synthetic control, matching, natural experiments), and record the data and robustness checks needed to validate or falsify the claim.

Interactive Tool

Causal Hypothesis Builder

This guided worksheet helps you turn an observation into a testable causal hypothesis and a practical identification plan. Use it to state a precise causal question, sketch the causal pathway and likely confounders, pick feasible quasi-experimental strategies, and record the data and robustness checks you'll need to defend or refute the claim. Save your work so you can iterate, share, or export a pre-analysis checklist.

Write the causal effect you want to estimate in a single sentence. Example: 'Does offering free maintenance (treatment) increase 6-month equipment uptime (outcome) among small manufacturers (population)?'
Be specific: variable name, unit, aggregation frequency, and any transformation (e.g., log, binary).
Describe what counts as 'treated' vs 'control', timing, intensity, and any eligibility rules.
Who is included? Any inclusion/exclusion criteria? Start and end dates for observations?
Sketch the plausible causal chain from treatment to outcome. Note mediators and potential indirect effects. A short bullet list or sentence is fine.
List obvious and less-obvious confounders. For each, note whether you can measure it and how well.
Describe any ways selection into treatment or loss to follow-up could bias estimates (and how you'll detect or address them).
Select one or more feasible strategies. After selecting, justify your primary choice below.
Explain why the chosen strategy is feasible, what assumption it relies on (e.g., parallel trends, valid instrument, no manipulation at cutoff), and what data elements support it.
Check items you have or can reasonably obtain. Missing items should drive feasibility decisions.
Select the assumptions you'll document and show how you will test or falsify them.
List placebo outcomes, lead tests, or subgroups where you expect no effect as checks against spurious correlations.
E.g., alternate control groups, varying bandwidths (RDD), different covariate sets, clustering choices, trimming or weighting rules.
Record immediate actions needed to make the analysis possible and defensible.
How important or actionable is this question right now? Use this to prioritize analysis work.
1.0 10.0
Anything you want a colleague or reviewer to check or pay attention to.
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.