# Slide Outline: Causal Inference Workshop

## 1. Workshop Goal

- Turn one observational question into a criticizable identification memo.
- End with an estimand, DAG in words, assumptions, diagnostics, and sensitivity plan.

## 2. Start With The Estimand

- Treatment
- Control
- Population
- Outcome
- Time horizon
- Effect definition

## 3. DAG In Words

- Confounders
- Mediators
- Colliders
- Selection mechanisms
- Variables that should not be adjusted for

## 4. Identification Strategy

- Adjustment, matching, DiD, IV, RDD, or new experiment
- Why this strategy fits the data-generating story
- What assumption would break it

## 5. Team Exercise

- Draft the one-page memo.
- Mark the weakest assumption.
- Name one diagnostic and one sensitivity check.

## 6. Peer Critique

- What unmeasured confounder worries you most?
- Where is overlap weak?
- Would a randomized prompt or rollout be more credible?

## 7. Close

- Revise the claim from "X caused Y" to "Under these assumptions, the evidence suggests..."