Search for your outcome variable in the Variable Browser or ask Pulse "What predicts [X]?" to confirm it's in the model set.
Go to Discover →From the DV discovery card or IV Explorer, identify candidate IVs — variables with high partial coefficients that survive regularization.
In the Hypothesis Test, set your focal IV, add standard demographic controls (age, gender, race, education, income), and choose your election year(s).
Open Hypothesis Test →Confirm statistical significance of your IV. Review model fit (R²/AUC) and whether controls are behaving as expected theoretically.
Re-run the same specification on different election years. Stable coefficients across years strengthen causal inference.
The Lab is where raw ANES survey items are assembled into the composite indices that political scientists actually use in research — and where Pulse re-runs the landmark studies that established those constructs. It lets you inspect scale reliability (Cronbach's α), view the item distribution, and launch a hypothesis test with the index as a predictor (IV) in any model you build. Each index is a continuous variable wired directly into the Hypothesis Test engine.
Two tracks of self-paced tutorials — three on statistical reasoning, three on machine learning. Each builds on the previous. New users should start with Layer 1; researchers familiar with regression can jump to ML Tier 1.