EP 01 · Confounding
A short animated walkthrough of the Causal Thinking Lab's "Classic Confounding" scenario: how a third variable inflates a crude risk ratio of 1.35, and how stratification recovers the true effect of 1.00.
Long-form pieces, animations, and infographics that extend the book. Read, watch, or download — and submit your own work for the journal.
The journal welcomes short, methodologically-grounded pieces from students, instructors, and practitioners. Pitches under 200 words, full drafts 1,500–4,000 words.
A short animated walkthrough of the Causal Thinking Lab's "Classic Confounding" scenario: how a third variable inflates a crude risk ratio of 1.35, and how stratification recovers the true effect of 1.00.
Visualising what "adjustment" actually does: stratifying the population, computing within-stratum effects, and pooling them into a single adjusted estimate.
The grammar of directed acyclic graphs: nodes, arrows, paths, colliders, and how to spot a confounder at a glance.
A single-page reference: how non-random selection into a study distorts associations, with three canonical mechanisms drawn in.