Correlation and causation
Everyone can recite that correlation is not causation. The useful skill is knowing what would establish causation, and this course covers the alternatives to rule out and the designs that do it.
What you'll learn
10 chapters, then the quiz
- 01What correlation measuresMovement together, nothing more.
- 02The five explanationsA checklist before concluding.
- 03ConfoundersThe third factor.
- 04ExperimentsThe strongest evidence.
- 05Natural experimentsRandomness found rather than made.
- 06Weighing causal evidenceWhen experiments are impossible.
- 07Selection effectsWho ended up in the data.
- 08Regression to the meanA cause that is not there.
- 09Language and reportingWords that carry claims.
- 10Acting under uncertaintyDecisions cannot always wait.
More ai & data tests
See all →- Easy
Understanding AI
What machine learning actually does, what a language model is, and where both fall down.
8 questions7 min - Medium
Data analysis
Averages, spread, correlation and the traps that make numbers say the wrong thing.
8 questions7 min - Medium
Machine learning basics
Training, testing and the workflow behind a model that actually works.
8 questions8 min