Correlation Is Not Causation
Learn the everyday ways correlations mislead and what real causal evidence looks like
A taste of a lesson
Users who turn on notifications churn 30 percent less. Should we force notifications on for everyone?
Careful. People who choose to turn on notifications are probably already more engaged, and engagement itself predicts staying. So the feature may be a sign of commitment rather than a cause of it. Forcing notifications on could even annoy less engaged users and increase churn. The way to find out is an experiment: randomly prompt some new users to enable notifications and compare churn with those not prompted. Can you name one other difference between users who opt in and those who do not?
Written by the teacher as an example. In your lesson the tutor answers your own questions, and like any AI it can be wrong.
What you will be able to do
- List alternative explanations for any observed correlation
- Identify confounders, reverse causation and selection effects
- Explain Simpson's paradox with a simple example
- Describe what kinds of evidence support causal claims
- Rewrite causal sounding claims into accurate language
Lesson plan
- 1 Why things move together List the possible reasons behind any correlation. Start
- 2 Common causes Recognise confounders in everyday and business data. Start
- 3 Arrows the other way Spot reverse causation in claims. Start
- 4 Chance and selection Understand coincidental and selection driven correlations. Start
- 5 Simpson's paradox See how combining groups can reverse a trend. Start
- 6 What causal evidence looks like Judge claims by the strength of their causal evidence. Start
Try asking
About this tutor
A free beginner tutor that builds the most useful habit in data literacy: asking why two things move together. You will meet confounders, reverse causation, coincidence, selection effects and Simpson's paradox through memorable examples, then learn what kinds of evidence make a causal claim believable, from randomised experiments to natural experiments and dose response patterns. Each lesson ends with a claim from business or the news to diagnose. It is a good first step before experiment design or causal inference, and useful for anyone who reads dashboards.
Reviews
4.5
4 ratingsSample
- Mira S.Sample
Short, lively lessons. Good to share with non technical colleagues.
- Tariq M.Sample
The fire station puzzle hooked me. I now ask 'what else could explain this?' in every dashboard review.
- Ruth A.Sample
Free and excellent. Simpson's paradox finally makes sense thanks to the hospital style example.
- Lucas G.Sample
Great for building intuition. I wanted a little more on natural experiments, but the next tutor covers that.
About the teacher
Experiments, causal questions and responsible models, explained for decision makers
9 tutors 437 lessons taught Sample
I help analysts and product people answer the question behind most data work: did this change cause that result? I teach A/B testing, experiment design and the basics of causal inference, plus the responsible side of modelling: fairness checks and explaining predictions. My background is in product analytics and experimentation, where I learned that a clear sentence to a decision...
See Kojo's profile and tutorsMore like this
Other tutors on the same or nearby topics.