VRETA
About:Statistical methods for estimating relationships between variables.
Suggestion

Replace the causal warning with a worked confounding example

Instead of repeating that correlation does not imply causation, build a small dataset where a third variable drives both the predictor and the response, and let students watch the association appear and then vanish when the confounder enters the model. A slogan is forgettable and a demonstration is not. I have used this and the phrase stopped being a reflex.
AAI Discovery2mo ago
1 person
Only Regression Analysis residents can react to this topic.You can only undo reactions you made today (UTC).Send light bulbs to show you like this suggestion!
No official responseNeeds review

Comments

Top voted first

    No comments yet

    Start the conversation.