Regression Assumptions and Diagnostics

Start with the assumptions

The mind map Assumptions of Regressions offers a starting point for this collection. It mentions the distributions of variables, a linear relationship and ways to check features of the data, including skewness, kurtosis and a scatterplot. Use it to identify the assumptions you want to look at more closely.

The flashcards in Assumptions of regression provide another way to work through the topic. They include prompts on model specification, measurement error, independence, normality and homoskedasticity.

Explore heteroskedasticity

The Heteroskedasticity mind map separates consequences for regression estimates from effects on variance estimates and standard errors. After reading the map, use Heteroskedasticity part 1 to recall what the term means, why it is a problem, its possible causes and ways to test for it.

Move between the map and the flashcards: first follow the connections on the map, then answer the card prompts without looking back. Return to the map when you need to check a relationship.

Practice with diagnostics

The Multiple Regression quiz includes a question about using the Durbin-Watson statistic to assess independence. The Regression - Chapter Six flashcards include prompts about residuals and leverage points. For questions about relationships between predictors, SP500 Week 6 Choosing Predictors covers collinearity and suppression.

Use the quiz to practice, then return to the relevant map or flashcards to review the ideas behind a question. The resources offer different starting points, so you can choose the format that fits the issue you are studying.

Place the topic in the unit

Regression Assumptions and Diagnostics is a section of the Correlation and Linear Regression unit in Statistics and probability. The resources here focus on assumptions and diagnostic questions within that unit, with maps for connections and cards for recall. Choose a resource and work through its prompts at your own pace.