While discussing the assumptions of classical linear regression model, one of the assumptions was of equal variance of the error term i.e. E(ui^2) = σ^2, in these videos, we will figure out, what will happen if this assumption is not met.
Hetroscedasticity is the problem in cross-section data. Imagine a dataset of income of households in some city, there will be both rich households and poor households. Do you think that expenditure patterns of both kinds of households are same? No. This is the problem of hetroscedasticity. We discuss the following in these videos:
- What is hetroscedasticity?
- What are the consequences of ignoring hetroscedasticity?
- How do you test for hetroscedasticity? LM Test and White’s test for hetroscedasticity
- What are the assumptions about the pattern of hetroscedasticity?
Consequences of ignoring hetroscedasticity
LM Tests for Hetroscedasticity
Whites test for hetroscedasticity
Assumptions about the pattern of hetroscedasticity