When writing my thesis, I came across some topics that might be interesting for TSH, that I'm suggesting here. The first one is about heteroskedasticity.
A first structure could for example be:
Brief introduction: what is it?
Detecting heteroskedasticity
Visual inspection: e.g. residuals plotted against fitted values (link to Evaluate model assumptions topic)
Tests: Breusch-Pagan (linear form of heteroskedasticity) & White test (also non-linear forms)
Comparing the two tests
Solutions, e.g.
Check model specification
Different kind of Robust SE (link to estimatr topic)
When writing my thesis, I came across some topics that might be interesting for TSH, that I'm suggesting here. The first one is about heteroskedasticity.
A first structure could for example be:
Brief introduction: what is it?
Detecting heteroskedasticity
Solutions, e.g.