Heteroskedasticity & Breusch-Pagan Test
Heteroskedasticity means residual variance is not constant — errors are small for some X values and large for others. This violates OLS assumptions and inflates standard errors.
Symptoms:
Symptoms:
- Funnel-shaped residual scatter (wider spread at high X)
- Breusch-Pagan p < 0.05
- Log-transform Y or large X variables
- Add or remove predictor variables
- Use weighted regression (WLS)
Try it in the app
Try: Run regression → check the Breusch-Pagan result below the coefficients table. If flagged (p < 0.05), try LOG10 transform on Y.