Skip to content
Quantoros

A heteroscedasticity test does not reject. Is constant variance proven?

StatisticsLevel: IntermediateRegression diagnostics

A heteroscedasticity test does not reject. Is constant variance proven?

Hints
  • Failure to reject is not proof.
  • Consider the tested alternative.

Answer

No. Power may be low or the chosen variance predictors may miss the pattern. Combine test results with residual plots and sensitivity-aware inference.

Source and licence: statsmodels — model diagnostics · BSD-3-Clause · answer: Quantoros editorial

Related questions