Q104: Which test is not used for the presence of Heteroscedasticity? - Research Methodology and Econometrics - Agricultural Economics- ICAR JRF/SRF 2024 | Agrigyan

• Diagnostic Tests for Heteroscedasticity vs. Autocorrelation:
• Durbin-Watson Test: Used exclusively to detect the presence of first-order autocorrelation (serial correlation) among regression error residuals ($u_t = \rho u_{t-1} + \varepsilon_t$), not heteroscedasticity.
• Standard Tests for Heteroscedasticity:
• Goldfeld-Quandt Test: Tests whether error variances differ across two sorted subsamples.
• Glejser Test: Regresses absolute residuals $|e_i|$ on explanatory variables.
• Park Test: Regresses logarithm of squared residuals $\ln(e_i^2)$ on $\ln(X_i)$.
• White's General Test & Breusch-Pagan-Godfrey Test.

Correct option: (4) Durbin Watson Test