Autocorrelation of the residuals may affect the reliability of the t-values for the estimated coefficients of the predictors X₁, X_2,…, X_k.

Question

Quizplus · Accepted Answer

Autocorrelation of the residuals violates the assumption of independent errors in linear regression, which results in biased standard errors of the estimated coefficients and unreliable t-values. This means that the significance of the predictors X₁, X_2,…, X_k may be overestimated or underestimated, leading to incorrect conclusions about their importance in the model.

Autocorrelation of the Residuals May Affect the Reliability of the T-Values