A QQ plot is used to assess which regression assumption?

A QQ plot is used to assess which regression assumption?

• A. Linearity.
• B. Normality of errors.
• C. Homoscedasticity.
• D. Independence of errors.

Answer: (B)

A QQ plot tests whether the regression errors are normally distributed. It does this by plotting the ordered residuals against the expected quantiles of a normal distribution. If the residuals follow normality, the points align closely along a straight line. Deviations from that line reveal departures from normality, such as skewness or heavy tails, which can affect the reliability of inferential statistics like standard errors and t-tests in ordinary least squares. This plot isn’t about whether the relationship is linear, whether the variance of residuals is constant across fitted values, or whether errors are independent. Linearity would show up as patterns in residuals versus fitted values; heteroscedasticity shows as increasing or decreasing spread in residuals; independence relates to correlations among residuals (e.g., in time series, via Durbin-Watson). The QQ plot specifically focuses on the distribution of the errors.

A QQ plot tests whether the regression errors are normally distributed. It does this by plotting the ordered residuals against the expected quantiles of a normal distribution. If the residuals follow normality, the points align closely along a straight line. Deviations from that line reveal departures from normality, such as skewness or heavy tails, which can affect the reliability of inferential statistics like standard errors and t-tests in ordinary least squares.

This plot isn’t about whether the relationship is linear, whether the variance of residuals is constant across fitted values, or whether errors are independent. Linearity would show up as patterns in residuals versus fitted values; heteroscedasticity shows as increasing or decreasing spread in residuals; independence relates to correlations among residuals (e.g., in time series, via Durbin-Watson). The QQ plot specifically focuses on the distribution of the errors.