hard20 pts · first answer countslinear-regressionstatistics

Error spread grows with the prediction

In an OLS model of spend, residuals spread out much more for high-income customers. What is the main consequence for the coefficient estimates?

A.The coefficients become biased towards zeroB.R-squared becomes negativeC.The model can no longer be fittedD.Coefficients stay unbiased, but the usual standard errors and p-values are unreliable

← A link that only appears in a subset Direction of omitted variable bias →