When there is heteroskedasticity, this means that C] The OLS estimates of the coefficients are biased [:J You can no longer interpret the R2 in the same way [:J The estimate of the standard error is no longer valid. C] The usual t-statistic is still valid