All of the following are assumptions underlying the validity of linear regression output except:

A) The errors are normally distributed.

B) The mean of the errors is zero.

C) Certainty.

D) The standard deviation of the errors is constant.