Since the dependent variable is approximately normally distributed, the simple linear model allows for unbiased estimation and there is no need for nonlinear alternatives, such as ordered logistic regression, that are difficult to interpret and impose arbitrary assumptions about the functional form.