Topic: Common regression model misspecifications

In this essay, I am not asking you simply for definitions of multicollinearity or serial correlation or conditional heteroscedasticity. Rather, I want you to discuss:

(a) Ways that models can be misspecified;

(b) The impact that various misspecifications can have on the estimated values of the coefficients (such as biased or inconsistent estimates);

(c) What types of misspecifications are the most serious; and

(d) You should conclude with ways to detect/avoid/and correct for misspecification.