Text: Question 1 [15 marks in total, 3 marks for each part] Answer each of the following. In each case, state whether the statement is True, False, or Uncertain and explain your answer.
(i)
The presence of omitted variables calls into question the use of a simple linear regression model y = Po + βx + u unless one uses homogeneous subsamples as a means of controlling for other variables. So, in a wage equation where the focus is on the impact of education, the sample could be restricted, say, to females with the same occupation and level of experience.
(ii)
In a regression study of electricity demand, the price response was found to be insignificantly different from zero. Therefore, the price is not an important determinant of electricity consumption.
(iii)
Suppose the regression equation y = Po + βx + β2z + u is estimated, and heteroskedasticity related to z is indicated; then a possible solution is to remove z from the regression equation.
(iv)
Suppose you are analyzing how long doctoral students in economics take to complete their PhDs. Using data on a large sample of students who commenced studies eight years ago, you have many relevant explanatory variables, but some students have not yet completed their degrees. The best solution is to drop these observations and analyze those for whom you know the duration of their studies.
(v)
A dependent variable with a deterministic trend that is weakly dependent can be used directly in multiple regression analysis.
