The data set is based on a sample of sale prices (965 apartments). Variables Number Total_Price Price_sqm Area Floor Location N_bedrm N_bathrm Exercise Description Number of residential area Total sold price ($US) Sold price per Sqm ($US) Total construction area of the apartment Floor that the apartment is situated Distance to the CBD increases (Km) Number of bedrooms in the apartment Number of bathrooms in the apartment 1 Estimate: P = f (Area) + ε. How much of the variation around the average price can be explained by the variation in the total construction area of the apartment? What do the t-values stand for? Economically interpret the coefficient. 2 Estimate: P = f (Location) + ε. How much of the variation around the average price can be explained by the variation in the Location of the apartment? What do the t-values stand for? Economically interpret the coefficient 3 Estimate: P = f (N_bedrm) + ε. How much of the variation around the average price can be explained by the variation in the number of bedrooms of the apartment? What do the t-values stand for? Economically interpret the coefficient 4 Estimate: P = f (Floor) + ε. How much of the variation around the average price can be explained by the variation in the number floors of the apartment? What do the t-values stand for? Economically interpret the coefficient. 5 Incorporate location in the regression model. Estimate: P = f (Area, Location) + ε. Interpret the parameters. Is model 5 better than model 1 and model 2? 6 Estimate: P = f (Area, Location, N_bedrm) + ε. economically interpret the coefficients. Compare with the previous model 5. 7 Using model 6 above, predict the price of a property that has an area of 145sqm, 3 bedrooms, and is located 1km to the CBD. 8 Estimate: P = f (Area, Location, Floor, N_bedrm) + ε. Economically interpret the coefficients. Does it make any difference if you include N_bathrm instead of N_bedrm? Explain either way. 9 Based on all the models above, which of the explanatory variables has more impact on the price of the property in the area? Why? 10 If you have successfully carried out the first 7 models and the time permits you to do one more model, you may consider this one. So far we have not considered the effect of floor variable on the price for different floors. In other words, we have assumed that the price that the buyers are willing to pay for the apartment is similar for all floors. Empirical studies have shown that buyers pay higher premium on top floors compare to lower floors. In order to be prudent in our estimation, you need to construct dummy variables for the lowest, highest floor or middle floors (i.e. 1, 2, 3, floor). Be careful on what is called “dummy variable trap”. Estimate: P = f (Area, Floor, Location, N_bathrm, middlef, highestf) + ε Good Luck!