You are interested in running a “hedonic pricing model” to examine the determina

Question

You are interested in running a “hedonic pricing model” to examine the determinants of housing prices. This just means that price is on the left-hand side.

A?Run a regression of housing price on the number of bedrooms and the total size of the house (sqrft). How do we interpret the constant and the slope coefficients in this regression? (DO NOT answer!!!)

C. How does restricting the regression line to pass through the origin affect your estimates from part A? Do you believe the model as specified in part A or part C makes more sense in this context? Explain.

Hint: you can type “help reg” to look up the options available when running a regression

D. If sqrft were omitted from the regression part A (with a constant), would that produce any sort of bias in your estimation of bdrms? Explain why or why not.

E. How much additional variation in price is explained when both bdrms and sqrft are included in your regression as opposed to simply bdrms? Calculate this using your model from part A.

F. Suppose you think a log model is more appropriate in this case. Run a regression of the log of housing price on the total number of bedrooms and the log of the total area of the house. How do we interpret the slope coefficients in this regression?

Answer question C,D,E,F only

SS MS Number of obs 88 Source df F (2, 85) 72.96 2 290004.576 Prob F Model 580009.152 0.0000 Residual 337845.354 85 3974 65122 squared 0.6319 Adj R squared 0.6233 Total 917854.506 87 10550.0518 Root MSE 63.045 Coef Std Err. P It 1953 Conf. Interval] 15.19819 9.483517 bdrms 1.60 0.113 3.657582 34.05396 0.000 1009495 Sqrft 1284362 0138245 9.29 1559229 -19.315 31.04662 0.62 0.536 81.04399 42.414 Con

Explanation / Answer

C.

Considering 0 intercept will not change the coefficient estimates of other independent variables. This approach will be much better due to the fact that when there is no size & no bedroom in the house, there should be no (or zero) price. (Negative price is not making any sense)

D.

If sqrft is ommitted from equation then only bdrms will remain in regression equation. Considering the fact that sqrft and bdrms are correlated, it would cause collinearity issue in regression equation. So it is recommended to remove one of them. No biasness will be there. (Check VIF for both the variables and remove one of them).

E.

Full information for this is not given. However, just run regression considering bdrms only find R-sq and then include sqrft also in equation and find R-sq. Find the diff between these 2 R-sq. That's the ans.

F.

It is difficcult to interpret. Still:

Unit increase in no. of bedroom increases log of price by 1 unit

Unit increase in log of sqrft increases log of price by 1 unit

Related Questions

Navigate

• Browse (All)
• Browse Y
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.