You are predicting the cost of a house (in thousands of dollars) in Boulder based on the lot size
in acres, the size of the house (in square feet) and the location of the house (city, suburbs, or
rural). The cost of the house is log-transformed to satisfy normality requirements. The size of
the lot and the size of the house have been centered around their respective means, which are 10
and 2000.
1. Write out your linear regression model with meaningful names for each of the
parameters.
2. You analyze your data set, and get this summary:
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 415.260000
0.195028 2129.230 < 2e-16
lotsize
0.410955
0.104446
3.935 0.00017
housesize
0.198820
0.002278
87.289 < 2e-16
suburb
0.374750
0.274860
1.363 0.17635
rural
-0.548615
0.277987
-1.974 0.05168 .
--Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’
0.1 ‘ ’ 1
Residual standard error: 1.059 on 85 degrees of
freedom
Multiple R-squared: 0.9891, Adjusted R-squared:
0.9886
F-statistic: 1931 on 4 and 85 DF, p-value: < 2.2e-16
3.
4.
5.
6.
What is your baseline group?
Write out your linear model based on this summary.
Interpret each of your covariate effects in terms of the original cost of the house.
Is the location of the house a significant predictor of the cost of a house? Which
locations are significantly different?