Predicting home sales price

(a) Propose a straight-line model to relate the appraised property value to x to the sale price y
for residential properties in this neighborhood.
From the scatterplot, we can find the linear relation between 𝑥 and 𝑦. i.e.,
𝑦 = 𝛽0 + 𝛽1 𝑥1 + ε
(b) The R scatter plot of the data is shown above. Does it appear that a straight-line model
will be an appropriate fir to the data?
Yes, the linear pattern between 𝑥 and 𝑦 suggests that a straight-line model can be an
appropriate fit to the data.
(c) The R simple linear regression printout is also shown above. Find the equation of the
best-fitting line through the data on the printout.
𝑦̂𝑖 = 1.358 + 1.408𝑥𝑖
(d) Interpret the y-intercept of the least squares line. Does it have a practical meaning for this
application? Explain.

𝛽̂0 = 1.358, is the estimate of the sale price of a property when the market value is set at
𝑥 = $0. The y-intercept does not have a practical meaning in this home sales price
example because a zero market value of a property does not make sense.
(e) Interpret the slope of the least squares line. Over what range of 𝑥 is the interpretation
meaningful?
 As the appraised market value increases by one unit, the sale price is expected to
increase by 1.408 units.
 From the scatter plot, the property’s market value ranges from about $150K to
$1300K. So, it is meaningful to estimate the sales price of property whose market
value is between $150K and $1300K. However cannot be used to predict the
value of y when the value of x fall outside the range of the sample.
(f) Estimate the mean sale price of a property appraised at $300,000. [Note: Both sale price
and total market value are shown in thousands of dollar, i.e., 𝑥 = 300 means $300,000 of
the market value].
𝐸(𝑦|𝑥 = 300) = 𝛽0 + 𝛽1 𝑥𝑖 = 𝛽0 + 𝛽1 300, our estimate of 𝐸(𝑦) given 𝑥 = 300 is 𝛽0 +
𝛽1 300 = 1,358 + 1.408(300) = 423.758. The mean sale price of a property appraised
at $300,000 as $423,758.
(g) Use the R printout above to determine whether there is a linear relationship between
appraised property value 𝑥 and sale price 𝑦 for residential properties sold in this
neighborhood. That is, determine if there is sufficient evidence (at 𝛼 = 0.01) to indicate
that 𝛽1, the slope of the straight-line model, is not zero.
The p-value for the slope parameter 𝛽1 is 2 × 10−16 , much less than 𝛼 = 0.01.
Therefore, we reject 𝐻0 : 𝛽1 = 0 in favor of 𝐻1 : 𝛽1 ≠ 0.