Microsoft Office Tutorials and References
In Depth Information
Multi-Regression
FIGURE 14.5 Regression Results
We can use this equation to calculate predicted price based on Size and the
number of Bathrooms. If, for example, an apartment has 1,000 square feet and 1
bathroom we can plug it in the equations and predict:
¼
:
þ
:
;
þ
:
Price
64
175
0
4544
×
1
000
60
65
×
1
2. It is important to note the p-values in cells E18 and E19. The p-values have to be
below 0.05 (5 percent) for each one of the independent variable coefficients. (Refer
to your statistics notes or book for detailed explanation. It is the
in the t test.)
3. R 2 has improved tremendously over the Simple Regression model. Now it is
0.8043 compared to the previous 0.4236 calculated on the first model. See cell B5.
4. The Standard Error in Cell B7 is 21.77. We can say that 95 percent of the time
our forecast will be at the predicted/calculated value
α
1.96 Standard Errors.
This means that we will be 95 percent of the time off by
þ
×
21.77). [1.96 is a key Z value in Statistics that yields the 95 percent Confidence
Interval. Again, refer to your Statistics notes.] The above example calculated
\$450,875 for the apartment, which is in the range of \$450,875
þ
42.67 (
¼
1.96
þ
\$42,670 with
a 95 percent Confidence Level.
Using the above model, we can say that every square foot adds on the average
0.4544 or \$454.40 to the price. Every bathroom adds approximately 60.65 or
\$60,650 to the price.
Search JabSto ::

Custom Search