Question #1
Among the following, which is NOT a part of a "generalized procedure" for
developing a multiple regression model?
•
Proceed with model estimation even if the model lacks a
theoretical basis.
•
Draw scatter diagrams to determine whether a linear model is
appropriate.
•
Gather data for the variables in the model
•
Proceed with model estimation using statistical software.
Question #2
This question is based on the ski resort example again, from
pages 8586
(slides 1823) of your course packet. This time, you would like to undertake
a model building exercise. If you need it again, the data is available
here
.
Since autocorrelation occurs over time, the inclusion of a time variable in
the model may correct the problem. So, include a new independent
variable 'Years.' You can do this simply by creating a column of serial
numbers going 1, 2, 3.
..till however many observations you have.
Then run the model again and answer the following questions.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document1.
The calculated value of the DurbinWatson test statistic is
1.884987482
2.
Then, based on the upper and lower critical values 1.54 and 1.10,
your decision should be:
There is no autocorelation, the
problem stands corrected.
3.
The model is significant as a whole.
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '09
 PETRY
 Regression Analysis, predicted profit margin

Click to edit the document details