Statistical Methods in Geography
Lab 5 – Chapter 7
Due November 28, 2005
1. Begin with the ‘country’ data set you used in Labs 3 and 4. Now, you’ll be using this
data set to come up with a multiple regression equation that best predicts life expectancy.
Your independent variables include the eight other variables you entered in last time.
Perform the following:
First try the kitchen sink method. You’ll use the same regression feature as last
time, except now you’ll be putting the ‘kitchen sink’ (all eight variables) in to
the equation. Calculate the parameters below. Interpret the equation and list
any outliers. If there is any multicollinearity, report which variables.
Regression equation (similar to last time)
R2 value
Collinearity statistics (VIF): tick ‘Collinearity’ in ‘Statistics’ box
Outliers via residuals: tick ‘Residuals: Standardized’ under ‘Save’
box
Repeat what you did above with the stepwise method. Under ‘Method’, change
to ‘Stepwise’. Report on the different steps it takes in adding and deleting
variables. Again, interpret the equation and list outliers. See if there is any
multicollinearity in this one.
Do a final regression. You may delete outliers if you wish, and eliminate from
consideration one or more variables that are collinear. (To delete an outlier, just
delete the value for that country’s life expectancy in the data sheet.) Report the
final equation and what you did. (Note there’s no single right answer here!)
2. Problem 2, page 150.
3. Problem 3, page 151.
4. Problem 4, page 151.
