
34 BUSINESS APPLICATIONS OF MULTIPLE REGRESSION
have a weak, positive correlation. This is supported by the resulting cor-
relation value of 0.5350.
Transit Ridership
Table 1.5 shows the largest urbanized areas by population, unlinked
passenger trips, and passenger miles for 2008. That dataset has three
variables:
1. Unlinked passenger trips in thousands
2. Passenger miles in thousands
3. Population from the 2000 Census
Figure 1.6 shows that the data had an outlier in the values for New York.
Its values in all three categories far exceed the values for any other transit
system. The single outlier does not affect correlation ...