The first step is to select a data type and time period so that you can see
what the actual migration was and what statistics are available to build
your model with. First, find the Use data by list and select
either County, to use statistics gathered for each county in
the United States, or State, to use statistics gathered for
each state. Then, find the Time Period list, and select one
of the available time periods. The statistics database for that time
period will be downloaded, along with a map of the actual migration
during that time period, labelled Actual Migration
A number of statistics have been gathered for each region over the course of
your selected time period. To build your model of migration, you will need to
select which ones you believe were important factors in the migration of that
time period. To get a list of variables from which to choose, press
Select Variables.
To select one of the variables from the list, simply click on it. This will
deselect all your other variables, however. To select, or deselect, a variable
without changing the status of all the other variables, hold down the
Control key as you click the variable. To select an entire range
of variables, click on the beginning of the range, then hold the
Shift key, and click on the end of the range. When you have
selected all the variables you want, press the OK button.
The variables you select will appear in the Variables list on the
left of the screen. Underneath the name of the variable, there is a 'slider'
control, labelled Predicted Best Value. This control allows you to
set a value that you think was most attractive to migrants during that decade.
The slider starts at the mean value of that variable. The value of the current
setting appears above the slider control area.
To change the setting, simply drag the the thumb piece of the slider left or right. Alternatively, you can click on the track at the position you want to move the thumb piece to. In either case, the updated value of the variable will be displayed above the slider.
If you have selected more than one variable, a slider control labeled
Importance will appear to the right of the variable name. This
control lets you adjust the relative importance of each variable in your
model. You may make a variable less important by dragging the slider to the
left or more important by dragging it to the right.
To see the migration based on your model, press the Run Model
button. After some calculation, a map will appear in the lower right with the
migration that would have occurred during that decade if your model was what
actually affected migration.
At the top of the Modelled Migration map, a value labeled
Match appears. This number indicates the correlation of the
migration between the Actual Migration and the
Modelled Migration maps. The Match number ranges
from -1 to 1, where -1 indicates that
the maps are completely opposite and 1 indicates a perfect match.
Generally, a match is considered good if it is greater than 0.3.
Look at the differences between the Actual Migration and the
Modelled Migration maps. Think about likely causes for those
differences. Then adjust the Predicted Best Value and
Importance sliders. Run the model and readjust the sliders until
you get as good a match as possible. Also, try other variables that you think
may have an impact.
You can also compare a model between different time periods. After a model is built, you can select another time period in which to run your model.
Prev: Getting Started Next: Control Descriptions