By default, PLOT simply plots cases for the Y
variable against the X variable in two-dimensional space.
If users also want bivariate statistics printed out to
summarize the relationship between Y and X, then the
PLOT commanad must be followed by this subcommand:
/FORMAT =
REGRESSION
Somewhat confusingly, the PLOT command also
demands to have a separate PLOT subcommand for each plot
generated (or set of plots using similar variables). So
after the FORMAT= subcommand above, the first set of
variables plotted must be preceded by
PLOT = [varlist]
WITH [varlist]
where [varlist] stands for the variable(s) to
be plotted against the other variable(s). Additional
plots using different variables must be preceded with
another PLOT= subcommand.
PLOT also allows users to specify one separate
"identification" variable (let's call it Z) whose code
numbers are printed where each case is plotted. This
helps identify which cases lie above and below the
regression line.This feature is invoked with thise
command, following the second [varlist]
BY [varname]
where [varname] is the name of the identifying
variable.
To illustrate, the following command
/PLOTS = Y1 Y2 WITH X1
X2
will create four plots, Y1 with X1, Y1 with X2, Y2
with X1, and Y2 with X2.
Use semicolons to specify separate multiple plot
lists. Consider:
/PLOTS=CDBG89 WITH
POP86; POP86 WITH UCI80
The first list produces a plot of CDBG89 with POP86.
The second produces a plot of POP86 with UCI80.