Help System (web edition)

VAR <y_variable>=#<function>(<expression>) TO <data> 
makes various special transformations of <expression> to <y_variable>.
In most cases <expression> is simply a variable.
The following #<functions> are available:
      #RANK   (ranks of values, tied observations with average rank)
      #NRANK  (ranks of values, tied observations with lowest rank)
      #NORMAL (normalized values with original mean and variance)
      #STD    (standardized values with mean=0, std.dev=1)
      #CENTER (centered values with mean=0)
      #PROPORTION (values divided by their total sum)
      #PERCENT    (100 times values divided by their total sum)
      #TRUNCP (outliers specified by P are replaced by missing values)
      #WINSP  (outliers specified by P are replaced by border values)
      #TRUNCL (outliers specified by LEVEL are replaced by missing values)
      #WINSL  (outliers specified by LEVEL are replaced by border values)

In #TRUNCP and #WINSP 100(1-P)% of the observations are considered
outliers. P is given by the specification P=<value> and default is 0.95.
In #TRUNCL and #WINSL an observation X is an outlier if
abs(X-mean) > level*stddev.
Level is given by the specification LEVEL=<level> and default is 1.96.

For example,
VAR Y=#PERCENT(X) TO DATA1 / gives Y values as 100*X/Sum_of_X's
VAR Y=#NORMAL(X) TO DATA1  / computes normalized X values as variable Y.

  V = More information on VAR operation