LEVEL Keyword (CATREG command)
The following keywords are used to indicate the optimal scaling level:
SPORD. Spline ordinal
(monotonic). This is the default for a variable listed
without any optimal scaling level, for example, one without LEVEL
in the parentheses after it or with LEVEL
without a specification. Categories
are treated as ordered. The order of the categories of the observed
variable is preserved in the optimally scaled variable. Categories
will be on a straight line through the origin. The resulting transformation
is a smooth nondecreasing piecewise polynomial of the chosen degree.
The pieces are specified by the number and the placement of the interior
knots.
SPNOM. Spline nominal (non-monotonic). Categories are treated as unordered. Objects in the same category obtain the same quantification. Categories will be on a straight line through the origin. The resulting transformation is a smooth piecewise polynomial of the chosen degree. The pieces are specified by the number and the placement of the interior knots.
ORDI. Ordinal. Categories are treated as ordered. The order of the categories of
the observed variable is preserved in the optimally scaled variable.
Categories will be on a straight line through the origin. The resulting
transformation fits better than SPORD
transformation, but is less smooth.
NOMI. Nominal. Categories are treated as unordered. Objects in the same category
obtain the same quantification. Categories will be on a straight line
through the origin. The resulting transformation fits better than SPNOM
transformation, but is less smooth.
NUME. Numerical. Categories are treated as equally spaced (interval level). The order
of the categories and the differences between category numbers of
the observed variables are preserved in the optimally scaled variable.
Categories will be on a straight line through the origin. When all
variables are scaled at the numerical level, the CATREG
analysis is analogous to standard multiple regression
analysis.