MDGROUP Subcommand (MRSETS command)
/MDGROUP NAME= setname {LABEL= 'label' }
{LABELSOURCE=VARLABEL}
CATEGORYLABELS={VARLABELS }
{COUNTEDVALUES}
VARIABLES= varlist
VALUE= {value }
{'chars'}
The MDGROUP
subcommand defines or modifies a multiple dichotomy set. A name,
variable list, and value must be specified. Optionally, you can control
assignment of set and category labels.
NAME. The name of the multiple dichotomy set. The name must follow IBM® SPSS® Statistics variable naming conventions and begin with a $. If the name refers to an existing set, the set definition is overwritten.
LABEL. The label
for the set. The label must be quoted and cannot be wider
than the limit for variable labels. By default, the set is unlabeled. LABEL
and LABELSOURCE
are mutually exclusive.
LABELSOURCE. Use the variable
label for the first variable in the set with a defined variable label
as the set label. If none of the variables in the set
have defined variable labels, the name of the first variable in the
set is used as the set label. LABELSOURCE
is an alternative to LABEL
an
is only available with CATEGORYLABELS=COUNTEDVALUES
.
CATEGORYLABELS
= [VARLABELS | COUNTEDVALUES]. Use variable
labels or value labels of the counted values as category labels for
the set.
VARLABELS
uses the defined variable labels (or variable names for variables
without defined variable labels) as the set category labels. This
is the default. COUNTEDVALUES
uses the defined value labels of the counted values as the set category
labels. The counted value for each variable must have a defined value
label and the labels must be unique (the value label for the counted
value must be different for each variable).
VARIABLES. The list of elementary variables that define the set. Variables must be of the same type (numeric or string). At least two variables must be specified.
VALUE. The value that indicates presence of a response. This is also referred to as the “counted” value. If the set type is numeric, the counted value must be an integer. If the set type is string, the counted value, after trimming trailing blanks, cannot be wider than the narrowest elementary variable.
Elementary variables need not have variable labels, but because variable labels are used as value labels for categories of the MD variable, a warning is issued if two or more variables of an MD set have the same variable label. A warning is also issued if two or more elementary variables use different labels for the counted value—for example, if it is labeled Yes for Q1 and No for Q2. When checking for label conflicts, case is ignored.