/TABULATE <data> BY <var1>[,<var2>,...] TO <list> / K.Vehkalahti
makes various types of cross-tabulations of <data> according to selected
variables and saves the results in <list>. All variables can be either
numeric or string variables. /TABULATE employs the standard TAB operation
of Survo automatically in different ways.
In addition to the TAB operation, /TABULATE produces tables of expected
frequencies and contributions to chi-square. Also the tables of means and
standard deviations have a slightly different form.
The grouping variables are given as parameters. Other variables are
activated with mask characters, see below. They are tabulated with the
grouping variables one at a time. At least one grouping variable must be
given. It will be the column variable of all tables. Up to 7 grouping
variables may be used, thus forming maximum of 8-dimensional tables. The
second grouping variable (if given) will be the first level row variable,
the third will be the second level row variable and so on.
The classifications of the grouping variables must be given following the
syntax of TAB (see TABLE?). The main idea of /TABULATE is that the classi-
fications of the activated variables are found automatically. On the other
hand, any number of classifications may be readily given, to ensure proper
number of classes, or to give names for the classes, for example.
The automatic classification is based on the information given by STAT,
which is run once for all activated variables. Maximum number of classes is
given by the CLASSMAX=<#_of_classes> specification (default CLASSMAX=30).
IND, CASES and SELECT specifications can not be used and the <data> can not
be a data table in the edit field. Use FILE COPY, for example, to create a
data file, possibly with selected observations.
The variables are activated by FILE ACTIVATE using any of the mask columns.
The proper mask column is pointed out by MASK=#<col>, the default being
MASK=#1. VARS and the usual MASK specifications are not available.
The sums of columns and rows are computed for each table.
The resulting <list> can be browsed by LIST SHOW <list> and printed e.g. by
*PRINT CUR+1,CUR+2
- [Courier(10)][line_spacing(11)][margin(300)]
- list <list>.LST
By default the sizes of the edit fields correspond to REDIM 350,150,50.
This is changed by the specification
LISTDIM=<#_lines>,<#_columns>,<max_#_shadow_lines>.
In summary, the results are selected as follows:
Mask Task
f Frequencies (one table)
r Frequencies + percentages in rows (two tables)
c Frequencies + percentages in columns (two tables)
t Frequencies + percentages from the grand total (two tables)
e Frequencies + expected frequencies (two tables)
x Frequencies + contributions to chi-square (two tables)
m Means, standard deviations and frequencies (one table)
F (=f+r+c+t) (four tables)
X (=f+e+x) (three tables)
T = More information on TAB operations
Help System (web edition)