Help System (web edition)

BURT TABLE <multiway_table>   or   BURT DATA <data>     / K.Vehkalahti 1998
creates a Burt's contingency table, which is a symmetric matrix
consisting of cross-tabulations of two or more classifiers. It can be
used as an input for multiple correspondence analysis, for example.
There are two different operations, depending on the type of input:

BURT TABLE <multiway_table> 
    creates a Burt's table from the given multiway frequency table.
    Multiway tables are usually created by TAB operations (see: TABT?).

BURT DATA <data> 
    creates a Burt's table from the active part of <data>.
    The classifications of numeric variables may be given following the
    syntax of TAB operation. If no classification is given, all distinct
    values of a variable are used. The maximum number of classes is by
    default 30. It can be changed by CLASSMAX specification, if necessary.

By default the resulting Burt's table is saved as a matrix file BURT.M and
a vector of the number of classes per classifier as CLASS.M . Alternative
names can be chosen by specifications BURT and CLASS.
By specification Z=<name> a binary matrix of all classifier combinations
for each observation is saved. This matrix is not saved by default, as it
can easily be quite large. However, it could be useful in some situations,
as it represents a 'dummy-variable' form of the original data.

These operations are best used by the sucros /BURT-TABLE and /BURT-DATA .

For more information on multiple correspondence analysis, see MCORRESP? .

References: Burt, C. (1950). The Factorial Analysis of Qualitative Data.
              Journal of Statistical Psychology, Vol. 3, No. 3, pp. 166-185.
            Lebart, L., Morineau, A., & Warwick, K.M. (1984).
              Multivariate Descriptive Statistical Analysis. Wiley.
            Gower, J.C., & Hand, D.J. (1996). Biplots. Chapman & Hall.

  T = More information on TAB operations