CONTENTS Subcommand (MATRIX DATA command)

CONTENTS defines the record types when ROWTYPE_ is not included in the data. The minimum specification is a single keyword indicating a type of record. The default is CORR.

  • CONTENTS is required to define record types and record order whenever ROWTYPE_ is not specified on VARIABLES and its values are not in the data. The only exception to this rule is the rare situation in which all data values represent pooled correlation records and there are no factors. In that case, MATRIX DATA reads the data values and assigns the default ROWTYPE_ of CORR to all records.
  • The order in which keywords are specified on CONTENTS must correspond to the order in which records appear in the data. If the keywords on CONTENTS are in the wrong order, MATRIX DATA will incorrectly assign values.

CORR. Matrix of correlation coefficients. This is the default. If ROWTYPE_ is not specified on the VARIABLES subcommand and you omit the CONTENTS subcommand, MATRIX DATA assigns the ROWTYPE_ value CORR to all matrix rows.

COV. Matrix of covariance coefficients.

MAT. Generic square matrix.

MSE. Vector of mean squared errors.

DFE. Vector of degrees of freedom.

MEAN. Vector of means.

STDDEV. Vector of standard deviations. SD is a synonym for STDDEV. MATRIX DATA assigns the ROWTYPE_ value STDDEV to the record if either STDDEV or SD is specified.

N_VECTOR. Vector of counts. N is a synonym for N_VECTOR. MATRIX DATA assigns the ROWTYPE_ value N to the record.

N_SCALAR. Count. Scalars are a shorthand mechanism for representing vectors in which all elements have the same value, such as when a vector of N’s is calculated using listwise deletion of missing values. Enter N_SCALAR as the ROWTYPE_ value in the data and then the N_SCALAR value for the first continuous variable only. MATRIX DATA assigns the ROWTYPE_ value N to the record and copies the specified N_SCALAR value across all of the continuous variables.

N_MATRIx. Square matrix of counts. Enter N_MATRIX as the ROWTYPE_ value for each row of counts in the data. MATRIX DATA assigns the ROWTYPE_ value N to each of those rows.

COUNT. Count vector accepted by procedure DISCRIMINANT. This contains unweighted N’s.

PROx. Matrix produced by PROXIMITIES. Any proximity matrix can be used with PROXIMITIES or CLUSTER. A value label of SIMILARITY or DISSIMILARITY should be specified for PROX by using the VALUE LABELS command after END DATA.

Example

MATRIX DATA VARIABLES=V1 TO V3 /CONTENTS=MEAN SD N_SCALAR CORR.
BEGIN DATA
  5  4  3
  3  2  1
  9
  1
 .6  1
 .7 .8  1
END DATA.
LIST.
  • ROWTYPE_ is not specified on VARIABLES, and ROWTYPE_ values are not in the data. CONTENTS is therefore required to identify record types.
  • CONTENTS indicates that the matrix records are in the following order: mean, standard deviation, N, and correlation coefficients.
  • The N_SCALAR value is entered for the first continuous variable only.

Example

MATRIX DATA VARIABLES=V1 TO V3 /CONTENTS=PROX.
BEGIN DATA

data records

END DATA.
VALUE LABELS ROWTYPE_ 'PROX' 'DISSIMILARITY'.
  • CONTENTS specifies PROX to read a raw matrix and create a matrix data file in the same format as one produced by procedure PROXIMITIES. PROX is assigned the value label DISSIMILARITY.