DIAGONAL Subcommand (FACTOR command)

DIAGONAL specifies values for the diagonal in conjunction with principal axis factoring.

• If DIAGONAL is omitted or included without specifications, FACTOR uses the default method for specifying the diagonal.
• DIAGONAL is ignored with extraction methods other than PAF. The values are automatically adjusted by corresponding variances if METHOD=COVARIANCE.
• If more than one DIAGONAL subcommand is specified, only the last one is in effect.
• If any ANALYSIS subcommand is explicitly specified, DIAGONAL subcommands specified before the last ANALYSIS subcommand are ignored. If no DIAGONAL is specified after the last ANALYSIS subcommand, the default is used.
• Default communality estimates for PAF are squared multiple correlations. If these cannot be computed, the maximum absolute correlation between the variable and any other variable in the analysis is used.

valuelist. Diagonal values. The number of values supplied must equal the number of variables in the analysis block. Use the notation n* before a value to indicate that the value is repeated n times.

DEFAULT. Initial communality estimates.

Example

FACTOR VARIABLES=V1 TO V12
  /DIAGONAL=.56 .55 .74 2*.56 .70 3*.65 .76 .64 .63
  /EXTRACTION=PAF
  /ROTATION=VARIMAX.

• The factor analysis includes all variables between and including V1 and V12 in the active dataset.
• DIAGONAL specifies 12 values to use as initial estimates of communalities in principal axis factoring.
• The factor pattern matrix is rotated using varimax rotation.