MINI Rectangular Smallest Space Analysis : MINIRSA

performs a non-metric multidimensional unfolding analysis.

 

DATA: Two-way, one-mode  (row-conditional, rectangular)                         
TRANSFORMATION: Monotnonic

MODEL: Euclidean distance

Consider a set of subjects and a set of stimuli where the subjects indicate their preferences for the stimuli (the judgements need not be of preference; any asymmetric relation is acceptable). The aim of the program is to position both stimuli and subjects as points in a space of minimum dimensionality so that, for each subject, the rank order of the distances from his or her point of maximum preference in the space (the "ideal point") to the stimuli matches the subject's preference ordering as closely as possible.

MINIRSA expects data in a 'row-conditional' format. The data matrix consists of N rows each of which reflects a particular subject's order of preference for the stimuli. There are p columns. The various ways in which these may be presented are detailed below.

MINIRSA does not accept paired-comparisons data as such but will take the row sums of such matrices (see MDPREF).

Data may be input to MINIRSA as ranks, or as scores, in the order most-preferred to least-preferred, or the reverse.

MINIRSA also allows for missing data. The value to be regarded as indicating a missing value should be specified in the PARAMETERS command with the parameter MISSING : e.g. if 9 is the code for a missing datum then MISSING(9) is the appropriate instruction.

INPUT COMMANDS

Keyword                                                        Function
N OF STIMULI    [number]                            Number of stimuli in the analysis
N OF SUBJECTS [number]                            Number of subjects/cases
DIMENSIONS     [number]                            Dimensions for analysis
LABELS             [followed by a series            Optionally identify the stimuli
                        of labels (<= 65 chars          and subjects. The labels should
                        each on a separate line]       identify first the stimuli (columns),  
                                                                  then the subjects (rows), of the data
                                                                  matrix.

NOTES
1. N OF SUBJECTS may be replaced by N OF ROWS.
2. N OF STIMULI may be replaced by N OF COLUMNS

PARAMETERS
Keyword         Default Value      Function
DATA TYPE         1              1: Data are ranks (I-scales) of column
                                            indices in decreasing order of preference
                                         2: As 1 but in increasing order of
                                             preference.
                                         3: Data are scores in order of column
                                             indices - high score means high preference.
                                         4: As 3 but high scores mean
                                                 low preference.

MINIMUM ITERATIONS 6     Sets the minimum number of iterations to
                                        be performed before the convergence test.

MISSING DATA           0     sets the data value which is to be regarded
                                        as missing data.

MATFORM                   0      (Relevant only when 'READ CONFIG' is used.)
                                         0: The input configuration is entered
                                              subjects and stimuli (rows).
                                              by dimensions (columns).
                                              subjects are entered before
                                              stimuli.
                                         1: The input configuration is entered
                                              subjects and stimuli (columns)
                                              by dimensions (rows).

PRINT options (to main output file)
Option                   Form                   Description
INITIAL          N x r          Two matrices are produced: the
                     p x r          coordinates of the subject points;
                                      the stimulus points in the
                                      required dimensions.
FINAL             N x r          Similarly two solution matrices are printed:
                     p x r
DISTANCES   N x N          Three matrices are printed:  
                     p x p          1. the distances between the subject points;
                     N x p          2. the distances between the stimulus points; and
                                      3. the distances between the subjects and stimuli.
FITTING         N x p          The matrix of disparities (DHATs)
RESIDUALS    N x p          The matrix of residuals is printed
HISTORY                         This keyword generates an extremely detailed
                                      history of the iterative process. Users are
                                      warned that this option generates a
                                      large amount of output!
By default only the final configuration and the final STRESS values are output.

PLOT options (to main output file)
Option                      Description
SUBJECTS           A plot of the subject points only
STIMULI              A plot of the stimulus points only
JOINT                 The configuration of subject and stimulus
                          points is plotted.
SHEPARD           The Shepard diagram is produced.
STRESS              A plot of STRESS values by iteration
                         is produced.
POINT                The contribution of each subject to the
                         overall STRESS value is plotted.
RESIDUALS        A histogram of residual values is produced.

By default, a Shepard diagram and the joint space only are plotted.

PUNCH options (to secondary output file)
Option                     Description
SPSS                   A file suitable for input to SPSS is produced.
                           The following values appear:
                           I : the subject index no.
                           J : the stimulus index no.
                           IFR : no. of repeat orderings
                           n : the stimulus index no.
                           INPUT : the datum according to I,J
                           FITTING : the corresponding DHAT value
                           DIST : the solution distance between I & J
                           RESID : the corresponding residual value
                           The format of the file is (4I4, 3F10.4)
STRESS               The STRESS values at each iteration are
                           output in a fixed format.
FINAL                   A text file of the final configuration
                           is produced.

By default, no secondary output file is produced.

PROGRAM LIMITS
Maximum no. of subjects = 100
Maximum no. of stimuli = 100
Maximum no. of dimensions = 5

See also

  • The NewMDSX commands in full