Here is an example of input for the SPSS procedure found under analyze:scale: multidimensional scaling.
The input are pairs (i,j) of times of one subject to correctly respond to semantic category j immediately after identifying category i. The matrix is asymmetric and contains dissimilarities (the default) as higher RTs suggest more dissimilarity between a pair of stimuli.
stimulus |
a |
b |
c |
d |
e |
f |
||||||
a |
48 |
56 |
23 |
123 |
45 |
460 |
||||||
b |
78 |
84 |
79 |
210 |
39 |
25 |
||||||
c |
80 |
237 |
200 |
145 |
178 |
127 |
||||||
d |
78 |
160 |
150 |
478 |
256 |
290 |
||||||
e |
65 |
250 |
67 |
312 |
4 |
345 |
||||||
f |
410 |
340 |
312 |
324 |
367 |
367 |
||||||
The code below, which may also be run from the menu, runs the multidimensional scaling procedure specifying a 6x6 asymmetric matrix of user input for six stimuli and the ordinal level of measurement.
ALSCAL VARIABLES= a b c d e f /SHAPE=ASYMMETRIC /LEVEL=ORDINAL /CONDITION=MATRIX /MODEL=EUCLID /CRITERIA=CONVERGE(.001) STRESSMIN(.005) ITER(30) CUTOFF(0) DIMENS(2,2) .