B: SORTPAC-3: A program for analysis of Sortings data
An MS-DOS program (in C++) for performing analyses of sortings data elaborated in SDCA , including:
Input of data in:- "preferred data format" (p18) each sorting is a row vector where the jth entry is the group/category number in which object j is located. Groups are sequentially but arbitrarily numbered), or
- "Anthropac" format (each sorting is input as a set of delimited groups, each consisting of the object number of constituent objects)
Conversion of each sorting into a co-occurrence matrix:
- (0,1) for simple co-occurrence (M1: Burton's F, p.46)
- weighting by category size (M2: Burton's F', p.47)
- weighting by reciprocal of category size (M3; Burton's G, p.46)
- information theoretic measure (M4:Burton's Z, p.48)
Aggregation (sum) of co-occurrence matrices
- Operations on aggregate matrix (e.g. add constant; scale by divisor; invert into dissimilarity measure)
Output:
- Conversion of an aggregated matrix into a complete Syntax or RunScript file for analysis by NewMDSX (MINISSA, HICLUS) or SPSS (Proxscal)
SORTPAC-3 is fully documented in: Sortpac v3.0 (Coxon & Coxon)