Module 4: INTERPRETING CONFIGURATIONS
|
|
|
This module might better be entitled “Searching for Structure”, as the main focus is upon detecting structural features in a configuration. That done, the question of “What does it mean?” arises, and here we enter methodological hot waters. For some, it is the researcher to decide what it “means”; for others of a cognitive persuasion, it is more a matter of using what the respondents say or mean when they generate the data, and using this as a resource to try to interpret it. In either event, the key concepts are:
- Internal vs External interpretation (using the same data as for the scaling, or using material from outwith the input data to assist interpretation)
- Simple structures (especially graphical investigation of lines/vectors, series, non-linear structures such as horseshoes)
- Differential density and clustering (We shall concentrate on Hierarchical Clustering, but there are other ways)
- “Cleaning up” by seeing the effect of removing ill-fitting point/s, and removing liaison point/s to reveal clustering
- Matching (using essentially verbal material as a resource to interpret the structures found)
The material in this module is covered primarily in TUG, chapter 4 (which read now ), and the main resources (apart from paper and pencil!) are:
- PERMAP’s diagnostic point and link Map;
- in NewMDSX, the programs PROFIT and PERMAP IV (linear property fitting), PERMAP III (ideal point property fitting) and HICLUS (also in SPSS) for clustering.
You should run these programs with test data to become familiar with what they do, ( and even become adventurous and use your own data!) and finally ...
- use Simple PINDIS to “put Humpty Dumpty together again” if you have several configurations you want to compare.
All files are in Acrobat PDF format: Double click on the file title to read on line, or use your right mouse button to "Save target as" to save the file to your disk.