In the file matching process described in section 6.3 above, we prepared a data file which matched the units from the original data file with appropriate scores on the basis of their treatment as a revised and autorecoded variable. To match these scores with the index file, we simply need to match the data files on a one-to-one basis after removing duplicate occupations from the original data file. In addition, it is useful to include information on whether or not a unit was represented in the original construction, and how many cases in total, before and after excluding pseudo-diagonal combinations, represented it.
Additionally, it is of considerable interest to know the population wide average CAMSIS derived scores for various population subgroups. This serves a purpose in the immediately next stage - in the imputation of scores to non-represented cases - and also in the production of the final index file, where such averages can be used to map scores to cases with incomplete occupational information. We thus also calculate variables which show the various subgroup averages within each scale. Illustrative SPSS syntax for achieving all of these requirements are shown in the accompanying file.
Last modified 14 February 2002
This document is maintained by Paul Lambert (paul.lambert@stirling.ac.uk)