7.2.1 Generating the file: details
The first step is the creation of a database which will serve as the start
of our index file. This requires one case for every possible occupational unit,
with the occupational unit numeric value assigned to each case. Later, each case
will then have scores matched to it, separate variables indicating the scores
applicable to men and women, and those applicable to title-by-status or title-only
base units. Additionally, we may also choose to add other variables which indicate,
for example, the number of cases which represented the given occupational base
unit during the scale construction; if and how the occupational base unit was
merged with other base units during the CAMSIS construction; or the proportion
of cases treated as pseudo-diagonals.
The first variables of the index
file can be created by listing the full set of occupational title numeric values
(perhaps from the same source as the list of title value labels), then stacking
successive full columns of these lists parallel to a single digit indicator of
occupational status, then finally deriving a third column which is equal to the
title*10 + status values. This can be readily achieved in Microsoft Excel, in
a way similar to that described in section
2.3 on adding occupational labels. Such a plain text file can subsequently
be read into SPSS, value and variable labels added, and occupational subgroups
calculated, in each case repeating methods discussed previously. The accompanying
file provides sample SPSS syntax illustrating
how these results are achieved.