6.3 Descriptive statistics and graphical displays: details
If we assume, from the sections above, that we now have an output table, for instance in Excel or a plain text file, which shows the occupational unit numeric values adjacent to columns indicating the number of cases each unit is represented by, and columns indicating the derived scores for men and women on the different versions, we can immediately derive descriptive statistics and graphical displays to summarise the estimated score distributions. The shape of these distributions is significant, because it could indicate whether the distribution of social interaction-derived locations is apparently a smooth hierarchy or more akin to a disjointed categorical structure within the first dimension estimated. The accompanying file gives sample SPSS syntax which moves from reading the original data files through to performing such relevant analyses. It assumes that we have two plain text data files (perhaps pasted vial Excel) which refer to the title-only and title-by-status base units, each having successive columns which indicate (1) the numeric value of the base unit, then (2) columns indicating the total number of men then women representing the unit, then (3) the total number of men then women representing the unit after all pseudo-diagonals were excluded, then (4) finally, columns representing the derived scores for men then women. We also assume that, as a by-product of the data setup processes, we have created a dataset, latestdata.sav, which is a table data file of husband wife combinations that contains the original occupational codes plus all the successive data revisions and appropriate 'square autorecoded' unit values Additional variables, - eg psdm1, psdm2 - on that file will indicate whether a particular combination was treated as a pseudo-diagonal or not.
Return to Construction guide
Last modified 14 February
This document is maintained by Paul Lambert (email@example.com)