Introduction Harmonised Dataset

CAMSIS homepage

Long-run changes in the significance of social stratification in Britain

This project was funded by the UK ESRC between 2002 and 2005. We examined a large number of survey datasets, spanning 1963 to 2005 in Britain, in order to establish the nature of changes in the significance of social stratification. Our analyses concentrate on social mobility, particularly the role of education, and political attitudes and behaviour.

Introduction Main features of the project
Papers Papers and references
Dataset description Details on the components of the dataset used
Contact and citation

  Introduction to the project

Are those structures of social stratification or inequality that we observe broadly steady over time, or have they been changing according to a discernible pattern? The dominant paradigm in recent British research has been that of constancy, for example, the claim that the relative social mobility chances of children from different age cohorts have remained roughly the same.

However much of the research which leads to this position has focussed upon a relatively short historical timespan and there are good reasons to believe that changes to stratification structures, when they do occur, happen very slowly and gently. Thus, it is quite plausible that those studies which find no firm evidence of time trends in Britain, have simply not been able to analyse a sufficiently long period to allow them to distinguish genuine change from sampling fluctuation.

The main contribution of this project is to expand substantially the range of datasets in order to construct a pooled sample which spans a longer time period and involves many more cases. In this project we concentrate on British datasets from the last 40 years, which can be accessed via the UK Data Archive at the University of Essex. An initial phase of the research, therefore, is the construction and specification of such a dataset (further notes on this issue are below).

Preliminary work on this project produced an initial dataset and findings which clearly suggested a slight but steady rise in father-to-son occupational social mobility rates over the period considered, as summarised in the 2002 conference paper.

Extension of this investigation takes two main forms. One is the expansion of the harmonised dataset by including more cases and more detail on more variables, particularly the inclusion of data on individuals' educational levels in the analysis of the relationship between their own and their father's occupation. The second is a parallel analysis of the relation between social stratification and political attitudes and behaviour and class identification.

In analysing this dataset we make the assumption that we can use occupational data, both on individuals' own jobs and on their fathers', as indicators of social stratification position. Precisely how we formulate such an indicator is, of course, open to contention. Rather than adopting one or other of the 'social class' schemes, our preference is to use the 'CAMSIS' measure of occupational position, which assigns a scale value to each occupation. The value indicates the relative position of typical incumbents of occupations within the wider structure of social stratification, as it is revealed through patterns of social interaction. Extensive further details on the CAMSIS measures can be found in associated webpages. The key points for this investigation, are that the CAMSIS measures indicate relative locations within historically specific structures, and that they highlight the key hierarchical element of inequality which, we suggest, dominates social stratification patterns.

Our main tool of analysis is the construction of structural equation models which describe the interelations of stratification-related variables in the context of time period, age cohort, and other relevant factors. For some introductory resources on structural equation modelling, see for instance the StatSoft textbook.



The following papers relate to the work of this project (follow links to download):

P.S. Lambert, K. Prandy and W. Bottero (2007) 'By Slow Degrees: Two centuries of social reproduction and mobility in Britian', Sociological Review Online, 12(1).

[Earlier conference version: K. Prandy, P.S. Lambert and W. Bottero (2002) By Slow Degrees: Two centuries of social reproduction and mobility in Britain. Presented at ISA RC-28 On Social Stratification and Mobility, Nuffield College, Oxford, 11-13April]

K. Prandy, M. Unt and P.S. Lambert (2004) 'Not by Degrees: Education and social reproduction in twentieth century Britain', Paper presented to ISA RC28, Neuchatel, 7-9 May 2004 (draft paper pdf)

K. Prandy, P.S. Lambert and M. Unt (2003) Stability, Trended Change or Data Artefacts: Pooling forty years of social mobilty enquiries in the UK. Presented at ISA RC-28 On Social Stratification and Mobility, New York University, New York, 22-24August

A wider review of literature concerning the use of CAMSIS measures and related issues can be found in the bibliographic review section of the CAMSIS webpages.


Dataset description

Key to the project is the collation of data from multiple survey sources spanning the last 40 years in Britain, and their preparation in a harmonised analysis file suited to the meta-analyses described above. A table describing components of the dataset we use is available at this link.

There are several methodological issues which arise when attempting to analyse such a harmonised resource. The contributing datasets cover a sufficiently long time period for issues of historical comparability in measures of occupational and educational level to be raised - one reason for favouring the CAMSIS measures is their anticipated sensitivity to relative position given a particular historical location. Secondly, there is a chance that specific features of each component survey's collection, or data preparation, introduce spurious structures to the dataset, and it is thus necessary to consider methods of accounting for this. Finally, it is very important that we attempt to account for the historical context of each invididual record considered - both in terms of wider economic circumstances, and, for instance, in terms of the career context of the current position.


Contact and citations

Researchers involved in this project are :

Prof Ken Prandy, Cardiff School of Social Sciences, Cardiff University, UK

Dr Paul Lambert, Department of Applied Social Science, Stirling University, UK

Marge Unt, Tallin Pedagogical University, Estonia, and Cardiff School of Social Sciences, Cardiff University Uk (between Jul-Dec 2003).

Anybody wishing to cite this page or any other files derivative from the CAMSIS project webpages can use:

Prandy, K. and Lambert, P.S. (YEAR), CAMSIS Project Webpages, , University of Stirling (accurate at [insert date]).

An appropriate summary reference for this project is:

Prandy, K. and Lambert, P.S. (2002) R000223899 Long run changes in the significance of social stratification in Britain, Swindon, ESRC Society Today.



Last modified 25 May 2008.
This document is maintained by Paul Lambert (