|INTRODUCTION||BIBLIOGRAPHIC REVIEW||SCALE CONSTRUCTION|
CAMSIS measures : Compatibility with the datasets of the CHER study
After an 'IRISS' visit to CEPS in September 2002, we have produced working CAMSIS versions appropriate to 11 countries from the CHER datset which have relevant occupational information.
- Introduction and description of the CAMSIS-CHER files
- Downloadable resources for CHER
- Other comments
- Contact and citation
The CHER (Consortium of Household Panels for European Socio-Economic Research) datasets are a collection of annual household panel surveys covering 17 European countries (with the possible later addition of data from Sweden, the US and Canada). The files have been harmonised on major variables by staff at the CEPS research institute, producing a data file structure where harmonised records cross different years and countries. At time of writing production of the CHER dataset is near completion and scheduled release of a CD with the data is in January 2003.
In addition, it is possible to access the CHER data by visiting the CEPS institute in Luxembourg and working on the data there without retaining it. It was in this capacity that four members of the CAMSIS project (Ken Prandy and Paul Lambert from the UK, Erik Bihagen from Sweden and Marge Unt from Estonia) visited CEPS under the 'IRISS' scheme in September 2002, where one of our undertakings was to produce data files linking CAMSIS scores with the CHER datasets. Here we describe how other users can make use of these scores by accessing the data and documentation files downloadable below.
The large majority of the documentation on the CHER dataset is available through the CHER webpages at CEPS (after following the appropriate links). This includes a very useful interface for browsing harmonised variable names and their categories, the 'CHER viewer'. The main CHER file structure involves year-by-year datasets where individual or household records can be matched by unique identifiers with the appropriate records from other years. For the main panel records, in each year the data from multiple countries is combined onto a single year-specific file, with the 'country' variable indicating nation. However users should note that a number of further datasets contain other related information, perhaps the most significant being the 'meta' data files which are not year specific but carry time constant data on the CHER respondents (including core variables such as gender). To combine information from such multiple files users need to match records based upon key unit identifying variables (documented on the CHER webpages).
Because of concerns over data confidentiality and 'disclosure risk', however, some data from the CHER files which could potentially allow users to identify specific individuals has been suppressed. This includes the occupational title data which, unfortunately to our minds, is released only at the level of ISCO-88 submajor (2-digit) group (up to 27 categories). (Though in one exception, the German occupational data is available in ISCO-68 3-digit minor group units). These truncations mean that it is not possible to associate full CAMSIS versions with the CHER datasets, though we can still calculate scores at the level of the sub-major groups. We chose to proceed with this strategy, and to our minds the scoring of the 27 units incorporates enough linear gradation for these 'working approximation' of CAMSIS scales to be effective variable instruments.
In addition, not all countries within the CHER project have occupational data available for them, and of those that do, we do not have completed CAMSIS versions for all of them. During our visit at CEPS, for the 4 CHER countries for which we did have final CAMSIS versions (Germany, Switzerland, the UK, and Ireland) we calculated weighted subgroup means which could be associated with ISCO-88 sub-major groups (see relevant notes from the versions page of the CAMSIS website). Then, a further 7 CHER countries had occupational information, and for those examples (Belgium, Hungary, Luxembourg, Poland, Denmark, France and Portugal) we derived our own CAMSIS versions, using the Husband-Wife data of the CHER project itself, at the level of ISCO-88 sub-major groups only. A fuller report on these version derivations, with subsequent assessments of their meanings, is downloadable below.
Although use of sub-major group level working approximations to CAMSIS scores ultimately goes against our general principles of using the maximum level of occupational differentiation, we should admit that it simplifies the use of the subsequent CAMSIS index files! During our visit to CEPS we produced a basic cross-country and cross-gender index file, downloadable below, which contains appropriate CAMSIS scores by country and gender for each CHER study for which occupational information was available. Users of the cross-country CHER files then simply have to sort their CHER datasets by country, gender and occupation, then match in with our index file. Example SPSS syntax to achieve this is included in the 'readme' file below; we may also add further examples to this webpage at a future date.
The only notable complication to this method concerns the German occupational data, which alone of the CHER datasets is stored in 3-digit ISCO68 units rather than 2-digit ISCO88. For compatibility, we have suggested first translating from ISCO68 to ISCO88 3-digit units (using the macros-provided from Harry Ganzeboom's ISMF site, see also our own 'occupational units' page). The downloadable data files below are then indexed, solely for the case of Germany, to 3-digit ISCO88 units (and 2-digit for all the other countries). Most of the length of the example SPSS syntax in the readme file below is solely concerned with utilising this translation macro.
In the table below, then, users can access the CAMSIS-CHER index file, the 'readme' documentation associated with it in Word format, and [planned Oct02] the longer report describing the CAMSIS-CHER files. These files may also be included within the release of the CHER CD in January 2002.
Downloadable resources for CHER
camsis_cher.sav SPSS index data file (6 variables) camsis_cher.dat Plain text index data file (6 variables) camsis_cher_readme.doc 'Readme' documentation for index file [expected Oct02] Report describing CAMSIS CHER properties
Contact and citations
We'd be interested in hearing from anyone who uses the CAMSIS scores on the CHER datasets, please contact us with any news. Paul Lambert had the most involvement in preparing the CAMSIS CHER files, so he is the suggested contact point for any queries.
Users wishing to cite this page or any other files derivative from the CAMSIS project webpages can use:
Prandy, K. and Lambert, P.S. (2002) "CAMSIS Project Webpages" http://www.camsis.stir.ac.uk : Cardiff School of Social Sciences, Cardiff University (accurate at [insert date]).
At time of writing (Oct 02), the most appropriate published summary reference for the CAMSIS project and its derivative scales is:
Prandy, K. and Jones, F.L. (2001) "An International Comparative Analysis of Marriage Patterns and Social Stratification." International Journal of Sociology and Social Policy 21: 165-83.
Stirling University Home Page