[Other comments]Before downloading the tables, users are recommended to read:
Accessing and using CAMSIS scale scores
DOWNLOAD CAMSIS FILES US2000SOC00 version 1.0; author: Paul Lambert; released on: 12 July 2004 US90OCC90 version 1.0; author: Paul Lambert; released on: 22 January 2003 US90ISCO88 version 1.0; author: Paul Lambert; released on: 22 January 2003 US60OCC60 version 1.0; author: Paul Lambert; released on: 22 January 2003 'SONOCS' versions (2000) Versions constructed by Paul Lambert and Dave Griffiths as part of the 'SONOCS' project; released 23 May 2011, as described in Griffiths and Lambert 2011.
Data 2000 Census / 1990 Census / 1960 Census, 1% samples provided by IPUMS
Cases (see note) 2000: 700,000; 1990: 363,421; 1960: 335,287 Occupational classification 2000 Occupational Classification ;
1990 Occupational Classification ;
1990 scores also linked to ISCO-88 categories;
1960 Occupational Classification
Additional scales for the 2000 unit, microclass units, and occupation-by-education combinations are available from the US CAMSIS 'SONOCS' versions (as described in Griffiths and Lambert 2011)
Status in employment 2000 and 1990 : Five categories;
1960: Four categories,
Plus all converted to Standardised Employment Status (see below).
2000 and 1990: For the 2000 and 199 versions, the index files of scores allow for status in employment to be coded in two ways, as listed below. The first uses those categories that are conventionally distinguished in contemporary US occupational data and named 'USEMPST' on our index files; the second uses those constructed categories that we derive in order to provide scores that are compatible with other groupings that are used in international comparative studies (see the fuller discussion in the web page on Status in Employment). (Note: in the 1990 scale construction, but not in 2000, the USEMPST categories 4 and 5 were combined).
USA 2000 (USEMPST) Standardised (STDEMPST) 0 Not available, not known etc. (derived from average of categories 1 to 5) Not available, not known etc. (US category 0) 1 Self-Employed, not incorporated Self-employed (all) (derived from average of US categories 1 to 3) 2 Family worker (unpaid) Self-employed (principal) (derived from average of US categories 1 and 3) 3 Self-employed incorporated 4 Government employee (combines federal, state and local government) 5 Private sector employee (combines private and non-profit) Family assistant (US category 2) 6 Employee (derived from average of US categories 4 and 5)
1960: Similarly for the 1960 versions, the index files of scores allow for status in employment to be coded in two ways, as listed below. The first uses those categories that are conventionally distinguished in US occupational data; the second uses those constructed categories that we derive in order to provide scores that are compatible with other groupings that are used in international comparative studies (see the fuller discussion in the web page on Status in Employment).
USA 1960 (US) Standardised 0 Not available, not known etc. (derived from categories 1 to 4) Not available, not known etc. (US category 0) 1 Self-Employed (principals) Self-employed (all) (derived from US categories 1 and 2) 2 Family worker (unpaid) Self-employed (principal) (equal to US category 1) 3 Private employee 4 Government employee 5 Family assistant (US category 2) 6 Employee (derived from US categories 3 and 4)
The 2000 construction used the US 2000 Census classification; the 1990 CAMSIS construction is based upon the American 1990 census occupational classification OCC90, and the 1960 construction based upon the 1960 census occupational classification OCC60. According to the IPUMS website, most US occupational classifications of recent years (there is a different classification every decade since 1950), can be considered reasonably closely related to either the 1990 or 1960 census classifications.
We have also prepared a file which links the CAMSIS scores, as derived on OCC90
units, to appropriate ISCO-88 occupational categorisations, by using the files
distributed by Harry Ganzeboom on the ISMF project which link the two schema.
For the 2000 and 1990 versions, the scores were based upon analysis of a subsample of the 363421 cases of both-currently-working husband-wife couples as obtained from the IPUMS census subsample (for 1990, 363421 cases were used from all available husband-wife combinations in the 1% sample; for 2000, 700000 cases were used as a random subsample from all available cases in the 5% sample [more cases would have prevented successful implementation of the association models on the lEM software used]).
For the 1960 version, however, the number of both-working couples available from the same subsample would have been significantly smaller, so our units of analysis were ever working intra-household cross-gender couples. The first extension means that many cases were included on the basis of their last job if not currently working. The second means that as well as husband-wife couples, any other male-female combinations of household sharers were also used if they contributed appropriate occupational data. Between these two extensions, we were able to more than double the available sample size to 335287 couples. So far at least, all evidence suggests that such different data resources invariably lead to the same derived CAMSIS structures of social association within a model.
Back to National Versions.