11.7.05 CAMSIS LIS FILES CURRENTLY UNDER REVISION, PLEASE CONTACT PAUL LAMBERT FOR LATEST NEWS

---

• Main CAMSIS LIS page (below):
  • Download CAMSIS - LIS datasets [zip file - latest version April 2003]
  • Introduction
  • Accessing the CAMSIS files stored at the LIS-Project
  • Latest studies covered by country
  • **Notes applicable to particular countries or studies**
  • Other variables included in files
  • Comment : Dealing with cross-gender populations
  • Citations
• Examples page: Example syntax and analysis for files for each country.

 

On this page we provde access to. and guidance on, a group of data files which enable the derivation of CAMSIS scale scores for selected countries and time periods within the datasets covered by the Luxembourg Income and Employment studies.

The Luxembourg Income and Luxembourg Employment Studies (LIS and LES) comprise the cross-nationally harmonised longitudinal (repeated cross-sectional) survey data collection. Topics covered include information on individual and family income, individual employment conditions, and many other related variables. They are constructed and maintained at the Luxembourg Income Study central offices in Luxembourg (see their 'lisproject.org' webpages), also receiving backing and support from a group at Syracuse University, New York. The studies are made openly available to researchers throughout the world upon registration with the LIS project (full details are found at the project webpages, which also include extensive documentation, and information on research using LIS and LES). The studies therefore represent a very accessible resource for cross-nationally comparative social science research.

Access to the LIS and LES studies is organised remotely over the internet : researchers submit by email command files in a given software package, then these files are run on the data files and computers which host the studies, and the output from the files subsequently returned by email. One consequence of this is that matching CAMSIS scores with occupational data via the file matching techniques advocated elsewhere on these webpages is not readily possible for use with the LIS and LES studies. Thus, in an ongoing project, we are in the process of supplying the appropriate CAMSIS score matching files to the LIS project (as and when we complete the scale constructions ourselves).

In previous years, these CAMSIS data files were stored on the LIS machines in Luxembourg, from where they could subsequently be called upon within a user's command file.

At present [notes last revised July 2007] the files are not located on the LIS machines, and any user wishing to exploit them will need to ask permission individually at LIS to deposit the CAMSIS data files on the LIS servers. The relevant CAMSIS data files can be downloaded here.

This document describes the availability of CAMSIS scoring files suitable for working on LIS. These notes were briefly revised in July 2007, but the main content of these pages was last updated in 2003. It is therefore possible that some additional LIS datasets could be linked with CAMSIS schemes which are not mentioned in this webpage. The most relevant section of this page for most readers will be the notes on LIS/LES files relevant to specific countries.

The CAMSIS data files prepared for the LIS machines since January 2003 represent release 2 of the CAMSIS LIS files. In line with other revisions made to the files downloadable from the CAMSIS webpages at the same time, a slight modification was made to the handling of employment status information, which affects both the scale derivation process and subsequent use of the derived CAMSIS scores. Resources had been supplied to the LIS machines for 4 countries (UK, US, Germany and Switzerland) prior to January 2003, and those versions were thus revised for the 'revision 2' update. The changes between releases are very small; we hope that the only appreciable difference will be a slight improvement in the consistency between scores given to the same occupations within countries, when different degrees of employment status information are available.

Download: CAMSIS project LIS/LES datasets (Zip file)

For each specific LIS and LES study for which it is possible, we have provided 'index' files which are stored with the LIS-Project in Luxembourg, and contain the appropriate CAMSIS scores for the relevant occupational units. In general those files have names related to the LIS and LES datasets, but full details are below.

The index files are available in two formats, as SPSS data files named *.sav ; or as plain text data files named *.dat (which can be used for analyses using STATA and SAS). Whichever software package a user has submitted their LIS / LES job in (there is a choice of those three types), it should be possible to call the appropriate index file and match it with the data in the current LIS or LES data file. Before March 2003 we also supplied many files in STATA datafile (*.dta) format. It seems, however, that it is easier for us simply to supply plain text files which may be read directly into STATA programmes..

To exploit these data files, users of the LIS and LES datasets will first need to contact the LIS/LES administration individually and request depositing the above data files in an appropriate space on their machine. Then, they will need to specificy the full directory path in which the files are located on the LIS machines, which may be something like :

$mydata\camsis\*.* (as aliased in STATA)

or

u:/camsis/*.* (as aliased in SPSS)

Matching from these files to the main LIS / LES datasets is done by creating matching variables, for employment title and status, in the current data file, which should be named after the corresponding index variables used in the CAMSIS matching file (see table of country / study notes). This variables should be an occupational unit indicator with a numeric value equal to the employment title unit, and an employment status indicator, with a numeric value representing the index file category to which the LIS / LES data allows us best to approximate (the value 0 if we have no information). See the notes for specific countries for details of the meanings of title and status units by LIS and LES countries and datasets.

The syntax used to match the CAMSIS files with the LIS / LES datasets obviously differs by packages (access examples on this page). However the basics are as follows, where the variables OCC and STAT represent the derived occupational title and status numeric variables, and CAMSISF the name of the CAMSIS index file:

• In SPSS :

  sort cases by OCC STAT.
  match files file=* /in=tempvar /table=CAMSISF.sav /by=OCC STAT.
  select if (tempvar=1).

• In STATA :

  
  sort OCC STAT;
  save tempdat, replace;
  infile OCC STAT MCAM FCAM using $CAMSISF.dat;
  merge OCC STAT using tempdat;

• In SAS :

  [to be added]

If wishing to match CAMSIS occupational scores to more than one variable on the same file (for instance, in the household file to match scores to both the head of household's occupation and the occupation of the spouse of the head of household), the syntax above needs to be repeated, with the additional use of renaming commands in order to distinguish the variable names of the respective scores.

One word of warning on using the LIS files in this application: The CAMSIS measures assume we are matching a score value to the occupation of an individual, but users should note that the LIS and LES files contain household as well as individual level data. This is dealt with easily enough - to use household level information we just need to be clear which householder the (occupation) refers to. However on some sections of the LIS files the data format can be confusing, such that without care some household level variables are easily mistaken for individual level ones. This is particularly common in some of the older datasets, where information pertaining to the head of household is sometimes stored on individual records.

 

(last update to this table : 2.4.03) 

 

• Austria 1987,95 LIS and 1991 LES
• Britain 1991 LIS and 1989,97 LES
• Czech Republic1992, 96 LIS and 1994 LES
• Germany 1989,94 LIS
• Hungary 1991, 1994 LIS and 1993 LES
• Ireland 1994,95,96 LIS
• Russia 1992,95 LIS
• Slovakia 1992 LIS, 1995 LES
• Slovenia 1994 LES
• Sweden 1990 LES
• Switzerland 1992 LIS and 1997 LES
• USA 1999,94,97 LIS and 1990 LES

---

Austria (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1991 LES (base unit: ISCO-88 unit groups, in 4 digit form)
STATA	csates91.dat	isco88 empst	mcam	fcam
SPSS	csates91.sav	``	``	``
SAS	csates91.dat	Columns: isco88 empst mcam fcam siops isei
1995 LIS (base unit: ISCO-88 unit groups, in 4 digit form)
STATA	csatis95.dat	isco88 empst	mcam	fcam
SPSS	csatis95.sav	``	``	``
SAS	csatis95.dat	Columns: isco88 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Austria:

Base units:

The 1991 LES and 1995 LIS studies contain occupational information at the level of ISCO-88 unit groups (4 digit codes). This data was actually the basis of the CAMSIS scale construction itself, see the CAMSIS-Austria page. The CAMSIS LIS index file thus uses the same occupational units and should give full coverage of the 1991 LES and 1995 LIS occupations.

The key linking variables on the distributed files are all called isco88 and empst. LIS and LES users will need to derive these two variables, by calculating them as new variables equal to the ( isco88) variables on the LIS / LES files, and for empst, as a numerical indicator of employment status category following the Austrian CAMSIS category schema described on the CAMSIS webpages for the contemporary Austrian version . The necessary employment status categories for empst are : 0 - Unknown; 1 - All Self-Employed; 6 - Employees.

[Notes by country ]

Additional variables:

1991 LES csates91	ISEI; SIOPS;
1995 LIS csatis95	ISEI; SIOPS;

For both files we have included variables 'ISEI' and 'SIOPS' showing, respectively, Socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained by recoding the ISCO-88 unit (using a macro provided at Harry Ganzeboom's ISMF site). See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant Austrian datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

Britain (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1991 LIS (base unit: 1990 SOC)
STATA	csukis91.dat	soc90 empst	mcam	fcam
SPSS	csukis91.sav	``	``	``
SAS	csukis91.dat	Columns: soc90 empst mcam fcam siops isei
1989 LES (base unit: 1988 ISCO)
STATA	csukes89.dat	isco88 empst	mcam	fcam
SPSS	csukes89.sav	``	``	``
SAS	csukes89.dat	Columns: isco88 empst mcam fcam siops isei
1997 LES (base unit: 1990 SOC)
STATA	csukes97.dat	soc90 empst	mcam	fcam
SPSS	csukes97.sav	``	``	``
SAS	csukes97.dat	Columns: soc90 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Britain:

Base units:

The 1991 LIS and 1997 LES files link the SOC-90 UK codes on the respective datasets directly to the CAMSIS scores derived for SOC-90 from 1991 census data. The 1989 LES dataset however has occupational detail to the ISCO-88 level; CAMSIS scores for the UK have been assigned to ISCO-88 units, see the 'national versions' section of this website, after employing an approximation linking ISCO-88 to SOC-90, as obtained from the Occupational Information Unit (see comments on the 'occupational units' section of this site).

The key linking variables on the distributed files are called soc90 / isco88 and empst (for SOC-90 and ISCO88 units, and employment status, respectively). LIS and LES users will need to derive these 'title-by-status' variables, by calculating them as new variables equal to the (soc90 / isco) variables on the LIS / LES files, and for empst, as a numerical indicator of employment status category following the UK CAMSIS category schema described on the CAMSIS webpages for the contemporary British version (standardised version). The necessary employment status categories are : 0 - Unknown; 2 - All Self-Employed; 3 - Self-employed with employees; 4 - Self-employed own account; 6 - Employees. The distributed files are intended to cover every possible {SOC90 or ISCO88}-by-status combination, although some units may have missing values (indicated -999) on the ISEI or SIOPS variables (see below), due to a lack of derivation of such scores for the relevant (probably rare) units.

[Notes by country ]

Additional variables:

1991 LIS csukis91	ISEI; SIOPS;
1989 LES csukes89	ISEI; SIOPS;
1997 LES csukes97	ISEI; SIOPS;

For all three files we have included variables 'ISEI' and 'SIOPS' showing, respectively, Socio-economic status and prestige scores allocated to the relevant occupations. For the case of the 1989 LES study these variables were obtained by recoding the ISCO-88 unit (using a macro provided at Harry Ganzeboom's ISMF site). For the 1991 LIS and 1997 LES datasets, the scores were obtained by approximating ISCO-88 units for the 1990 SOC occupational units in the data (following the macros available from our occupational units page), then subsequently recoding the ISCO-88 units as above. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant British datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

 

Czech Rebuplic (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1994 LES (base unit: ISCO-88 minor groups, in 4 digit form)
STATA	csczes94.dat	isco88 empst	mcam	fcam
SPSS	csczes94.sav	``	``	``
SAS	csczes94.dat	Columns: isco88 empst mcam fcam siops isei
1992 LIS (base unit: ISCO-88 submajor group, in 2 digit form)
STATA	csczis92.dat	isco88 empst	mcam	fcam
SPSS	csczis92.sav	``	``	``
SAS	csczis92.dat	Columns: isco88 empst mcam fcam siops isei
1996 LIS (base unit: ISCO-88 submajor group, in 4 digit form)
STATA	csczis96.dat	isco88 empst	mcam	fcam
SPSS	csczis96.sav	``	``	``
SAS	csczis96.dat	Columns: isco88 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Czech Republic:

Base units:

The 1994 LES study contains occupational information at the level of ISCO-88 minor groups (3 digit codes presented in 4 digit form, eg minor group 933 is given value 9330). This data was actually the basis of the CAMSIS scale construction itself, see the CAMSIS-Czech page. The CAMSIS LIS index file thus uses the same occupational units and should give full coverage of the 1994 LES occupations. The 1992 and 94 LIS datasets both restrict occupational information to ISCO-88 submajor groups (2-digit values). The corresponding CAMSIS LIS index files for each year link the submajor group average CAMSIS and ISEI/SIOPS values with each such occupational unit. We can thus expect a loss of detail in the value of the CAMSIS scores for these LIS datasets, reflecting the averaging process.

The key linking variables on the distributed files are all called isco88 and empst. LIS and LES users will need to derive these two variables, by calculating them as new variables equal to the ( isco88) variables on the LIS / LES files, and for empst, as a numerical indicator of employment status category following the Czech CAMSIS category schema described on the CAMSIS webpages for the contemporary Czech version . The necessary employment status categories for empst are : 0 - Unknown; 1 - All Self-Employed; 6 - Employees.

[Notes by country ]

Additional variables:

1994 LES csczes94	ISEI; SIOPS;
1992 LIS csczis92	ISEI; SIOPS;
1996 LIS csczis96	ISEI; SIOPS;

For all three files we have included variables 'ISEI' and 'SIOPS' showing, respectively, Socio-economic status and prestige scores allocated to the relevant occupations. For the case of the 1994 LES study these variables were obtained by recoding the ISCO-88 unit (using a macro provided at Harry Ganzeboom's ISMF site). For the 1992 and 1996 LIS datasets, the scores were calculated as subgroup averages from the constituent unit groups of each submajor group. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant British datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

 

Germany (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1989 LIS (base unit: 1968 ISCO minor group)
STATA	csdeis89.dat	is68min empst	mcam	fcam
SPSS	csdeis89.sav	` `	` `	` `
SAS	csdeis89.dat	Columns: is68min empst mcam fcam siops isei
1994 LIS (base unit: 1968 ISCO minor group)
STATA	csdeis94.dat	is68min empst	mcam	fcam
SPSS	csdeis94.sav	` `	` `	` `
SAS	csdeis94.dat	Columns: is68min empst mcam fcam siops isei

[Notes by country ]

Notes specific to Germany:

Base units:

The 1989 and 94 CAMSIS LIS files link the ISCO-68 2-digit 'minor' group occupational codes of the respective LIS datasets, to CAMSIS scores derived at the ISCO-88 3 digit minor group level from 1995 micro-census data. This linkage involves two levels of approximation. First, the ISCO-88 unit level scores were linked with corresponding ISCO-68 base units by using macros available from Harry Ganzeboom's ISMF website. Second, average ISCO-88 based scores for ISCO-68 minor groups were caculated, and it is these averages which are associated with the German LIS data which classifies to the level of ISCO-68 minor groups (2-digits). The basic nature of the CAMSIS scores is likely to be robust to such approximations, but many of the finer details of the approach may be lost by this averaging process.

The key linking variables on the distributed files are called is68min and empst (standing for ISCO-68 minor group, and employment status). LIS users will need to derive these variables, by calculating new variables equal to the ISCO-68 minor group units, for the first, and the employment status categorical variable for empst. The numerical categories on empst follow the German CAMSIS (standardised) category schema described on the CAMSIS webpages for the contemporary German ISCO version. The relevant categories are : -999 - Status unknown; 1 - Self-employed (all); 2 - Self-employed (principals); 3 - Own account; 4 - Employer; 5 - Family assistant; 6 - Employee.

The distributed files cover most ISCO68 minor groups which occur within the LIS datasets, but, because they are derived from an ISCO-88 schema which when back-translated does not link directly to some ISCO-68 categories, the scores in the CAMSIS index file do not cover every combination. As a partial solution we have also calculated ISCO-68 1-digit (major group) units and the associated average scores by employment status category, and added them to our index file with the numerical value 200 + major group number (a direct numerical value would coincide with some minor group values). The syntax example files show how these averages can be used to fill in scores to ISCO-68 minor groups which do not in themselves have values matched to them.

[Notes by country ]

Additional variables:

1989 LIS csdeis89	ISEI; SIOPS;
1994 LIS csdeis94	ISEI; SIOPS;

For both files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated Socio-economic status and prestige scores allocated to the relevant occupations. For both cases these variables were obtained by recoding the ISCO-88 unit, then averaging the scores of ISCO-88 into repsective ISCO-68 minor groups (using macros provided at Harry Ganzeboom's ISMF site). This latter averaging step means that the ISEI and SIOPS variables should be regarded as approximations. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant German datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

Hungary (also see examples page for example command file syntax)

Format	Filename	Index vars	Male CAMSIS	Female CAMSIS
1991 LIS (base unit: 1968 ISCO unit group)
STATA	cshuis91.dat	isco68 empst	mcam	fcam
SPSS	cshuis91.sav	` `	` `	` `
SAS	cshuis91.dat	Columns: isco68 empst mcam fcam siops isei
1994 LIS (base unit: 1988 ISCO unit group)
STATA	cshuis94.dat	isco88 empst	mcam	fcam
SPSS	cshuis94.sav	` `	` `	` `
SAS	cshuis94.dat	Columns: isco88 empst mcam fcam siops isei
1999 LIS (base unit: 1988 ISCO unit group, uses 1994 LIS index files)
STATA	cshuis94.dat	isco88 empst	mcam	fcam
SPSS	cshuis94.sav	` `	` `	` `
SAS	cshuis94.dat	Columns: isco88 empst mcam fcam siops isei
1993 LES (base unit: 1988 ISCO unit group)
STATA	cshues93.dat	isco88 empst	mcam	fcam
SPSS	cshues93.sav	` `	` `	` `
SAS	cshues93.dat	Columns: isco88 empst mcam fcam siops isei
1999 LES (base unit: 1988 ISCO unit group, uses 1993 LES index file)
STATA	cshues93.dat	isco88 empst	mcam	fcam
SPSS	cshues93.sav	` `	` `	` `
SAS	cshues93.dat	Columns: isco88 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Hungary:

Base units:

The 1994 and 99 LIS, and 1993 and 99 LES, CAMSIS files link the ISCO-88 unit group occupational codes of the respective LIS datasets, to CAMSIS scores derived at the ISCO-88 unit level from 1996 labour force survey data. Coverage of the ISCO88 units in the 1994 LIS file by CAMSIS scores is very high, but there are quite a few (about 7% of valid cases) occupations listed in ISCO88 units on the 1993 LES files where there is no corresponding unit in the CAMSIS index file.

The 1991 LIS CAMSIS file links the ISCO-68 unit group occupational code of the LIS dataset with CAMSIS scores derived for ISCO-68 units via a translation from the ISCO-88 1996 Hungarian CAMSIS derivation. This translation (obtained from Harry Ganzeboom's ISMF website, see notes on our 'occupational units' page) is to some extent an approximation and not every possible ISCO-68 unit is attributed a score on the basis of the ISCO-88 unit group it would be a member of. The basic nature of the CAMSIS scores is likely to be robust to such translations between unit group schema, but the loss of valid cases in this LIS file (only around two thirds of occupations in the 1991 LIS file are successfully translated into a valid isco88 occupation), may be more restrictive.

The key linking variables on the distributed files are called isco68 / isco 88 and empst (standing for ISCO-68 / 88 unit groups, and employment status, respectively). LIS users will need to derive variables on the LIS files which share these variable names and hold the appropriate values. Examples of how to do this are in the syntax examples. For the isco68 / 88 units, it is simply necessary to compute a variable equal to the occupational unit information on the LIS datasets. For the empst units, it is necessary to calculate appropriate categories by recoding on the basis of LIS information on employment and labour market status, in line with the employment status categories described in the main section of the CAMSIS webpages for the Hungarian version. The necessary schema for the variables used in the Hungarian CAMSIS LIS files is : 0 - Unknown employment status; 1 - Self-employed; 2 - Employee. The Hungarian LIS data seems to allow full information on employment status, but a default failsafe is to use the category '0' for unknown employment status.

[Notes by country ]

Additional variables:

1991 LIS cshuis91	ISEI; SIOPS;
1994,99 LIS cshuis94	ISEI; SIOPS;
1993, 99 LES cshues93	ISEI; SIOPS;

For all files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained by recoding the ISCO-88 unit (then in the case of the LIS91 file, averaging the scores of ISCO-88 units into repsective ISCO-68 minor groups). This was undertaken using macros provided at Harry Ganzeboom's ISMF site. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant datasets with the supplied CAMSIS files.

[Notes by country ]

---

Ireland (also see examples page for example command file syntax)

Format	Filename	Index vars	Male CAMSIS	Female CAMSIS
1994 LIS (base unit: 1988 ISCO submajor group, expressed in 4 digits)
STATA	csieis94.dat	isco88	mcam	fcam
SPSS	csieis94.sav	` `	` `	` `
SAS	csieis94.dat	Columns: isco88 mcam fcam siops isei
1995 LIS (base unit: 1988 ISCO submajor group, expressed in 4 digits)
STATA	csieis94.dat	isco88	mcam	fcam
SPSS	csieis94.sav	` `	` `	` `
SAS	csieis94.dat	Columns: isco88 mcam fcam siops isei
1996 LIS (base unit: 1988 ISCO submajor group (expressed in 4 digits)
STATA	csieis94.dat	isco88	mcam	fcam
SPSS	csieis94.sav	` `	` `	` `
SAS	csieis94.dat	Columns: isco88 mcam fcam siops isei

[Notes by country ]

Notes specific to Ireland:

Base units:

The 1994, 95 and 96 LIS CAMSIS files link ISCO-88 submajor group averages with the submajor group units covered by the LIS files (note that the LIS data is recorded in 4 digits but only has 2 digit submajor group detail, for example submajor group 72 is recorded as 7200). Coverage of the ISCO88 units in the 1994 LIS file by CAMSIS scores is [.......]very high. The CAMSIS scores supplied are based on the population weighted averages of ISCO-88 scores within each submajor group for Ireland, obtained by translating the original CAMSIS Irish dataset's 'CSO96' occupational units into ISCO-88 units (see the CAMSIS-Ireland page). There are thus two levels of approximations involved when using the LIS index file CAMSIS scores (ISCO-88 translation and submajor group averaging), so this application of CAMSIS scores may be less accurate than other versions.

The key linking variables on the distributed files is called isco88. LIS users will need to derive a variable on their LIS files which shares this variable name and holds the appropriate values. For the isco88 units, it is simply necessary to compute a variable equal to the occupational unit information on the LIS datasets. Examples of how to do this are in the syntax examples page.

[Notes by country ]

Additional variables:

1994, 95 and 96 LIS csieis94

ISEI; SIOPS;

For all files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained as the submajor group average of the ISEI and SIOPS scores assigned to constituent ISCO-88 unit groups, obtained by using macros provided at Harry Ganzeboom's ISMF site. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant datasets with the supplied CAMSIS files.

[Notes by country ]

---

Russia (also see examples page for example command file syntax)

Format	Filename	Index vars	Male CAMSIS	Female CAMSIS
1992 LIS (base unit: 1988 ISCO unit groups, 4 digits)
STATA	csruis92.dat	isco88	mcam	fcam
SPSS	csruis92.sav	` `	` `	` `
SAS	csruis92.dat	Columns: isco88 mcam fcam siops isei
1995 LIS (base unit: 1988 ISCO unit groups, 4 digits)
STATA	csruis95.dat	isco88	mcam	fcam
SPSS	csruis95.sav	` `	` `	` `
SAS	csruis95.dat	Columns: isco88 mcam fcam siops isei

[Notes by country ]

Notes specific to Russia:

Base units:

The 1992 and 95 LIS CAMSIS files link ISCO-88 unit group scores (which were themselves derived from a LIS-related dataset, see the CAMSIS-Russia page), with the units covered by the LIS files. Coverage of the ISCO88 units in the 1994 LIS file by CAMSIS scores should be complete.

The key linking variables on the distributed files is called isco88. LIS users will need to derive a variable on their LIS files which shares this variable name and holds the appropriate values. For the isco88 units, it is simply necessary to compute a variable equal to the occupational unit information on the LIS datasets. Examples of how to do this are in the syntax examples page.

[Notes by country ]

Additional variables:

1992 and 95 LIS csruis92

ISEI; SIOPS;

For all files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained as the submajor group average of the ISEI and SIOPS scores assigned to constituent ISCO-88 unit groups, obtained by using macros provided at Harry Ganzeboom's ISMF site. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant datasets with the supplied CAMSIS files.

[Notes by country ]
  

---

Slovakia (also see examples page for example command file syntax)

Format	Filename	Index vars	Male CAMSIS	Female CAMSIS
1995 LES (base unit: 1988 ISCO unit groups, 4 digits)
STATA	csskes95.dat	isco88	mcam	fcam
SPSS	csskes95.sav	` `	` `	` `
SAS	csskes95.dat	Columns: isco88 mcam fcam siops isei
1992 LIS (base unit: 1988 ISCO submajor groups, 2 digits)
STATA	csskis92.dat	isco88	mcam	fcam
SPSS	csskis92.sav	` `	` `	` `
SAS	csskis92.dat	Columns: isco88 mcam fcam siops isei
1996 LIS (base unit: 1988 ISCO major groups, expressed as 4 digits)
STATA	csskis96.dat	isco88	mcam	fcam
SPSS	csskis96.sav	` `	` `	` `
SAS	csskis96.dat	Columns: isco88 mcam fcam siops isei

[Notes by country ]

Notes specific to Slovakia:

Base units:

The 1995 LES CAMSIS files link ISCO-88 unit group scores (which were themselves derived from this LES dataset, see the CAMSIS-Slovakia page), with the units covered by the LES files. Coverage of the ISCO88 units in the 1994 LIS file by CAMSIS scores should be complete. The 1992 and 96 LIS files however only have occupational data in truncated forms: for 1992 in terms of ISCO-88 submajor group (2-digit) units, and in 1996 in terms of ISCO-88 major group (1 digit) units. CAMSIS scores are attached to the LIS based index file on the bases of subgroup averages for the two types of aggregated groups, but we do not expect the same level of accuracy from the two LIS based linkage files, in particular the major group only 1996 survey.

The key linking variables on the distributed files is called isco88. LIS users will need to derive a variable on their LIS files which shares this variable name and holds the appropriate values. For the units of each mentioned Slovak file, it is simply necessary to compute a variable equal to the occupational unit information on the LIS datasets. Examples of how to do this are in the syntax examples page.

[Notes by country ]

Additional variables:

1992 and 95 LIS csskis92

ISEI; SIOPS;

For all files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained as the submajor group average of the ISEI and SIOPS scores assigned to constituent ISCO-88 unit groups, obtained by using macros provided at Harry Ganzeboom's ISMF site. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant datasets with the supplied CAMSIS files.

[Notes by country ]
  

---

Slovenia (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1994 LES (base unit: ISCO-88 unit groups, in 4 digit form)
STATA	cssles94.dat	isco88 empst	mcam	fcam
SPSS	cssles94.sav	``	``	``
SAS	cssles94.dat	Columns: isco88 empst mcam fcam siops isei
1999 LES (base unit: ISCO-88 submajor groups, in 4 digit form)
STATA	cssles99.dat	isco88 empst	mcam	fcam
SPSS	cssles99.sav	``	``	``
SAS	cssles99.dat	Columns: isco88 empst mcam fcam siops isei
1997 LIS (base unit: ISCO-88 submajor groups, in 4 digit form: uses 1999 LES files)
STATA	cssles99.dat	isco88 empst	mcam	fcam
SPSS	cssles99.sav	``	``	``
SAS	cssles99.dat	Columns: isco88 empst mcam fcam siops isei
1999 LIS (base unit: ISCO-88 submajor groups, in 4 digit form: uses 1999 LES files)
STATA	cssles99.dat	isco88 empst	mcam	fcam
SPSS	cssles99.sav	``	``	``
SAS	cssles99.dat	Columns: isco88 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Slovenia:

Base units:

The 1994 LES study contains occupational information at the level of ISCO-88 unit groups (4 digit codes). This data was actually the basis of the CAMSIS scale construction itself, see the CAMSIS-Slovenia page. The CAMSIS LIS index file thus uses the same occupational units and should give full coverage of the 1994 LES occupations. The 1999 LES study and the 1997 and 1999 LIS studies all contain occupational information at the level of ISCO-88 (two digit) submajor groups, expressed in 4 digit codes with trailing zeros (eg submajor group 72 becomes 7200). The CAMSIS scores in the 1999 LES index file, which is used for linking scores to all three files, are calculated as the appropriate submajor group averages of constituent unit group scores.

The key linking variables on the distributed files are all called isco88 and empst. LIS and LES users will need to derive these two variables, by calculating them as new variables equal to the ( isco88) variables on the LIS / LES files, and for empst, as a numerical indicator of employment status category following the Slovenian CAMSIS category schema described on the CAMSIS webpages for the contemporary Slovene version . The necessary employment status categories for empst are : 0 - Unknown; 1 - All Self-Employed; 6 - Employees.

[Notes by country ]

Additional variables:

1994 LES cssles94	ISEI; SIOPS;
1999 LES cssles99 (and 1997,99 LIS by default)	ISEI; SIOPS;

For both files we have included variables 'ISEI' and 'SIOPS' showing, respectively, Socio-economic status and prestige scores allocated to the relevant occupations. These variables were obtained by recoding the ISCO-88 unit (using a macro provided at Harry Ganzeboom's ISMF site). See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant Slovene datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

Sweden (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1987 LIS (base unit: Swedish NYK minor group)
STATA	csswis87.dat	nykmin; empst	mcam	fcam
SPSS	csswis87.sav	` `	` `	` `
SAS	csswis87.dat	Columns: nykmin empst mcam fcam siops isei
1992 LIS (base unit: Swedish NYK minor group)
STATA	csswis92.dat	nykmin; empst	mcam	fcam
SPSS	csswis92.sav	` `	` `	` `
SAS	csswis92.dat	Columns: nykmin empst mcam fcam siops isei
1995 LIS (base unit: Swedish NYK minor group)
STATA	csswis95.dat	nykmin; empst	mcam	fcam
SPSS	csswis95.sav	` `	` `	` `
SAS	csswis95.dat	Columns: nykmin empst mcam fcam siops isei
1990 LES (base unit: 1988 ISCO unit group)
STATA	csswes90.dat	isco88 empst	mcam	fcam
SPSS	csswes90.sav	` `	` `	` `
SAS	csswes90.dat	Columns: isco88 empst mcam fcam isei siops

[Notes by country ]

Notes specific to Sweden

Re-release April 2003:

A minor correction was made to the main CAMSIS index files for Sweden in Feburary 2003 (see the Swedish-CAMSIS page). In April 2003 the index files on the LIS machines were updated in line with this correction. Differences between the previous and latest releases, though, are very slight.

Base units:

The Swedish 1990 LES file gives data in ISCO-88 units, although it should be noted that this classification is itself, on the LES file, an aggregation, which only spans around 60 ISCO-88 titles. CAMSIS scores are then attached which are distributed to ISCO-88 units based upon the 1990 CAMSIS scale derivation. This scale was calculated on Swedish national NYK83 units, then a translation to ISCO-88 undertaken using the macro, presented in the occupational information section of these pages, which was produced by Erik Bihagen. There is thus a double translation in the data construction of this example, which may to count against the strength of the CAMSIS scores for this Swedish dataset. We do, nevertheless, expect the CAMSIS scores to be broadly robust to the approximations involved in deriving the ISCO-88 units.

All three Swedish LIS files give occupational data in a Swedish classification of approx 60 categories. These are almost, but not quite, equivalent to the NYK83 Swedish minor group classifications shown in the main CAMSIS Swedish notes. To assign the scores given to those NYK83 minor groups to the LIS data, a manual recode of the NYK83 categories was undertaken to derive a variable comparable to the LIS categorisation. The changes needed to the categories are fairly small so this stage of translation could be expected to be quite reliable. Users should beware, however, that the LIS category groupings are unlikely to be good indicators of occupational differences in themselves, as they are based most on industrial sectors and cross-cut many of the occupational divisions more usually regarded as significant. Thus, we would not expect the Swedish LIS CAMSIS scores matched through this process to be as informative as in other datasets.

The key linking variables on the distributed files are called isco88/nykmin and empst (standing for ISCO-88 unit group or NYK minor group, and employment status). LIS users will need to derive variables on the LIS files which share these variable names and hold the appropriate values. Examples of how to do this are in the syntax examples. For the isco88 units on the LES file, and the NYK minor group units on the LIS files, it is simply necessary to compute a variable equal to the occupational unit information on the respective datasets. For the empst units, it is necessary to calculate appropriate categories by recoding on the basis of LIS information on employment and labour market status, in line with the employment status categories described in the main section of the CAMSIS webpages for the Swedish version. The necessary employment status schema for the variables used in the CAMSIS LIS Swedish files is : 0 - Unknown employment status; 1 - Self-employed; 2 - Employee.

[Notes by country ]

Additional variables:

1990 LES cssees90

ISEI; SIOPS;

For the Swedish 1990 LES file, the ISCO-88 values have had the variables 'ISEI' and 'SIOPS' added. These show, respectively, approximated socio-economic status and prestige scores allocated to the relevant occupations, following the discussion by Ganzeboom et al 1996. The values were derived with the macros provided at Harry Ganzeboom's ISMF site. It was not possible to derive equivalent variables based upon the NYK minor group occupational classifications used in the Swedish LIS files. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant Swedish datasets with the supplied CAMSIS files.

[Notes by country ]
 

---

Switzerland (Also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1992 LIS (base unit: 1990 Swiss National 'BTAP' minor group)
STATA	cschis92.dat	btapm empst	mcam	fcam
SPSS	cschis92.sav	` `	` `	` `
SAS	cschis92.dat	Columns: btapm empst mcam fcam siops isei
1997 LES (base unit: 1988 ISCO)
STATA	csches97.dat	isco88 empst	mcam	fcam
SPSS	csches97.sav	` `	` `	` `
SAS	csches97.dat	Columns: isco88 empst mcam fcam siops isei

[Notes by country ]

Notes specific to Switzerland:

Re-release April 2003: A minor correction was made to the main CAMSIS index files for Switzerland in Feburary 2003 (see the Swiss-CAMSIS page). In April 2003 the index files on the LIS machines were updated in line with this correction. Differences between the previous and latest releases, though, are very slight.

The 1992 LIS contains employment status data which conforms with that used in the CAMSIS scale constructions. Whilst the title unit is the Swiss 'BTAP' schema, it is truncated to the level of minor groups. The index file supplied to the LIS project, which utilises the CAMSIS 1990 Swiss scores for BTAP units, has correspondingly been set up at the level of minor group units (by taking the average BTAP unit scores within minor groups, over the national population). The CAMSIS scores associated with the LIS index files are therefore approximations applied to minor group units rather than precise occupational unit codes.

The 1997 LES contains employment status data which conforms with that used in the CAMSIS scale constructions. The title unit is the 'ISCO-88(COM)' base unit, which is also one of the units used by the CAMSIS scale for Switzerland. The index file supplied to the LIS project is thus available at the level of ISCO unit groups, using CAMSIS scores for the Swiss 1990 dataset of ISCO unit groups.

The key linking variable on these distributed files are the employment title units - called btapm or isco88- and the employment status indicator empst. LIS users need to derive variables with these names on the relevant datafiles, then match in the index files using these variables as index keys. The occupational title variables just have the numeric value of the occupational unit. The empst variable is a numerical indicator variable for employment status category following the Swiss CAMSIS (standardised) schema as described on the CAMSIS webpages for the contemporary Swiss versions . The relevant categories are : 0 - Status unknown; 1 - Self-employed (all); 2 Self-employed (principals); 5 Family worker; 6 - Employee. The distributed files should cover every possible unit-by-status combination, although some units may have missing values (indicated -999) on the ISEI and SIOPS variables (see below), due to a lack of derivation of such scores for the relevant (probably rare) units.

[Notes by country ]

Additional variables:

1992 LIS cschis92	ISEI; SIOPS;
1997 LES csches97	ISEI; SIOPS;

For both files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated Socio-economic status and prestige scores allocated to the relevant occupations. For both cases these variables were by recoding ISCO-88 units into the two measures (using macros provided at Harry Ganzeboom's ISMF site). The 1992 LIS files required in addition a translation to ISCO-88 from the PBER schema (the file used by the CAMSIS project, described on the 'occupational units' page, was used), and, additionally, the averaging of ISEI and SIOPS scores within the minor group base units, meaning that the ISEI and SIOPS scores for this LIS file should be regarded as approximations. See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant Swiss datasets with the supplied CAMSIS files.

[Notes by country ]

 

---

 

USA (also see examples page for example command file syntax)

Format	Filename	Index var	Male scales	Female Scales
1991 LIS (base unit: 1990 SOC)
STATA	csusis91.dat	occ90 empst	mcam	fcam
SPSS	csusis91.sav	` `	` `	` `
SAS	csusis91.dat	Columns: occ90 empst mcam fcam siops isei
1994 LIS (base unit: 1990 SOC)
STATA	csusis94.dat	occ90 empst	mcam	fcam
SPSS	csusis94.sav	` `	` `	` `
SAS	csusis94.dat	Columns: occ90 empst mcam fcam siops isei
1997 LIS (base unit: 1990 SOC)
STATA	csusis97.dat	occ90 empst	mcam	fcam
SPSS	csusis97.sav	` `	` `	` `
SAS	csusis97.dat	Columns: occ90 empst mcam fcam siops isei
1990 LES (base unit: 1988 ISCO)
STATA	csuses90.dat	isco88 empst	mcam	fcam
SPSS	csuses90.sav	` `	` `	` `
SAS	csuses90.dat	Columns: isco88 empst mcam fcam siops isei
1997 LES (base unit: 1980 SOC)
STATA	csuses97.dat	soc80 empst	mcam	fcam
SPSS	csuses97.sav	` `	` `	` `
SAS	csuses97.dat	Columns: soc80 empst mcam fcam siops isei

[Notes by country ]

Notes specific to USA:

Base units:

The 1991, 94 and 97 CAMSIS LIS files link the SOC-90 occupational unit codes of the respective LIS datasets, to CAMSIS scores derived for the SOC-90 units from 1990 census data.

The 1990 CAMSIS LES files link the ISCO-88 occupational unit codes of the LES database to CAMSIS scores derived for the SOC-90 units from 1990 census data. The linkage is achieved by recoding the SOC-90 units of the CAMSIS index file to corresponding ISCO-88 units, following the index files provided by Harry Ganzeboom at the ISMF website (in fact because this is an often needed recode, an index file linking ISCO-88 and CAMSIS scores for the US is included in the downloadable CAMSIS archive of US version, available from the US CAMSIS page).

The 1997 CAMSIS LES files links the SOC-80 occupational unit codes of the respective LES dataset, to CAMSIS scores derived for the SOC-90 units from 1990 census data. The linkage is achieved by recoding the SOC-80 units of the LES file to corresponding ISCO-68 then ISCO-88 units, then matching those units with CAMSIS scores linked to ISCO-88 units as described above. These successive translations were achieved by using files provided by IPUMS, and by Harry Ganzeboom at the ISMF website. We must assume that there is some capacity for error in these translations so the linked CAMSIS scores should only be regarded as approximations in this case, for the US 1997 LES file. In addition, a number of SOC-80 units present on the LES file could not sucessfully be associated with any ISCO-88 unit and hence with a relevant CAMSIS score - the index file shows a score of -999 for these missing cases.

The key linking variable on these distributed files are the employment title units - called occ90, isco88 or soc80 - and the employment status indicator empst. LIS users need to derive variables with these names on the relevant datafiles, then match in the index files using these variables as index keys. The occupational title variables just have the numeric value of the occupational unit. The empst variable is a numerical indicator variable for employment status category following the US CAMSIS (standardised) schema as described on the CAMSIS webpages for the contemporary USA SOC 90 version . The relevant categories are : 0 - Status unknown; 1 - Self-employed (all); 2 Self-employed (principals); 5 Family worker; 6 - Employee. The distributed files should cover every possible {SOC-90 unit}-by-status combination, although some units may have missing values (indicated -999) on the ISEI and SIOPS variables (see below), due to a lack of derivation of such scores for the relevant (probably rare) units.

[Notes by country ]

Additional variables:

1991 LIS csusis91	ISEI; SIOPS;
1994 LIS csusis94	ISEI; SIOPS;
1997 LIS csusis97	ISEI; SIOPS;
1990 LES csuses90	ISEI; SIOPS
1997 LES csuses97	ISEI; SIOPS

For all files we have included variables 'ISEI' and 'SIOPS' showing, respectively, approximated Socio-economic status and prestige scores allocated to the relevant occupations. For both cases these variables were obtained by first translating the occupational unit to ISCO-88 if necessary, then recoding ISCO-88 units into the two measures (using macros provided at Harry Ganzeboom's ISMF site). See also an additional note on these variables.

[Notes by country ]

Example Command File Syntax:

Follow this link to access a page which includes links to example command files matching the relevant US datasets with the supplied CAMSIS files.

[Notes by country ]

 

---

---

 
 

When possible we have also included on the LIS and LES files other variables regularly used as occupational location indicators in social stratification research. Although there are many further examples of national specific classifications, we are only aware of two other examples of systematic work allowing for cross-nationally comparable coding of occupatonal units (corrections welcome!).

First is the series of files made available by Harry Ganzeboom at his ISMF site (also see notes on our 'occupational information' page). These allow for the coding of ISCO units into the Trieman 1977 'SIOPS' measure (Standard International Occupatonal Prestige Scale), the Ganzeboom et al 1992 'ISEI' measure (International Socio-Economic Status Index), and into the Erikson and Goldthorpe 1993 'EGP' or 'CASMIN' (Erikson, Golthorpe and Portacero, or Comparative Analysis of Social Stratification and Mobility in Industrialised Nations) class scheme. Both the ISEI and SIOPS measures are easily obtained as a linear translation of ISCO units, and are included in most of our index files. There is one complication to the SIOPS / ISEI values provided above, however. In a number of LIS/LES studies the unit of occupational information is resticted to subgroup averages (eg, an ISCO-88 unit in 2 digits only). In such cases the SIOPS/ISEI scores which we add to our CAMSIS index files are the averages of the unit group scores of the constituent group, and thus are not necessarily equivalent between countries. An alternative set of cross-nationally equivalent subgroup average scores can be obtained from the ISMF distribution files, where scores for 4-digit ISCO-88's which end in zeros represent appropriate subgroup averages (eg, 7200 for submajor group 72).

lThe EGP categories provided from the same ISMF source also require information on employment status details, but since the criteria required are not usually exactly equivalent to those details incorporated in the CAMSIS employment status variables, we do not usually include EGP units with the CAMSIS index files (though interested users could calculate their own derivations using the ISMF files).

The second representation is the discussion of ISCO subgroups and skill categorisations within the ISCO schema presented by Peter Elias (for example this paper). Here for instance a 4-fold categorisation of skill levels, and 10 fold categorisation of major groups, is presented. In this case however, the categorisation of levels within the ISCO schema is so easily obtained that we do not include it in our index files. For illustration, the following SPSS syntax generates ISCO major groups and skill levels according to this schema from base ISCO-88 units.

* ISCO major groups, ISCO-88 variable is 'isco'.

compute ismaj=trunc(isco/1000).

* ISCO skill levels.

recode ismaj (1,0=-999) (2=4) (3=3) (4,5,6=2) (7,8=1) into 'isskill'.

It should be noted that unlike the CAMSIS scale scores, all of the above representations are imposed as equivalent between genders and different countries - the relative merits of the alternative approaches for comparative research are discussed briefly in this ISA 2002 conference paper.

We may in the future add other indicator variables of occupational representations if they become available.

 

 

As mentioned elsewhere, it is argued to be a strength of the CAMSIS project that male and female CAMSIS scales are not forced to be equal, but instead in each case reflect the relative positioning of occupations within the gender group. Therefore, the CAMSIS approach is ideally only suited to within-gender analyses (also see the comment on the associated 'useofscores' page.

However analyses of CAMSIS scores on cross-gender populations is perfectly defensible through two alternative methods. Firstly, every case could be analysed in terms of the male (or female) hierarchy of occupational locations - reflecting a 'conventional' approach to stratification representations. Secondly, and probably our recommendation, we can take advantage of the fact that the male and female scores have been parameterised to the same population averages, and compute a variable which reflects the male scores for men, and the female scores for women. The interpretation of the scores assigned to individuals on this variable need care - positions reflect relative location within the gender group, and ideally interaction terms with gender should additionally be considered - however the basic interpretation of gender sensitive stratification location is valid.

To achieve either of these workarounds in the LIS and LES studies requires simple manipulations to the matched variables. Using the former strategy, users need only omit the 'gender awareness' recoding command during file matching and proceed with the relevant variable covering all cases. To adopt the latter strategy, users can simply calculate the combined variable, illustrated by the command syntax below following the earlier variable conventions.

(EG: CAMSIS scores are mcam for men, fcam for women, with gender indicated by psex (male=1, female=2).

*******************************************************;

* Example STATA compuation of cross-gender CAMSIS variable;

gen cgencs=-999;

replace cgencs=mcam if (psex==1);

replace cgencs=fcam if (psex==2);

*******************************************************;

***********************************************.

* Example SPSS compuation of cross-gender CAMSIS variable.

compute cgencs=-999.

if (psex=1) cgencs=mcam.

if (psex=2) cgencs=fcam.

***********************************************.

 

 

All works using the LIS and LES studies are obliged to follow the conditions of access to these datasets, details of which can be found here.

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 (Aug 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.