ISSP 1992 - "Social Inequality II" - ZA No. 2310

Errata:

Stefan Svallfors from the  University of Umea, Sweden made us aware of a problem with the 1992 ISSP data and documentation as distributed from GESIS-ZA. It touches variables V77 to V83 for Sweden in this ISSP 1992 data-set

The Swedish data in the 1992 data file are not "real" ISSP data, since they come from a one-off replication (conducted in 1991) of the 1987 ISSP survey. To the Swedish survey were also added a number of questions from an Australian survey, and since some of these questions also became included in the general ISSP92, it was decided to merge the Swedish data with the 1992 data file.

The problem is that these questions are not strictly comparable between Sweden and other countries, since some amendments and changes were made in the 1992 ISSP questionnaire compared to the Australian one. 

Here is the list of the original items from the Australian survey which we used for the Swedish translation:

In deciding how much people ought to be paid, how important should each of these things be?

1. How much responsibility that goes with the job (included as v77 in the   ISSP92 codebook)
2. How complex and difficult the job is
3. The number of years spent studying (included in modified form as v78 in the ISSP92 codebook)
4. whether the job requires supervising others (included as v79 in the ISSP92 codebook)
5. how much practical experience it takes
6. how old the worker is
7. whether he or she is married
8. whether the worker is a man or a woman
9. whether he or she has children to support (included as v81 in the ISSP92 codebook)
10. how dangerous the job is
11. how dirty and unpleasant the working conditions are

The answering scale in Australia was "The most important single thing", "extremely important", "very important", "fairly important", "not very important", "Should not matter at all"

In Sweden, the answer scale does not contain the "extremely important" category.

In addition, the Swedish item battery contains an item "how much is needed for the family to have a decent life". (which is included as v80 in the ISSP92 codebook, but which is really different)

In the translation of item (i) the Swedish translation runs as "whether he or she has children to take care of" (included as v81 in the ISSP92 codebook; again very different from the real ISSP item)

In summary, there are problems regarding

1. item formulations
2. answer categories
3. context effects

The Swedish data for items v77 - v83 will be taken out of the data file completely. Until that time these variables should not be used for analyses without considering the problems.

(26 Oct. 2001)

POLAND: There is an error in the variable V128 - Urban-Rural - in data and documentation:

Current (wrong) distribution for this variable V128 (unweighted) is:
1. Urban - 595 respondents
2. Rural - 1041 respondents

Correct distribution for this variable (unweighted) should be:
1. Urban - 1041 respondents
2. Rural - 595 respondents

(11 Oct 1999)

A user of ISSP data informed us about some equal respondent numbers in the files produced until now. We publish the list of these cases and leave it to your decision whether these should be excluded from secondary analyses. Most of these cases have different responses and thus they are probably really different respondents:

Great Britain: all cases have equal numbers

(04 Mar 1996)