|
Archiving Data and Documentation of the ISSP at GESIS-ZA
Contents
Introduction
The first ISSP module - Role of Government I - was conducted in 1985 in
six countries. Since then, the ISSP was extended continuously, both in
terms of modules and of member institutions.
From the beginning, the Zentralarchiv, now GESIS-ZA, has served as
official archive of the
ISSP. Since 1997 the ZA is supported by the Spanish ISSP partner ASEP
in Madrid. The central task is to produce international data sets with
appropriate documentation (codebooks). When dealing with international
comparative data, data sets and documentation must be seen as two sides
of the same coin. Even though every effort has been made to produce
comparable data in each country, and even though GESIS-ZA works closely with member
countries to clarify and harmonise deviations from the basic coding
schemes, there still remains a remarkable amount of country specific
peculiarities in the final, international data set. These country specific
peculiarities have to be documented in order to supply secondary
analysts with vital information.
The format of the data and documentation supplied by each member
country to GESIS-ZA should
be as close to the format of the integrated international files as ever
possible. This may sound simple, but experience has shown that deviations
are frequent. And even seemingly minor deviations from the expected
standard cause
time-consuming efforts when combining the national data sets.
Therefore, the main task of this manual of Archiving Rules is to define
standards and formats for data and documentation of the ISSP modules to
be deposited at GESIS-ZA.
Time Schedule and Deadlines
The data and the complete documentation should have arrived at the
archive
by September 1st of the calendar year following the year of the data
collection for the respective module (e.g. data from the ISSP 2003 module are to be
completely archived
by September 1st, 2004 at the very latest). A data set delivered to the
archive can be only regarded complete if it is accompanied with sufficient documentation (coding
information plus study description plus population characteristics plus a
blank questionnaire).
All data sets being archived by this time (September 1st) are
guaranteed to be
integrated into the module. This international file
will usually be made available in a preliminary version at the following
annual ISSP conference. Comments and corrections from ISSP participants
can be incorporated right after the conference. The complete
international data set and full documentation will then be made publicly
available within a few months.
If single countries provide revised versions of data sets which have been archived
before, these revisions can only
be incorporated if delivered within the time-schedule and if the deviations
from previous versions are exactly documented.
Country data sets which are not completely deposited before April 15th
of the second year after the module's year can definitely not be
included in the final integrated file.
This schedule was agreed upon during different ISSP conferences and is
part of the ISSP Working Principles. It must be seen in the light of the actual
experiences. Changes in these time perspectives caused by changing
conditions will be discussed and decided at the ISSP meetings.
Data Quality and Filters
Some of the variables (especially some background variables) are
subject to filtering conditions, e.g. "Occupation of
spouse/partner". It should be documented clearly for which
respondents this question applies: for those married; only for those who
are married and live together; for those who have a permanent partner ...
The questions behind this demand are:
- How far can consistency checks carried out at GESIS-ZA go?
- What is the minimum acceptable level of data quality?
It is obvious that the highest level of consistency should be reached,
because the quality of analyses depends on the quality of data. However, it
is also obvious that each additional bit of quality consumes a
multiple of processing time. The extent of this additional effort depends on how consistent the
original data is and how well the filtering conditions are documented.
What constitutes an ISSP deposit?
In many cases, the ISSP questionnaire is part of a larger survey. However, the data set supplied to the archive should
only contain the variables from the ISSP questionnaire (including the
optional variables of such a module, if they have been asked, and of course
including the standard background variables).
The coding information for the complete ISSP data set (at least:
variable names and labels, value labels and variable positions in the data
set (data list) for each variable) should come together with the data set
and should be all in English. The data should be supplied in SPSS format
(portable file). In many cases, the necessary information is not described
sufficiently by the SPSS 'label information'. In these cases (e.g. open-ended questions, derived variables
etc.), additional detailed documentation is absolutely necessary!
A blank questionnaire as used during the field-work also belongs to the documentation of the study and will be archived as well.
It would be much appreciated if in addition to the mandatory hard copy version
an electronic version (PDF file, exact MS Word printing master of the
questionnaire, or scan image) could be supplied. A complete list with
links to the required document templates can
be found here.
Technical Formats for Data Delivery
The data deliveries can be made on all common media, that is, by CD-ROM, by e-mail to
isspservice<at>gesis<dot>org
or in special cases by FTP upload.
SPSS portable files with full documentation (VARIABLE LABELS and VALUE LABELS for
all variables) according to the SPSS or XML setup
circulated in advance for each module. The ISSP Standard Background
Variables should be coded according to the scheme supplied in the
respective document.
© GESIS Markus Quandt
2008-02-05
|