[EB] Problems Matching respondents of separate EB-files wit

[Meinhard Moschner]

Dear Otto Hermkens,

probably Nero already informed you about the following since he
has been in contact with me about your problems...

> I am a student of sociology presently employed at the Social and
> Cultural Planning Office in Rijswijk, Holland. I am working on my
> thesis, and using Eurobarometer datasets (EB's). My research focuses
> on trends in environmental concern. Therefore I am interested in
> using several EB's. I am trying to match environmental concern
> variables from separate EB-files (EB 5,6,10,20,28, 31a,35,36,40) to
> the cumulative EB-file (1970-1992), so as not to have to harmonize
> all demographic variables myself, but profit from earlier labor by
> the constructors of the cumulative file. The problem is that I am
> meeting serious matching problems:
> 
> (1) Duplicate respondent numbers in various files: eb 5,6 and 20. In
> EB 20 this problem can be solved by matching on several other
> variables as well, so it looks like different respondents with the
> same respondent number for eb20, but for eb 5 and 6 the respondents
> with the same respondent numbers seem to be complete duplicates (at
> least on all variables in the cumulative file), suggesting
> old-fashioned weighting by card duplication (a suggestion by Nero
> Matthias of the Steinmetz-archive).  

This is indeed old fashioned weighting. As far as I know French rural 
cases have been duplicated in EB 5 and 6. So I don't see a problem
here.    

> (2) Different respondent numbers between the separate EB-files and
> the cumulative file: this seems to go for EB's 20,28,31a, and for eb
> 36 for greece only. For all Eurobaromters, separate files have the
> same number of respondents as the cumulative file.

Several years ago I faced the same problem - although in those
days only up to Eurobarometer 21. I will send you the report
that I produced in respect. In short: respondent id problems are
not limited to the surveys that you mention ...

> My question to you is the following:
> Does any of you know of the above mentioned problems, and of a
> solution? I would be very pleased to hear from you.

More via direct e-mail
A nice weekend for all list members
Meinhard Moschner 


+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 Zentralarchiv fuer Empirische Sozialforschung
 Meinhard Moschner
 Bachemer Str. 40
 D-50931 Koeln  
                                            
 Tel.:  (+49)-(0)221-47694-21                              
 Fax.:  (+49)-(0)221-47694-44                          
 e-mail: moschner at za.uni-koeln.de
 URL: http://www.za.uni-koeln.de
 URL: http://www.social-science-gesis.de

 Eurobarometer information: http://www.za.uni-koeln.de/data/en/eurobarometer/index.htm                                             
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++