Full text: Wehner, Sigrid: Exploring trends and patterns of nonresponse

5. Final Discussion 
This thesis examined patterns of nonresponse and discussed strategies of bias corrections 
using the nonresponse study of the East German Life History Study. At the end of the analysis 
of the nonresponse study, I will repeat and discuss the general findings. 
In chapter 1, I pointed out the conflict between the ideal theoretical random sample 
and the practical world of surveys. Several inevitable sources of error always influence the 
quality of survey data. One of these influences is the nonresponse problem, the guiding 
subject for this thesis. Mainly the sample drop-out in the initial phase of sampling is a 
problem, as usually no individual information about the lost persons is available. The 
concentration on this topic emphasises its importance. One should be aware, however, that the 
nonresponse problem must be seen in a sequence of research steps. The careful control of 
other conditions (e.g. the sampling design or the formulation of the questions contributes to a 
high quality of the survey data. 
In chapter 2, I introduced the nonresponse study of the East German Life History 
Study. The accent of the chapter was the data exploration. I presented comparative 
descriptions of the nonresponse study, and I will briefly summarise the main results. 
First of all, it has to be underscored that the nonresponse study in principle is biased, 
as most of the refusals are missing again. This constraints generalisations about refusing 
behaviour (the overall refusal rate in the study is about 70%). Predominantly, one can assume 
to have information about hard-to-contact persons and some of the "soft" refusals. However, 
the big advantage of the nonresponse file is to have collected individual data using almost the 
same interview instrument as the main study did. 
The cohort design of the German Life History Study suggests a separate data 
inspection per birth cohort. It turned out that in fact the characterisations of the NRS persons 
differ by cohort. On the one hand, this can be interpreted as age-specific: persons of birth 
cohorts 1930, 1940, 1950, 1960 were about 60, 50, 40, 30 years old in the year of the initial 
sampling. On the other hand, the cohort differences reflect a different impact of historical 
situations during the life course. 
In a one-dimensional comparison, the NRS data is described by an extreme loss of 
birth cohort 1930, a loss of (older) men, and a higher proportion of single persons. In addition 
to these results, variables indicating a stable social position (better education, stable careers, a 
higher proportion of party membership for men) describe differences for the older cohorts.

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.

powered by Goobi viewer