Objectives
The "Life Course and Welfare Development" project was designed to provide quantitative information about the life courses of different historical birth cohorts. The aims of the project were, first, to enable more dynamic measurement of social inequality in terms of both cumulative effects across the life course and inequality of living conditions across cohorts. Second, the life course was to be conceived of as a self-contained and causally definable whole, in order that circumstances observed at any given point in life could be explained by reference to opportunities and risks earlier in the life course. Third, the life course was to be regarded as a series of events and role configurations, the sequencing and length of which are standardized, and which are shaped by institutional settings. The retrospective collection of continuous life history data in numerous domains was an innovative methodological approach.
Questionnaire Design
Retrospective data on key events in the domains of social background, siblings, general and vocational/professional education, employment history, residential history, partnership and family were recorded as comprehensively as possible using pencil-and-paper questionnaires (see Table 1). Download English version of the questionnaire
Table 1: Examples of Life History Data Collected on Various Domains of Life
| Domain |
Example Variables |
| Social Background |
Dates of birth and death; absences from biological parents; parents' education and employment; foster parents |
| Siblings |
Siblings' dates of birth and death; education, marital status, and occupation of all siblings |
General Education |
Dates of school entry, school transitions and graduation; qualifications |
| Vocational/Professional Training |
Dates of all training programs; qualifications attained; type and location of training establishment |
| Employment History |
Dates of all jobs; exact job title; responsibilities; size of company; sector; working hours; income; spells of non-employment |
| Marital History |
Dates of marriage, divorce, widowhood (all marriages); age, education, and occupation of partner(s) |
| Children |
Dates of birth and death; education, marital status, and occupation of all children; step and foster children; grandchildren |
| Residential and Migration History |
All moves; household types; sizes of towns/villages; types of dwelling; household members |
| Cross-Sectional Data |
Electoral behavior; political and religious orientations; assets and income; health status; satisfaction with various domains of life |
Codebook in English
Sample Design
The aim was to obtain a representative sample of German nationals born in the years 1929-1931, 1939-1941, and 1949-1951, with a particular view to comparing the life courses of men and women.
The 1929-1931 birth cohort was selected on the assumption that World War II had very adverse effects on the opportunities for training and labor market entry of those born around 1930. The 1939-41 cohort was selected on account of its particular demographic situation as the cohort with one of the highest birth rates.
Because it was not possible to draw a quota sample or a sample from the official registers (ZUMA-Nachrichten 10 [May 1982], p. 22f. ), it was decided to use a sample design based on random walk technique. First, a sufficient number of private households were drawn as a random sample. Second, all persons living in these households who belonged to the target population (i.e., all members of the selected birth cohorts) were identified (see Kirschner & Wiedenbeck, 1989, p. 84ff., pdf format, 83 KB).
Data Collection
The realized sample is very balanced in terms of the absolute numbers of cases per cohort and gender. However, because of deficits in certain geographic areas, additional quota interviews were carried out to achieve a better distribution of the target population across urban and rural areas. This means that the cohort and gender subgroups are no longer quite as balanced. Because only 132 quota interviews were conducted, however, it is unlikely that this change in the recruitment procedure led to much bias in the sample (see Brückner, 1989, p. 123ff., pdf format, 630 KB).
Table 2: Sample Coverage and Reasons for Non-Participation
| |
Birth cohort |
Total |
| |
1929-31 |
1939-41 |
1949-51 |
N |
% |
| Gross sample |
1,257 |
1,289 |
1,148 |
3,694 |
100.0 |
| Neutral non-response |
|
|
|
421 |
11.4 |
| Adjusted sample |
|
|
|
3,273 |
100.0 |
| Refusals |
|
|
|
763 |
23.3 |
| Non-contacts |
|
|
|
265 |
8.1 |
| Other non-interviews |
|
|
|
206 |
6.3 |
| Realized random sample |
681 |
681 |
677 |
2,039 |
62.3 |
| Quota interviews |
28 |
52 |
52 |
132 |
|
| Realized interviews |
709 |
733 |
729 |
2,171 |
|
Representativeness
The quality of the sample was assessed in a number of studies by comparing the distribution of certain socio-structural characteristics in the realized sample with the corresponding distributions in the official statistics (Blossfeld 1987, Papastefanou 1990, Huinink 1988, Allmendinger 1994). These analyses confirmed that the data are characterized by an exceptionally good retrospective reproduction of the social structure. The distributions of most variables correspond quite accurately with those of the official statistics (microcensus). Furthermore, there is no bias in favor of the middle class, i.e., the proportion of respondents with lower socio-economic status, particularly blue-collar workers, in the sample reflects their proportion in the population at large. Because the middle cohort is slightly underrepresented and the youngest cohort slightly overrepresented, it is advisable to weight the cohorts accordingly.
Data Editing
In contrast to cross-sectional surveys, where data cleaning is essentially limited to eliminating undefined codes and correcting filter errors, life history data permit consistency checks with respect to both content and timelines. The data were edited very thoroughly on the individual level to produce a coherent dataset with a consistent structure. Because data correction may adversely affect data validity, however, a targeted and consistent data cleaning procedure is essential to minimize the risk of arbitrariness. Standard procedures for data monitoring and correction were thus adopted (see: Editionsregeln, pdf format, 180 KB).
Notes on Data Analysis
Data were originally held as SIR databases. To facilitate research, they were released to the Central Archive for Empirical Social Research (ZA) in Cologne as public-use files in the form of SPSS data files, SPSS portable files, and STATA files. The dataset can be obtained from the Central Archive for Empirical Social Research (ZA Studies No. 2645).
For reasons of data protection, the public-use files were factually anonymized such that an unreasonable amount of time, expense, and labor would be required to identify individual statistical units. Any direct references to places and all open-ended responses were removed. The original "questionnaire number" was replaced by a new ID number produced by a random generator, such that no direct links can be made between the public-use files and the questionnaires themselves.
Documentation (Downloads, in German)
Karl Ulrich Mayer und Erika Brückner: Lebensverläufe und Wohlfahrtsentwicklung
Konzeption, Design und Methodik der Erhebung von Lebensverläufen der Geburtsjahrgänge 1929 - 1931, 1939 - 1941, 1949 - 1951. Materialien aus der Bildungsforschung Nr. 35. Max-Planck-Institut für Bildungsforschung Berlin 1989
Teil I: Methodenberichte zur Stichprobe, Durchführung und Datenaufbereitung der Pilotstudie und Haupterhebung 1980 - 82
