Research Questions

1. How do young men and women succeed in combining their occupational careers with family/partnership in today's increasingly mobile society?
Given the complexity of the interplay between the domains of working life, partnership/family, and residential history, few studies have as yet explored the sequencing and timing of events in these all domains. Family and gender research, for example, tends to focus on the impact of partnership and family on career trajectories (of women in particular). Although this approach does link up occasionally with the residential history (particularly when considering 'accompanying partners or wives'), the relationship between career trajectories, residential and partnership history has not yet been the subject of systematic research. Analyses of the causes and effects of individual regional mobility are rare, and there has been virtually no research on how these factors relate to events within the partnership and family .

2. How does the employment situation of women and their partners impact on childbearing?
The drop in the birth rate in East Germany coincided with the transition from a planned to a free-market economy, i.e., with the restructuring of the labor market and, in particular, a sharp increase in unemployment. It might therefore seem reasonable to assume a causal relationship between labor market developments and the "fertility crisis." The central assumption here is that couples will decide to start a family only in a stable economic situation and that economic insecurity (in terms of unemployment or short-term contracts, for example) will lead them to postpone this decision.

By contrast, the theoretical concepts proposed to explain the decline in the birth rate or the increase in the childbearing age in western Europe focus on women's employment behavior and career aspirations (Lesthaeghe 1992; Becker 1993). These concepts are based particularly on the neoclassical household model, the main assumptions of which are the gender-specific division of labor and the incompatibility of child care and work outside the home. Based on these assumptions, it may be hypothesized that better-educated women have fewer children (Becker 1993) or postpone starting a family during the "career planning" phase (Gustafsson 2001). An interesting question here is whether career-minded women respond differently to unemployment and economic insecurity than women with poor employment prospects in terms of their patterns of childbearing.

3. Can biographical calendars enhance recall in retrospective life history surveys?
In the panel study, numerous contradictions and inconsistencies were found between participants' new responses and the life history data they had reported in the previous wave of the study. As such, this follow-up study also served as a methodological experiment, in which the attempt was made to use the life history data already collected as the basis for the follow-up. Rather than using particular historical events as "anchors" to enhance respondent recall (e.g., December 1989 in the LV-Ost panel survey), we referred to events in the individual respondent's own life course, as reported in the previous wave of the study. This required a survey instrument that lightens the cognitive load on the respondents, and helps the interviewer to identify and clear up any inconsistencies promptly. Biographical calendars and timelines that put insights from cognitive psychology into practice lend themselves to the assessment of event data. Compelling evidence from comparative studies has shown that this kind of approach can be an improvement on traditional questionnaire procedures. A number of questions remain open, however. What is the precise effect of calendars and timelines in the interview situation in terms of gains in accuracy? For which respondents and contents is the greatest improvement seen? How do the costs of this innovation correspond to its yield? To assess life history data in this follow-up survey, insights from methodological and cognitive psychological research as well as from many years of practice at the Max Planck Institute for Human Development were applied using the "LDEX" program. The aim was to support the assessment of complete sequences of events by means of visualization and consistency checks. Because of the innovative approach used to assess life history data in a panel design, the effects with respect to data quality and survey costs are to be evaluated in a split ballot study.

Survey Instrument
In principle, the survey instrument was the same as that implemented in the previous life history studies, but in a 'slimmed down' version. No whole modules were omitted; rather, questions judged to be irrelevant given the objectives of the study were cut from the individual modules. To minimize errors, respondents were reminded of the last episodes they had mentioned in each domain in the previous wave of the study. If these episodes were still ongoing at the time of the follow-up interview, data were first collected on these episodes. If the episodes had since finished, respondents were asked to list the episodes that had occurred since the last wave of the survey in the domain in question. The modular nature of the instrument was upheld throughout, even with respect to the occupational history, i.e., every episode in the relevant domains (school, training, employment, unemployment, maternity leave) was assessed separately. After each of these modules had been assessed, responses were automatically checked for gaps and overlaps, and activities that had previously been forgotten or not yet mentioned were addressed. The "LDEX" program directed the interviewer to problems of this kind, and the interviewer was able to question the respondent specifically about the accuracy of problematic data. During this data checking sequence, it is possible to access the main modules again individually, and to make temporal corrections to existing episodes or new entries relating to activities that had previously been overlooked. The program guides the interviewer through this data checking sequence by providing information about the data recorded thus far (in chart and table form), identifying problematic periods, and displaying questions on gaps or overlaps as appropriate. After corrections have been entered, the tables, charts, and data checking sequence are automatically updated until all problems have been solved.

Schedule and Timeline

May 2004	Evaluation study pretest	(16 from LV-West 64)
May-June 2004	Evaluation study	(600 from LV-West 64)
June 2004	Pretest for the follow-up survey	(16 from LV-West 64)
July-August 2004	Follow-up survey	(all from LV-West 71 and LV-Ost 71)

Targeted Case Numbers

• Evaluation of the Survey Instrument
  Only the addresses of respondents in the LV-West 64 study were used in the evaluation study. The modules on residential history, maternity leave, training, employment episodes, and unemployment were assessed for 300 respondents with and 300 respondents without the support of a calendar.
  
• Follow-Up Study
  108 of the 1,435 respondents in the LV-West 71 sample were not prepared to participate in the panel study, giving a gross sample of 1,327 cases. Assuming a coverage rate of approx. 60%, in the order achieved by previous life history studies, the LV-West 71 follow-up study can be expected to yield approx. 800 usable interviews.
  Likewise, 44 of the 610 respondents in the LV-Ost 71 sample were not prepared to participate in the panel study, giving a gross sample of 566 cases. Again assuming a coverage rate of approx. 60%, the LV-Ost 71 follow-up study can be expected to yield approx. 340 usable interviews.