The transition to adulthood in a comparative perspective: a research program based on harmonized national panel studies from East Asia, North America, and Western Europe

Projects

The research program is carried out in two highly integrated research projects:

1. Within the international project “Life Course and Family Dynamics in a Comparative Perspective”, the transition to adulthood is studied as part of the comprehensive life course in China and Western Europe. The research project is funded by the German Research Foundation (DFG) and led by Bernhard Nauck (NA164/19-1), Chemnitz University of Technology, as well as Karsten Hank (Co-applicant), University of Cologne; the researcher is Barbara Fulda.
2. Within the long-term research project “Panel Analysis of Intimate Relationships and Family Dynamics (pairfam)”, the German Family Panel, as it is more commonly known, is harmonized in part with panel studies from East Asia, the United States of America, and Europe. The research project is funded by the German Research Foundation (DFG) and led by Bernhard Nauck (NA164/18-4), Chemnitz University of Technology; the researcher is Nicolai Gröpler.

Cooperation

The research program is embedded in several cooperative efforts:

1. Within the project “Life Course and Family Dynamics in a Comparative Perspective”, the research program cooperates with (a) a research team on “Child Development and Education”, located at Beijing University and directed by Yu Xie (Beijing University and Princeton University), (b) a research team on “Middle Life Security and Well-Being”, located at Oxford University and directed by Tak Wing Chan (University of Warwick and Oxford University), and (c) a research team on “Later Life, Intergenerational Exchanges and Well-Being”, located at Erasmus University Rotterdam and directed by Pearl A. Dykstra (Erasmus University Rotterdam). This cooperation aims at providing a comprehensive comparison of the life course in China and Western Europe.
2. Within the long-term research project pairfam, the program cooperates with the research group located at the University of Bremen, directed by Johannes Huinink, on the harmonization of biographical data and life histories. This cooperation aims at a coherent approach regarding the data handling of the German Family Panel.
3. Externally, the research program cooperates with the Generations and Gender Programme (GGP), located at the Netherlands Interdisciplinary Demographic Institute (NIDI), on their “harmonized histories” project, directed by Anne Gauthier. This cooperation is aimed at harmonized data management efforts of the GGP and their international data-harmonization-program, along with providing complementary data to the existing “harmonized histories” data set.

The Transition to Adulthood (TTA)

The research program examines the transition to adulthood through the lenses of the life course approach, which entails considering this transition as a time dependent, multidimensional, and multilevel process or set of processes. The transition to adulthood is a significant phase within the life course and a status passage, which involves several interdependent partial transitions, such as leaving the parental home, the transition from education to employment, entry into an intimate relationship, cohabitation, marriage, and parenthood. This perspective calls for the use of longitudinal data and corresponding methods of data analysis. The transitions are analyzed by means of event history, sequence and panel analysis. The former allows for the analysis of the timing of transitions, sequence analysis allows us to study the order and length of phases, and the latter enables us to conduct analyses of the causal relationships between these transitions and diverse outcomes, such as (mental) health, attitude and preference changes, and intergenerational relationships.

Data Sampling

At present, data from the following surveys have been harmonized for comparative analyses (for a description of the datasets along with an overview of the sub-samples utilized to examine the transition to adulthood in and across the respective countries, please click on the links embedded in the dataset names in the following list):

• the China Family Panel Studies (CFPS),
• the Taiwan Youth Project (TYP),
• the Japanese Life Course Panel Surveys (JLPS),
• the National Longitudinal Survey of Youth 1997 (NLSY97) from the United States,
• the German Family Panel (pairfam)

The following data preparations measures were carried out in order to harmonize the abovementioned datasets:

• Cohort Studies: TYP, JLPS, NLSY97, and pairfam
  • Full use of the complete panel sample in TYP, JLPS and NLSY97 along with the two youngest cohorts (81-83 and 91-93) in pairfam.
• Household Surveys: CFPS
  • Creation of a sub-sample, covering the age range of 12 – 35 in the first wave and capturing all the information for these cases in the subsequent waves
• Creating sub-files which only contain the relevant variables required for the analysis of TTA
  • Internal harmonization across waves (within the datasets)
  • Long- and wide-format
  • Data harmonization (across the datasets)
• Creation of an open-access data file, which contains the harmonized data for all incorporated countries (not yet available)

At this point it is important to note that we are not a data archive organization. Meaning, that we do not provide users with original data; all rights are reserved by the respective data management organizations in their respective countries (to access the original datasets please follow the links embedded in the dataset names in the table below). Instead, we aim to provide an open-access file to the scientific community. This access file will contain information and data regarding our harmonization efforts across the numerous countries and data sources. This file will enable researchers to utilize our harmonized data for their future research endeavors.

Original Datasets

The following is a design overview of the original datasets (to access the original datasets please follow the links embedded in the dataset names in the table below):

Publications:

Fulda, B. E. (2017). The transition to adulthood in individualistic and collectivistic cultures: Prevalence and timing of premarital cohabitation and first marriage in Germany and China. In: Scott, C. L. and Blair, S. L. (eds.). Intimate relationships and social change (Contemporary Perspectives in Family Research, Volume 11). Emerald Publishing Limited: 145-172.

Nauck, B., Groepler, N., and Yi, C.-C. (2017). How kinship systems and welfare regimes shape leaving home: A comparative study of the United States, Germany, Taiwan, and China. Demographic Research 36: 1109-1148.

Fulda, B. E. (2016). The diversity in longitudinal partnership trajectories during the transition to adulthood. Demographic Research 35: 1101-1134.

China Family Panel Studies (CFPS)

China Family Panel Studies (CFPS) is a nationally representative, annual longitudinal survey of Chinese communities, families, and individuals launched in 2010 by the Institute of Social Science Survey (ISSS) of Peking University, China. The CFPS is designed to collect individual­, family-, and community-level longitudinal data in contemporary China. The studies focus on the economic, as well as the non-economic, wellbeing of the Chinese population, with a wealth of information covering such topics as economic activities, education outcomes, family dynamics and relationships, migration, and health. The CFPS is funded by the Chinese government through Peking University. The CFPS provides for the academic community the most comprehensive and highest-quality survey data on families in contemporary China. The CFPS was designed by a team of interdisciplinary researchers led by Professor Yu Xie, jointly affiliated with Princeton University and Peking University, with help from an international academic advisory committee. The Institute of Social Science Survey (ISSS) of Peking University carries out the survey operation of the study.

The design of the CFPS contains the following features:

1. All members over age 9 in a sampled household are interviewed. These individuals constitute core members of the CFPS.
2. All current and future children of the CFPS are also considered core members of the CFPS. Theoretically, a core member can leave the study only through death.
3. Five provinces are chosen for initial oversampling (1600 families in each) so that regional comparisons can be made. The remainder of the CFPS sample (8000 families) is drawn from the other provinces so as to make the overall CFPS sample representative of the country through weighting (except for remote, sparsely populated areas of Tibet, Qinghai, Xinjiang, Ningxia, Inner Mongolia, and Hainan). The CFPS sample represents 94.5% of the total population of China in 2010.
4. The CFPS respondents are tracked through annual or biennial follow-up surveys. All the CFPS families, original respondents in these families, and their children are permanent members, to be interviewed in follow-up surveys. Other co-residing family members are also interviewed.

The sample for the 2010 CFPS baseline survey was drawn with implicit stratification. It was designed to be multi-stage so as to both reduce the operational cost of the survey and allow for studies of social contexts. Each subsample in the CFPS study was drawn through three stages: (1) county (or equivalent), (2) village (or equivalent), (3) household. Interviews were conducted using computer assisted personal interviewing (CAPI) technology, provided by the Survey Research Center (SRC) at the University of Michigan.

The nation-wide CFPS baseline survey in 2010 successfully interviewed 14,798 households, along with 33,600 adults and 8,990 children within these families, in 25 designed provinces, for an approximate response rate of 81%. The major follow up of CFPS in 2012 successfully interviewed 13.315 households, along with 8.620 children below age 16 and 35.719 adults. Data are provided as scientific use files to the academic audience and stored in five separate files for each wave: (1) a “community file” stores relevant context information, (2) a “family file” stores information about the respective family household, (3) a “family roster” defines the relationship between family members and contains basic information about each member, (4) the “children file” stores individual and proxy information about family members of age 0 to 15, and (5) the “adult file” stores individual and proxy information about family members from 16 onwards. Proxy information from adult household members is used especially for family members not living in the respective household and for children of younger ages.

Due to its design as an innovative combination of household- ad family network design and of personal and proxy informants, the age range in the resulting individual data sets range from 0 to 102. Because of the specific research topic of studying the transition to adulthood, it was necessary to select individuals of an age range, within which this transition occurs. To establish international comparability, an observation window from age 12 to age 35 was chosen.

In the first stage, all individuals who were 10 or older in the “children file” from wave 1 in 2010 were selected. This early age was chosen, because the 10 years olds will enter into the observation window from wave 2 onwards. Correspondingly, individuals in the “adult file” in wave 1 were selected up to an age of 35. Although the oldest individuals will have already left the observation window by wave 2, they were included to provide base-line cross-sectional information. Information from the “children” and the “adult” subsamples were combined into one data set. This selected sub-population will be followed in subsequent waves in their transition to adulthood. This technically also includes the transition of family members being treated first as “children” and subsequently as “adults”. Information of not included adults is used, too, but only in so far as it concerns their role as a father or mother of the targeted sample.

The following is a brief description of the sub-sample implemented for examining the transition to adulthood in mainland China (a complete overview of all waves is available upon request).

Taiwanese Youth Project (TYP)

The Taiwan Youth Project (TYP) is a longitudinal study on Taiwan’s youth led by the Family and Life Course Research Group of the Institute of Sociology in Academia Sinica. The project was first launched in 1999. The main goal of this project aims at comprehending the life course of Taiwan's youth by examining the interplay among three significant social institutions: family, school, and community -in order to explore youths’ different experiences while growing up and to identify possible factors causing the varieties in a sociological perspective. With assistance from Academia Sinica’s Institute of Sociology and the National Science Council’s Social Science Research Center, after preliminary surveys, TYP was put into practice between March and June 2000. The sample population is junior high school students from 40 schools (162 classes) in Taipei City/ County and Yi-Lan County. Since the first-wave survey was completed, Chin-Chun Yi serves as the director of TYP.

TYP observes two cohorts as the study subjects: the 1st year junior high students with an average age of 13 along with the 3rd year junior high students with an average age of 15. During phase I, the sampled youth have advanced from their early adolescence into early adulthood. After the end of phase I in 2009, both samples have become young adults of 22 (seniors in college) and 24 (college graduates) respectively. In phase II, TYP studies the development of the sampled individuals into adulthood. For this purpose, both cohorts are merged and surveyed with a uniform questionnaire.

Sampling

Due to the relatively high acceptance rate of Taiwan’s mandatory middle education, TYP includes the entire junior high school 1st and 3rd year students (excludes night schools, tutor schools, and continuation high schools) as the sampling population. For the purpose of having a sampling size large enough for statistical power, TYP collects 1000 students in junior high's 1st and 3rd year each from both Taipei City/ County and 800 students in junior high's 1st and 3rd year each from Yi-Lan County. The total sampling size is 5600 students. Interviews were conducted using either self-administered questionnaires or telephone interviews.

TYP employs a multi-stage stratified cluster sampling method, with the division between county and city as the first stratum and the division between town, county, city, and district respectively as the second stratum. After that, cluster sampling is applied to divide the second stratum into two sub-strata for random sampling. “School” is further used as the sampling unit in the first sub-stratum, while “class” is used as the final unit in the second sub-stratum. The first stage is based on proportional stratified sampling. TYP first divides the area of community into strata according to its degree of urbanization based on RCHSS's urbanized degree scale for town, county, and city as the standard. At this stage, TYP stratifies targeted cities/counties into 2-3 areas according to their urbanization degree, that is, this procedure reduces heterogeneity between sampled units across different strata while allowing sample-specific characteristics to be homogeneous in the same stratum. Furthermore, the number of samples for each stratum which should be taken is determined by the ratio of the student number in each stratum to the total student number. In addition, with the principle of sampling two classes in each school, the number of schools for sampling is determined thereafter. The second stage is based on cluster sampling. The sampling unit at this stage, “class”, enables all classes in each stratum to have an equal chance of being sampled. Proper number of schools and classes are then selected according to the number of students needed for each stratum. Besides, to enable sampled students representative of the students in a particular school’s particular grade, in each school’s 1st and 3rd grade two classes are respectively selected.

The following is a brief description of wave 1 and wave 11 (latest available wave) of the sub-sample implemented for examining the transition to adulthood in Taiwan (a complete overview of all waves is available upon request).

Japanese Life Course Panel Studies (JLPS)

The Japanese Life Course Panel Surveys (JLPS) are a panel study focusing on changes in lifestyles and living conditions during a period of societal transformations in Japan. The first wave of interviews was conducted in early 2007 with 4800 participants. The research project has been funded by the Japan Society for the Promotion of Science, by the Japanese Ministry of Health, Labour and Welfare and by Outsourcing, Inc. It is carried out by the Institute of Social Science at the University of Tokyo. The survey is designed to investigate change in lifestyles and attitudes in the context of demographic and labour market change. It comprises a wide range of questions regarding work, living conditions, attitudes, and socioeconomic status.

The study consists of two panels, with one of them covering young adults (20–34 years old) and the other one covering middle-aged adults (35–40 years old). The participants were selected from the resident registry by means of stratified random sampling. The sample is therefore stratified by age group and sex in order to evenly represent respondents of both genders in each of four age groups (20-24, 25-29, 30-34 and 35-40). Follow-up interviews are conducted annually.

The questionnaire was distributed by mail (PAPI). For the youth survey, 3,367 responses were obtained (response rate: 34.5%), and for the middle-aged survey, 1,433 responses (response rate: 40.4%). Follow-up surveys were conducted each subsequent year. Up until now, six waves of the survey are available for analysis.

The retention rate ranges between 82,5 % (wave 2) and 64,1 % (wave 7). An additional sample of 963 respondents was added to the study in wave 5 to counterbalance the effects of panel attrition on the sample size.

National Longitudinal Survey of Youth 1997 (NLSY97)

The National Longitudinal Survey of Youth 1997 (NLSY97) is a longitudinal survey designed to be representative of youths living in the United States in 1997 who were born during the years 1980 through 1984. Data collection for the study, which documents this cohort’s transition from school to work, has been conducted on an annual basis, but as of 2011 has taken place every second year only. The NLSY97 is sponsored as part of the National Longitudinal Surveys (NLS) program by the Bureau of Labor Statistics, U.S. Department of Labor, and conducted by the National Opinion Research Center (NORC) at the University of Chicago with assistance from the Center for Human Resource Research (CHRR) at The Ohio State University. Additional funding has been provided by the U.S. Department of Defense, the U.S. Department of Justice, the U.S. Department of Education, and the National Institute of Child Health and Human Development. Given the study’s substantive interest in the transition from school to work, the NLSY97 collects extensive information on youths’ labor market behavior and educational experiences. However, a variety of data on their family and community backgrounds is also included in the survey to provide researchers with a wide range of information on contextual factors that may have an impact on these labor market entrants. Other topics include health as well as attitudes and behaviors.

The sample design of the NLSY97 is defined by the following characteristics:

1. 75,291 randomly selected households in 147 primary sampling units were screened by interviewers to identify all youths eligible for the NLSY97. All household residents ages 12 to 16 as of December 31, 1996, were considered eligible. A sample of 8,984 respondents originating from 6,819 distinct households was selected. 1,862 households included more than one respondent.

2. Two samples were drawn: A cross-sectional sample representing the U.S. population born in the years 1980 through 1984, which contains 6748 individuals, and a supplemental sample of the black and Hispanic population born during the same period containing 2236 individuals.

The NLSY97 features a number of different survey instruments. Additional information was collected from one of the youth’s parents in round 1. Furthermore, the Screener, Household Roster, and Nonresident Roster Questionnaire complement the Youth and the Parent Questionnaire providing demographic and relationship information on household occupants and nonresident relatives. A computer-assisted personal interviewing (CAPI) system automatically guides respondents through the questionnaire taking into account their age and previous responses. A set of checks within the CAPI system lowers the probability of inconsistent data. To date, 17 waves of data have been made available to the public. The retention rate ranges between 93 % (wave 2) and 79 % (wave 17).

The following is a brief description of wave 1 and wave 17 (latest available wave) of the sub-sample implemented for examining the transition to adulthood in the United States of America (a complete overview of all waves is available upon request).

German Family Panel (pairfam)

The German Family Panel pairfam (“Panel Analysis of Intimate Relationships and Family Dynamics”) is a multi-disciplinary, longitudinal, multi-actor study for the analysis of partnership and family dynamics in Germany. The study was launched in 2008 with a nationwide random sample of more than 12,000 persons from three birth cohorts. The interviews with the anchor persons, their partners, parents and children are conducted annually. Since 2010, pairfam has been funded as a long-term project by the German Research Foundation (DFG). The project is realized in a cooperative effort by the University of Bremen, the Chemnitz University of Technology, the Friedrich Schiller University of Jena, the University of Cologne, and the Ludwig Maximilian University of Munich and is directed by Johannes Huinink (Bremen), Bernhard Nauck (Chemnitz), Franz J. Neyer (Jena), Karsten Hank (Cologne), Josef Brüderl (Munich I) and Sabine Walper (Munich II). The survey fieldwork for pairfam is being conducted by TNS Infratest Sozialforschung in Munich. The study has a focus on issues of partnership and the family. However, a variety of questions relating to other life domains are also included. The main topics covered by the survey are partnership, parenthood, intergenerational relationships, parenting and child development as well as social embeddedness.

The sample design of pairfam is defined by the following characteristics:

1. For the initial panel wave in 2008/09, 42,000 addresses were randomly drawn from the population registers of 343 randomly selected communities. 12,402 interviews were conducted with respondents from three birth cohorts: 1991-93 (N=4,338), 1981-83 (N=4,010) and 1971-73 (N=4,054). Thus, at the outset of wave 1 the age ranges between 14 and 18 in the youngest cohort, between 24 and 28 in the middle cohort and between 34 and 38 in the oldest cohort.

2. If consent was given, additional interviews were conducted with the anchor’s current partner and – from wave 2 onwards – with the anchor’s parents and one focal child aged between 8 and 15.

3. In order to increase panel stability, a non‐monotonic design with a maximum gap of one wave is employed for anchor persons. Starting with wave 3, respondents who dropped out in the previous wave due to “soft refusal” (no contact, no time etc.), are contacted again in the subsequent wave. Respondents who do not participate twice in a row are dropped from the panel.

 Anchors are interviewed with a Computer-Assisted Personal Interview (CAPI). In case of consent, partners and parents receive a Paper and Pencil-Questionnaire (PAPI), whereas selected children participate in a 15-minute CAPI interview. Starting with wave 4, participants from the child survey who grow out of the age range upon their sixteenth birthday are included in the anchor sample. The retention rate ranges between 74 % (wave 2) and 35 % (wave 9). The fully documented data from the first nine survey waves of the German Family Panel are available at no cost as a scientific use file (current release 9.0) for the global scientific community.

Because of its cohort-sequence design, pairfam is in principle well-suited for studying the transitions. However, given the substantive interest in the transition to adulthood, it was necessary to define an observation window within which the transition to adulthood occurs. Since respondents from the oldest cohort are at least 34 years old at the time of the first interview, they are considered irrelevant for this specific research topic. Correspondingly, the oldest cohort was excluded from analysis resulting in a transition-to-adulthood sub-sample of 8,348 respondents.

The following is a brief description of wave 1 and wave 9 (latest available wave) of the sub-sample implemented for examining the transition to adulthood in Germany (a complete overview of all waves is available upon request).