China Family Panel Studies

• 1. Founding And Launch

  ›

  The China Family Panel Studies project began its work in April 2010. The Institute of Social Science Survey at Peking University launched the initiative to document changes across Chinese society. This biennial longitudinal survey aimed to track shifts in economy, population, education, and health. Researchers targeted households within twenty-five regions of the People's Republic of China. They excluded Hong Kong, Macao, Taiwan, Xinjiang, Tibet, Qinghai, Inner Mongolia, Ningxia, and Hainan from the initial scope. The goal was to capture data representing ninety-five percent of the total Chinese population. A baseline collection period ran from early April through mid-September of that first year.

• 2. Methodology And Sampling

  ›

  Researchers implemented Probability-Proportional-to-Size Sampling with implicit stratification for the study design. Administrative units and socioeconomic status served as the main variables for grouping areas. Local gross domestic product per capita acted as the ordering index for socioeconomic status within each unit. If local GDP figures were unavailable, staff used nonagricultural population proportions or population density instead. The original target sample size reached sixteen thousand households. Eight thousand households came from five independent sampling frames known as large provinces. These provinces included Shanghai, Liaoning, Henan, Gansu, and Guangdong. Each of these sub-samples contained one thousand six hundred households. The remaining eight thousand households originated from an independent frame covering twenty other provinces. This structure allowed for regional comparisons while maintaining national representativeness. Primary Sampling Units were administrative districts or counties. Second-stage units were villages or neighborhood communities. Final selection involved systematic housing unit choices using random starting points.

• 3. Data Collection Modes

  ›

  Field teams conducted face-to-face interviews during the 2010 baseline survey. Computer-assisted personal interview software facilitated data entry in the field. Telephone interviews provided a secondary mode for follow-up contact. Computer-assisted telephone interviews offered another layer of remote interaction. Staff utilized mixed modes to maximize response rates across different demographics. In 2010, the household cooperation rate reached ninety-six point five eight percent. Individual contact rates hit ninety-six point seven percent that same year. Refusal rates remained low at two point six seven percent for households. Strict quality control measures applied to the baseline survey ensured data integrity. Teams used audio record checks and field checks to monitor interviewer behavior. Statistical analyses helped identify irregularities in collection processes. Follow-up visits and historical matching strategies minimized participant loss over time.

• 4. Research Themes And Topics

  ›

  The survey measured economic activities alongside non-economic well-being indicators. Educational attainment levels formed a core variable tracked across all sampled families. Family relationships and dynamics received detailed attention from researchers. Migration patterns and physical health status were recorded for every individual over age nine. Mental health assessments included standardized scales like CESD20 and the K6 questionnaire. The study captured data on debt, employment, expenditures, and financial assets. Land distribution and public transportation usage appeared as specific community-level metrics. Researchers also tracked attitudes toward children, parental ideologies, and social interaction frequency. Household composition changes and retirement planning were central themes throughout the project. Income levels and insurance coverage provided critical context for family stability analysis. The universe of observation covered individuals living within one of twenty-five regions excluding specific autonomous areas.

• 5. Principal Investigators And Team

  ›

  Yu Xie served as a principal investigator affiliated with both Peking University and Princeton University. Zhang Xiaobo led research efforts from his position at Peking University. Tu Ping directed operations while based at the Institute of Social Science Survey. Ren Qiang contributed to the team's leadership structure from the same university. These academic leaders managed the longitudinal tracking of households over multiple years. Their collective work ensured consistency in data collection protocols across different provinces. The team utilized software tools including Blaise Version 4.8 and SAS versions 9.1 and 9.2. Strict measures governed quality control during the initial survey phase. Supervisory checks and interview reviews maintained high standards among field staff. The group designed weights for individual, household, and community level analysis.

• 6. Funding And Institutional Support

  ›

  The National Natural Science Foundation of China provided financial backing for the study. Peking University offered institutional support to sustain long-term operations. This funding enabled the biennial collection schedule running from 2010 through the present day. Resources allowed researchers to maintain a sample size of sixteen thousand households initially. Financial support facilitated the use of advanced statistical software like SAS and Blaise. Grants covered costs associated with follow-up visits and supervisory checks. The partnership between the foundation and the university ensured project longevity. Data cleaning operations included consistency checking and text recoding procedures. Public access to datasets remains available through the official CFPS website. This support system underpins the reliability of findings regarding Chinese social changes.