Written and maintained by CASRAI Editorial Board
Last updated
The Panel Study of Income Dynamics (PSID) is the world’s longest-running household longitudinal survey. Directed by the University of Michigan since 1968, it provides empirical data for studying intergenerational mobility, poverty dynamics, and life-course development. Researchers using secondary survey data must navigate its genealogical design, cross-wave file structures, and data use rules to conduct valid analyses.
What It Is
The Panel Study of Income Dynamics (PSID) is a nationally representative longitudinal household survey in the United States. Begun in 1968 at the Survey Research Center within the Institute for Social Research (ISR) at the University of Michigan, the study evaluates family income, poverty transitions, and economic well-being over time. Instead of drawing a new cross-section at each wave, the PSID tracks original sample members and their descendants continuously.
What is PSID data?
PSID data consists of longitudinal survey records covering more than five decades of socioeconomic and demographic metrics. Collected variables span earned and unearned income, wealth, domestic expenditures, labor market participation, health conditions, and family composition. The public archive provides family-level files (reporting household income aggregates, housing costs, and food expenditures) and individual-level files (tracking individual wages, work hours, education, and relationship status).
The core innovation of the PSID is its genealogical sample design, governed by explicit following rules. When children or other household members from the original 1968 sample leave home, marry, or form independent households, they are designated as split-offs. The PSID interviews these split-off individuals as independent family units while continuing to interview the parent household. This genealogical rule enables multi-generational tracking across up to five generations within the same family line.
Interviews were conducted annually from 1968 to 1997, shifting to a biennial schedule in 1997. To evaluate how panel tracking compares with other longitudinal cohorts, consult our guide to longitudinal study design.
Who Runs It and What It Covers
The PSID is administered by the Survey Research Center (SRC) at the Institute for Social Research (ISR), University of Michigan. Data collection is funded primarily through research grants from the National Science Foundation (NSF; awards SES 1157698, SES 1623684) and the National Institutes of Health (NIH; grants R01 AG040213, R01 HD069609 via NIA and NICHD).
The baseline 1968 sample included approximately 4,800 families (over 18,000 individuals) across two components: an equal-probability national cross-section (~2,930 families) and an oversample of low-income families (~1,872 families) from the Survey of Economic Opportunity. Genealogical expansion and post-1968 immigrant refreshers (added in 1997/1999 and 2017) expanded the panel to over 11,000 families and 26,000 individuals.
Core measurement domains include:
- Income and Transfers: Wages, salaries, business income, pensions, Social Security, and public assistance.
- Wealth and Assets: Housing equity, vehicle values, business equity, transaction accounts, and debt burdens.
- Consumption Expenditures: Housing, utilities, food, transportation, childcare, and healthcare costs.
- Health: Chronic illnesses, physical limitations, activities of daily living (ADLs), and insurance coverage.
- Supplements: The Child Development Supplement (CDS), Transition into Adulthood Supplement (TAS), and Childhood Retrospective Circumstances Study (CRCS).
Public-use files, codebooks, and questionnaires are free to registered researchers through the online data repository. Sensitive variables carrying re-identification risk—such as geographic identifiers below state level and administrative health linkages—require an institutional Data Use Agreement and IRB approval.
How to Search or Use It
Researchers extract customized subsets of variables across survey waves rather than downloading monolithic multi-decade files. The primary extraction platform is the official PSID Data Center. Follow six steps to build and download an extract:
- Locate Sign-in and User Registration: Visit the official PSID portal to register a free profile for data access.
- Open the Data Center: Select the Data Center from the main navigation to open the variable extraction engine.
- Select Variables via Built-in Pathways: Choose among five selection paths: Variable Search (keyword query), Cross-Year Index (topical categories), File (annual rosters), Variable List (direct paste of codes), or Previous Carts (saved extracts).
- Assemble the Data Cart: Add chosen variables to the Data Cart, including annual family interview identifiers, Person Number (PN), and sampling weights.
- Configure Relational Linkages: For intergenerational analyses, use the Family Identification Mapping System (FIMS) to generate parent-child or sibling link files.
- Extract and Download Data Files: Check out the cart to download raw ASCII data alongside setup syntax for Stata, SAS, SPSS, or delimited formats to import and label variables.
Data curation practices adhering to FAIR data principles ensure analytical workflows remain transparent and reproducible.
Strengths and Limits
Strengths
- Intergenerational Depth: Tracks mobility and health transmission across up to five generations.
- Broad Scope: Simultaneous measures of income, wealth, consumption, and health capture responses to shocks.
- Historical Continuity: Five decades of panel data provide consistent baselines for fixed-effects models.
- Linked Supplements: Modules connect early childhood investments directly to adult economic outcomes.
Limits
- Panel Attrition: Cumulative sample loss requires longitudinal attrition weights to retain population representation.
- Biennial Interval: The 1997 shift to biennial waves limits high-frequency transition analyses.
- Confidentiality Top-Coding: Extreme income and wealth values are top-coded in public files to protect respondent privacy.
- Complex File Structure: Linking family and individual records across waves requires careful key management.
Alternatives
The table below compares the PSID against primary alternative longitudinal panels:
| Dataset / Platform | Sponsoring Organization | Sample Unit and Cadence | Core Focus | Primary Difference from PSID |
|---|---|---|---|---|
| Panel Study of Income Dynamics (PSID) | University of Michigan ISR / NSF / NIH | Households and split-offs; annual (1968–1997), biennial (1997–present) | Intergenerational income, wealth, health, and family dynamics | Longest-running genealogical panel tracking split-off descendants across generations. |
| National Longitudinal Surveys of Youth (NLSY79 / NLSY97) | U.S. Bureau of Labor Statistics (BLS) | Individuals in specific birth cohorts; annual to biennial | Labor market dynamics, education, training, and career progression | Focuses on specific age cohorts rather than genealogical whole families. |
| Survey of Income and Program Participation (SIPP) | U.S. Census Bureau | Continuous household panels; monthly tracking over 1 to 4 years | Government program participation and short-term income dynamics | Provides monthly tracking over short panels rather than multi-decade family histories. |
| Health and Retirement Study (HRS) | University of Michigan ISR / NIA | Individuals aged 50+ and spouses; biennial | Aging, retirement transitions, cognitive health, and medical spending | Focuses on older adults rather than life-course and child development. |
| Understanding Society (UKHLS) | University of Essex / ESRC | UK households and split-offs; annual | Socioeconomic transitions and health in the United Kingdom | UK geographic focus with a larger initial sample (~40,000 households). |
For cross-sectional attitude trends, researchers often pair panel data with repeated cross-sections from the General Social Survey (GSS) or historical designs covered in our overview of retrospective cohort studies.
FAQ
What is PSID data?
PSID data is a longitudinal household survey produced by the University of Michigan tracking income, wealth, employment, health, and family composition across multiple generations in the United States from 1968 to the present.
How to cite PSID data in academic publications?
The Survey Research Center requires researchers to cite the dataset in all publications. The standard citation formula is:
Panel Study of Income Dynamics, public use dataset [or restricted use data]. Produced and distributed by the Survey Research Center, Institute for Social Research, University of Michigan, Ann Arbor, MI [insert download year].
Manuscripts using NIH-supported modules (such as the CDS or health supplements) must also comply with the NIH Public Access Policy by obtaining a PubMed Central reference number (PMCID). For persistent digital citations, refer to our overview of CrossRef DOIs.
How to pay PSID payment?
In billing, tax systems, and public administration, “PSID” refers to a Payment Slip ID or Consumer ID generated on an electronic invoice. It has no connection to the University of Michigan research dataset. To submit a PSID payment, enter the voucher number directly into your bank portal, mobile wallet, or government agency payment system.
What is a PSID on a Samsung TV?
On Samsung televisions and consumer electronics, “PSID” stands for Product Serial ID, Persistent Storage Identifier, or Advertising Service ID embedded in device firmware. It is used for hardware management and device telemetry, unrelated to academic social science surveys.
Is PSID the same as psi?
No. In physics, “psi” denotes pounds per square inch, an imperial unit of pressure. In mechanical engineering, “psid” stands for pounds per square inch differential, measuring the pressure delta across a barrier. In social science, PSID refers to the Panel Study of Income Dynamics.
What is the difference between family-level and individual-level files in PSID?
Family-level files contain one record per responding household per wave, capturing aggregate measures like total family income, housing costs, and food expenditures. Individual-level files contain one record per person, capturing personal wages, work hours, education, and genealogical relationships.
Sources
- Survey Research Center, Institute for Social Research, University of Michigan. Panel Study of Income Dynamics Main Documentation and Study Overview. https://psidonline.isr.umich.edu/
- University of Michigan ISR. PSID Data Center User Guide, Extraction Tools, and Variable Selection Methods. https://psidonline.isr.umich.edu/Guide/
- University of Michigan ISR. Guidelines for Citing PSID Data and Acknowledging Federal Grant Support. https://psidonline.isr.umich.edu/Guide/HowToCite.aspx
- National Science Foundation. Division of Social and Economic Sciences. Award Abstract #1623684: The Panel Study of Income Dynamics. https://www.nsf.gov/
- U.S. Bureau of Labor Statistics. Longitudinal Surveys and Demographic Household Panels Reference Overview. https://www.bls.gov/
