META-INFORMATION FOR THE 1991 SOCIODEMOGRAPHIC SURVEY (ES-91)
(For replication of “Fewer Children, Better Futures: How War Shapes Family Choices”)

1. SOURCE AND OWNERSHIP
- Data title: Encuesta Sociodemográfica (ES-91)
- Year: 1991
- Producer: Instituto Nacional de Estadística (INE) – Spanish National Institute of Statistics
- Access point:
  - https://datos.gob.es/es/catalogo/ea0010587-encuesta-sociodemografica-1991-microdatos1
  - https://www.ine.es
- Licensing and conditions:
  The dataset is publicly available for research use upon formal request to INE or through the Spanish Government Open Data Portal. It is released under non-commercial academic use, and the delivery process may involve administrative verification.

2. DESCRIPTION OF THE DATASET
The Encuesta Sociodemográfica (ES-91) is a nationally representative retrospective survey that collected detailed life histories from the Spanish population aged 10 and over. It is a unique source for studying demographic and social transitions in twentieth-century Spain.

- Type of survey: Retrospective household survey
- Reference population: Individuals aged ≥10 residing in private households in Spain in 1991
- Sample size: Approximately 200,000 respondents (including all individual-level records)
- Sampling: Two-stage stratified sampling based on census sections
- Units of analysis: Individual and household

3. VARIABLES USED IN THIS STUDY
The analyses for “Fewer Children, Better Futures” use a subset of ES-91 variables, focusing on fertility, childhood conditions, and parental characteristics.

Category | Variable | Description
Identification | id, hhid, provincia | Individual ID, household ID, province code
Demographics | ano_nac, sexo, ano_nac_madre, ano_nac_padre | Year of birth, sex, parents’ years of birth
Fertility & siblings | num_hermanos, posicion_fraterna | Number of siblings, birth order
Education & early life | edad_escolarizacion, trabajo_infancia | Age at school entry, indicator of working before age 11
Child quality | supervivencia_hermanos | Indicator for sibling survival before age 3
Contextual variables (external) | urbanizacion, PIB_pc | Provincial-level urbanization rate and GDP per capita (from CSO2017-87597-R project)

All variables used are fully anonymized and can be reconstructed directly from the official microdata after access is granted.

4. DATA PROCESSING AND RECODING
The replication code (REPLICATION PACKAGE 3 – analysis.do) performs the following transformations:
1. Recodes individual-level years of birth into birth cohorts (1920–1950).
2. Constructs exposure variables identifying cohorts who started schooling during the Spanish Civil War (1936–1939).
3. Generates binary dependent variables for family size (small vs. large), and indicators for early work experience and sibling survival.
4. Merges contextual data at the provincial level for selected robustness checks.

5. LIMITATIONS ON DATA SHARING
The ES-91 microdata are not redistributable under INE’s data policy.
Researchers must request access directly from INE or through the Open Data Portal. The delivery process may require administrative approval, and waiting times can vary. Because this survey was conducted in 1991, access to the full microdata file may require additional verification steps by INE staff.
The replication materials include the original ES-91 questionnaire (as a scanned image) to help users identify variables and structure before requesting access.

6. SOFTWARE INFORMATION
All analyses were conducted in Stata 17.0 (64-bit).
The provided .do file replicates cleaning, modelling, and figure generation as presented in the article once the ES-91 data are available.

7. CONTACT
For questions about replication materials, contact:
Dr. Dulce Manzano
Institute of Public Goods and Policies (IPP-CSIC)
Email: dulce.manzano@cchs.csic.es