Reconstructing birth histories using linked household data and the 1911 Census fertility survey

Area of Science:

• Demography
• Historical Sociology
• Quantitative History

Background:

• Full birth histories are crucial for demographic analysis but are scarce in historical records.
• Existing methods for reconstructing birth histories have limitations with historical census data.
• Partial birth histories from surveys and censuses lack complete information on all births.

Purpose of the Study:

• To adapt and improve probabilistic methods for reconstructing full birth histories from historical data.
• To reduce missing observations in partial birth histories using linked census records.
• To enhance the accuracy of estimating duration-specific marital fertility in historical populations.

Main Methods:

• Utilizing a probabilistic imputation method to reconstruct full birth histories.
• Applying the method to a sample from the 1911 Census of England and Wales.
• Linking census records from 1881 onwards to supplement partial birth history data.

Main Results:

• The adapted reconstruction method successfully reduced missing observations in partial birth histories.
• Linked data significantly improved estimates of duration-specific marital fertility.
• Estimates were particularly enhanced for women whose childbearing occurred over 15 years prior to the survey.

Conclusions:

• Linked historical census data combined with probabilistic imputation offers a powerful approach to reconstructing full birth histories.
• This methodology enhances the study of historical fertility patterns, especially for long-term reproductive trends.
• The findings have implications for understanding women's reproductive behavior in past populations.