Evaluation of the Child-Ladder Method for Studying Migration Using Family Trees: A Comparison with Linked U.S. Census Data, 1850 to 1920

Maryam Torkashvand, University of Iowa
Caglar Koylu, University of Iowa
Alice B. Kasakoff, University of South Carolina

This article uses a population-scale family tree data set in the U.S. to construct interstate family migration flows in the United States from 1850 to 1920. Did migration improve well being in the US by allowing families to settle their children in farms of their own on cheaper land? Were larger families more likely to move? We use the child-ladder method, which is based on changes in birthplaces between successive births. Although child-ladder provides a more accurate estimate of migration timing, this method can only be used for families with at least two children. To evaluate the representativeness and usefulness of family trees for studying migration, we compare the results to migration data derived from changes of residences in linked household data between consecutive censuses (MLP). Even though the child-ladder information is biased towards large families, the geographic patterns are very similar to the linked census data and demonstrate the usefulness of the child-ladder method in migration research. Moreover, because many families moved more than once during the census interval, this method reveals a more detailed picture of migration paths than the linked census data. However, small families’ migrations are only represented by the linked census data. Both sources of data show that spatial patterns of migration differed by family size with very large families moving greater distances especially in the US North.

No extended abstract or paper available

 Presented in Session 56. Approaches to Studying Migration in Historical US and Japan