Matrilineal Kinship Networks in the United States 1850-1940

Matt A. Nelson, University of Minnesota

Previous analysis of patrilineal kin propinquity relied on surnames to identify the probability of non-random isonymy as a proxy for nearby kin. Two limitations of this approach include not knowing whether same surname matches were truly kin and losing matrilineal kinship networks due to women’s changing surnames upon marriage. Using linked census data from IPUMS MLP for 1850-1940, this project identifies true kin between censuses and better measures matrilineal kinship networks in the United States. Preliminary results confirm the previous life cycle characteristics of patrilineal kin propinquity and a similar pattern among known matrilineal kin propinquity. While IPUMS MLP is biased towards identifying people who lived near similar neighbors between censuses, the results indicate many of the known links in the same enumeration district lived extremely close. Further, because IPUMS MLP uses a probability score to identify linked persons between censuses, some kin are now identified because of slight surname differences between censuses. Matrilineal kinship networks appear to have undergone a similar decline as patrilineal kin propinquity in this time period, but particular nuances suggest that we should not treat these declines similarly.

No extended abstract or paper available

 Presented in Session 23. Household Structure and Kinship Networks