Do educational opportunities skip generations? Evidence from sibling and cousin correlations in the US
In a new article Professor Kristian Bernt Karlson discusses educational opportunities between generations
Cousins grow up in different households, with different parents, and often in different places - yet they end up with strikingly similar levels of education. Studies from the US, Sweden and the Netherlands have found that up to a quarter of the variation in educational attainment can be traced back to the extended family. This has often been read as evidence that grandparents and the wider family shape children's chances directly, so that educational advantage can skip a generation.
In a new article in Research in Social Stratification and Mobility, Professor Kristian Bernt Karlson tests whether that interpretation is needed. Building on Lundberg's (2020) framework, the article asks whether cousin similarity can arise from ordinary parent-to-child transmission alone: cousins' parents are siblings and therefore resemble each other, and similar parents tend to raise similar children.
The study uses three-generation data from the U.S. National Longitudinal Surveys (NLSY79 and the Child and Young Adult sample). Unlike earlier work, it estimates all the correlations the test requires - sibling correlations in both the parent and the child generation, as well as cousin correlations in the child generation. Combined with established reliabilities of educational attainment, these yield a model-implied cousin correlation that can be compared directly with the observed one.
The two match almost exactly. Cousin similarity in education is fully accounted for by a dynamic first-order Markov process, in which transmission from grandparents to parents was stronger than from parents to children. Once this shift in transmission strength and measurement error are taken into account, little scope is left for any residual grandparental or extended-family effect. The conclusion holds for Black and White respondents alike, even though cousin correlations are higher among White respondents.
Educational advantage, in other words, does not appear to skip generations. It passes down one step at a time - which is reassuring for the many studies that observe only parents and children. The article is part of the SIBMOB project, funded by the European Research Council.