Genes certainly play a role in children’s outcomes. The modern economy heavily rewards certain traits, such as intelligence and social skills, and decades’ worth of research demonstrates that these traits (and indeed most human traits) are partly genetic. A child’s chances for a prosperous future, therefore, go up if he has a parent who naturally possesses these traits, and even more if he has two. Some estimates suggest that genes constitute about two-fifths of the reason that rich parents have rich kids.

That still leaves three-fifths, of course, and the environment a child experiences matters, too. According to a massive, disturbing new analysis of cognitive-test scores, in fact, it may especially matter for poor children in the U.S.: Something seems to be keeping these kids from fulfilling their genetic potential. Exactly what in the environment makes poor (or rich) kids end up as poor (or rich) adults is an open question, but there are several major contributors that might help explain the trend.

One is family structure itself. Kids are better off — and, if poor, more likely to achieve upward mobility — when they are raised by two married parents. And illegitimacy is a major fault line between educational groups; again, the college educated are getting married and staying married, while less-educated parents often have their children out of wedlock. In America, unmarried parents typically do not stay together through the baby’s childhood, compounding the effects of assortative mating: Instead of two parents with low levels of education at home, the child gets only one, often with the mother’s subsequent romantic partners as well.

There is also a “parenting gap” at work. Less-educated parents talk and read to their kids less, discipline them more harshly, and so on. Assortative mating means that when a child ends up with one such parent, he’ll likely get another one too, again assuming two parents are even in the picture.