Author splices the data multiple ways, but ignores one obvious one which is African-American single-parent households (over 65% - [1]), and to a lesser degree Hispanic or Latino households (over 40%). The percentage is also high for American Indians (over 50%) but that has a smaller overall effect because of their population is an order of magnitude smaller.

This obviously has a large number of upstream causes and a large number of downstream effects.

[1] https://datacenter.kidscount.org/data/tables/107-children-in...