Father Absence and the Welfare of Children

Father Absence and the Welfare of Children  Sara McLanahan  (1999)  Aporia likes to bring back older articles that they believe have been neglected. 

Dividing the children of the NSFH into four groups - those with no family disruption, those who lost a parent to death, those whose parents divorced, and those born to never-married mothers - we find significant differences in educational outcomes. Those whose mothers divorced or never married clearly suffer the most negative effects. Adjusting for the factors that predate father absence and are known to influence school failure, we find that children in these two categories are several times more likely to drop out of school than their peers with intact families. The dropout risk is 37 percent for those with never-married mothers and 31 percent for those with divorced parents, in contrast with the 13 percent risk of those from families with no disruption. Significantly, the risk for children who lost a parent to death is 15 percent, virtually the same as that for children from intact homes. Clearly, children of a widowed mother enjoy economic and other advantages over their peers from households headed by divorced or never-married parents.

Or? Or? What could be another possibility why the children of widowed mothers might do as well as the children of intact homes? Seeing that it is not likely that widowed mothers enjoy that much economic advantage (life insurance, I suppose) and the other advantages are...what, exactly, compared to the divorced and never married?

Maybe it's not what I suspect.  But shouldn't the possibility of (forbidden answer) be considered rather than assumed to not be in operation?

Update: Donna B thinks I'm being too cute by not mentioning my guess and she is probably right.  Genetic differences between the fathers who died and the fathers who were still alive but not present (and the women they married) were not mentioned as even a possibility for the different outcomes. Study after study, what is measured is the environment, even when genetics are a distinct possibility.