Could graduate weddings widen the inequality gap?
It's not often that books about university get people so irritated.
But a book published in the US, Marry Smart, has stirred up controversy with the idea that women at university should be hunting for a husband.
It's been attacked and parodied for what its critics say is an outdated, 1950s view of women. And it's been bashed for a snobbish view that non-graduates are not marriage material for a clever woman.
The writer of this provocative advice book, Susan Patton, was labelled by The Times as "probably the most hated author in America right now".
But what if more graduates really are choosing to settle down with other graduates?
An intriguing and rather uncomfortable piece of research published earlier this year suggested that all-graduate couples were a driving factor in widening social inequality in the US.
We've become so used to thinking of education as an engine of social mobility that it's hard to envisage the idea of widening university access as being a mechanism for social division.
The study, produced by academics in the US, Germany and Spain for the National Bureau of Economic Research, shows evidence of how graduate wedding bells are linked to a widening income gap.
Looking between 1960 and 2005, it found an increase in so-called "assortative mating" - where people seek out a particular trait in potential partners.
And the "assortative mating" taking place was social selection according to levels of education.
In 1960, it was much more usual for couples to cross educational divides - a graduate might have settled down with an early school leaver.
In 2005, graduates were more likely to marry other graduates.
There are many reasons why that might be the case, including an increase in overall graduate numbers and women's wider career horizons.
But whatever the reason, the study found that it is increasing income inequality. It concentrates advantage. It amplifies a social divide, with levels of education strongly connected with a polarising jobs market.
In 1960, a US couple who left education after completing high school could expect to earn around the average household income. And if a high-school drop-out married a graduate, they could still expect to earn above average for a couple.
But by 2005 this had changed. A couple of high school graduates had slipped downwards to below the average household income. And the high school drop-out married to the graduate were also below average earnings.
The real winners have been the graduate couples. The rewards have never been greater for these power couples. And for a couple where both are postgraduates, earnings are more than double the average household income.
The study concludes that if people in 2005 had the same "mating patterns" as in 1960, which means marrying with much less regard to education, there would be a "significant reduction in income inequality".
So fewer graduate couples would mean less inequality? Or the irresistible law of unintended consequences?