In The Limits of Family Influence, David Rowe summarizes some evidence on spousal resemblance:

[H]igh spousal correlation coefficients did not appear to be the result of social influence in the marriage; people who had been married a long time were not more alike in their attitudes than newlyweds.  Initial assortment, rather than influence, thus seems to be the cause of spousal behavioral resemblance.*

This summary inspires Rowe to eloquent wisdom:

[W]e accept quite readily the idea that our spouses are hard to influence; it is easier to avoid an area of divergent opinion than to try to get our wives or husbands to agree.  How intuitive it is that our spouses are hard to change!  Yet the great change that occurs in children does not mean that their direction of change is any more malleable to our wishes than that of our spouses, to whom we also apply pressure by social example and by levers of reward and punishment, but to little advantage.

The practical lesson of behavioral genetics for parents, in my view, is to stop trying so hard to change your kids.  The practical lesson of Rowe’s evidence for singles, in contrast, seems to be that you should hold out for a very close match.  Once you accept that the person you marry is unlikely to “grow into” in the changes that you urge upon him or her, the sensible response is to rely more heavily on the power of selection.

In short, economics informed by behavioral genetics come down on the side of romantic idealism: Show the world your true self – and see who loves you just as you are.

* Incidentally, Rowe understates his case here.  If spouses with divergent attitudes were less likely to stay together, correlations would rise over time even in the absence of causal influence.