Brad DeLong points to Brad Plumer's post debunking the conservative allegation that liberals aren't talking about the relationship between poverty and family arrangements (specifically, single motherhood). Plumer writes a bunch, but I want to remark to the following section near the end:
One other thing: insofar as the fact of single motherhood itself is actually a "problem" (and I'm not convinced it is, but let's suppose...), there are basically two remedies. One, we can try to reduce the number of divorces by, say, making divorce harder to do, though that seems like a terrible option. Divorce is often very necessary, quite obviously, since even the best marriage counseling can't prevent every unhealthy or violent relationship. . . . [deletia]
So let's look behind door #2. And door #2 is... reducing out-of-wedlock births in the first place. This seems like a pretty unambiguously decent policy goal . . . the tried-and-true way to reduce unintended out-of-wedlock births involves teen-pregnancy prevention programs that emphasize, yes, condoms and other "icky" items. (Hell, they can teach abstinence too, since that seems to work, though "abstinence-only" programs pretty clearly do not.) Measures to reduce subsequent pregnancy, like "second-chance homes" for teen mothers, or home visiting programs, seem to have had some success. Oh, and abortion—which, at the moment, is effectively unavailable to a good number of low-income women. But these are all pretty well-known liberal policy goals, I daresay.
Plumer's obviously speaking to a liberal audience, so these will, indeed, be the two obvious options. But to social conservatives, there are actually a whole array of other options. Social conservatives genuinely believe that the "liberal" media's "glorification" of promiscuity is a significant cause of teen pregnancy. They also blame the decline in government sponsorship and acknowledgment of Christianity. Some of them blame the general climate of "moral degeneracy" stemming from the widespread acceptance of "immoral" behaviors, including, for example, homosexuality. These beliefs imply an array of other policy options: censorship, state-sponsored Christianity, and persecution of homosexuals , to name a few. Liberals, of course, believe these policy options would be ineffective, immoral, and unconstitutional, not to mention laughable. However, if one were (unlike Plumer) trying to convince conservatives, one would have to address and refute these ideas, rather than assuming that "our" policy options are the only ones.
 For those who wonder how homosexuality could possibly have anything do with single motherhood --- it's not supposed to make a whole lot of sense. You have to stop looking for objectively observable chains of causation, and think like a social conservative: when a culture is virtuous, good things happen; when a culture is sinful, bad things happen. Homosexuality is sinful, and single motherhood is bad. Ergo, when a culture accepts homosexuality and other forms of deviancy (pornography, pagan religions, interracial dating, women's suffrage, etc.), single motherhood increases.
Of course, I shouldn't paint with too broad a brush --- for any particular deviant behavior B and bad outcome O, plenty of enlightened social conservatives either don't believe B is deviant, or don't believe that B causes O. And many others believe on a gut level that B causes O, but also feel obligated to cook up some random gimcrackery about a chain of causation (which usually shatters under scrutiny) --- e.g., "If homosexual culture were not perverting all these men, they would stay straight, get married, and be good husbands to all these single mothers." Rarely does this form of conservative thought rear its head with the bare-naked stupidity of Jerry Falwell's statement that the September 11th attacks were caused by "pagans, abortionists, feminists, gays, lesbians, the ACLU and the People for the American Way." But if you poke around in most conservatives' worldviews, you will find this style of reasoning as a background assumption, against which their otherwise inexplicable blatherings become at least more explicable, even though they remain incorrect.