Of course, these numbers (cited often by conservatives) are probably one of the best reasons never to take politically isolated poll numbers too seriously, since black districts in the US elect some of the strongest secular, pro-abortion, pro-gay rights, pro-affirmative action legislators in Congress. (Oddly, as do the socially conservative Catholic districts of Boston)
{For those who hate discussing concrete party politics- stop reading)
Politics reflects the actually existing choices of coalition partners to form governing agreements. For all the scorn of anti-Democrats on this list, the core of the Democratic coalition became increasingly liberal in the last decades as groups of socially conservative working class groups (both blacks and many working class whites) sought alliances with the emerging feminist and gay rights groups. The older racist Southern populists - who were only economically progressive for their white supporters, not the general working class - became Republicans, while the new "conservative" force in the Democratic party were new suburban legislators representing the working class that had stepped up the economic ladder (largely based on New Deal programs). The result, given the defections of the Southern white populists, was a party more culturally progressive than much of its working class base and more pro-union than much of its new suburban base.
In the imagination of many third part progressives, there is a possible independent "working class" party, yet it seems unlikely that in real existing politics, such a group would retain the disproportionately feminist and pro gay rights formation that the Democratic party has evolved into in its clutzy direction. I've been reading a biography of Tip O'Neill -- one of the really decent classic Democrats of the old school who was still broad-minded enough to oppose the Vietnam War relatively early. It is facsinating to see the policies he promoted in his early years - opposing every form of family planning spending and even opposing federal education spending unless the parochial schools got their share - versus the politics he would support late in his career as the party incorporated these new post-McGovern forces. There is little question that the Democratic Caucus he presided over as Speaker in the 1980s was a far more progressive (if smaller) group than the often rightwing Dem machine ruled by Southern barons he first joined in the 1950s.
Polling changes in viewpoints account for parts of the shift, but as the polling on black attitudes - a bedrock of party voting - shows, that is only part of decades of progressive organizing (and yes, counter-organizing by conservative corporate forces) that shaped the operational coalitions of Democratic party governing.
My biggest problem with most third party activism is that is absents a large group of activists from that internal organizing work in favor of an idea that some kind of uniform "Democratic elite" will get a "message" from successful third party efforts, when the reality is one of much more fractured coalitions jockeying for advantage.
A lot of people of the list hate bringing up the muck of the actual diversity of political coalitions and how the ephemera of "public opinion" gets translated into actual politicians in real races in 435 districts around the country, since it's much more fun to pretend there are monolithic party elites betraying its befuddled voting members. An interesting counter to the left version of this is Michel Lind, who argues that Dem elites are betraying the socially conservative base of the working class party and longs for the return to the 1950s version of the New Deal (minus the hardcore racism). I think both versions miss the much more interesting process of long-term coalition building in our particular political system, where the aspirations of different geographic units get shaped through representatives having to negotiate with counterparts in other districts.
It's worth noting that this process in the United States is radically different from the centralized party list systems of Europe, where opinions are aggregated nationally through the parties. This is in sharp contrast to doing so than through local geographic segments in the US through individual politicians. It is the attempt to map national opinion polls onto fractured local races that causes much of the handwringing (and false political assumptions) that lead to seemingly odd results like the original discussion on conservative black cultural viewpoints.
-- Nathan Newman