Sunday, January 27, 2008

South Carolina Crosstabs

I continue to take my cross-tab results from the MSNBC page. They are simply easier to read than any other site. I think that the exit polls from this primary are very slightly under-reporting Obama's strength, with less than 1% error, but it is still important to note.

I keep track of the following crosstabs: Party, Sex, Race, Age, Issue, and Population. The reason I chose these (vs. Religion, Income, Political Spectrum, Candidate Qualities, etc.) is because they are easy to measure and relatively inflexible. They also tend to map more directly onto a voters' choice, while other tabs help to explain the particulars of a specific primary.

In my post last week, Crosstabs - Age, I said "I think South Carolina will look a lot like Iowa, with Edwards taking away a significant number of HRC votes, and Obama's showing among AA voters comparable to his margins in Iowa." This has proved true, except that Edwards was much weaker with Democrats than in Iowa, and that most of his lost votes went to Obama. For me, the big, under-reported story of South Carolina is the Edwards implosion. That Obama should win big is no surprise and requires no explanation. The key numbers in this primary are with HRC and Edwards.


The obvious tab for this contest is race because of SC's substantial AA population. Some people are making hay out of the fact that the % of Obama's support among AA voters dropped from 83% in Nevada to 78% in South Carolina. I don't think this is statistically very important. There was a much larger sample in SC, so it may be that the NV sample was not as accurate. As I mentioned above, I think that the exit polls are slightly undercounting numbers, too, so the actual percentage is probably closer to 80%. In any event, this is a non-issue. One percent of AA votes in SC is larger than the total of all AA votes in Iowa.

Obama pulled the level of AA support he was expected to pull. This demographic was 55% of the electorate, approximately 292,000 voters, and so explains all other numbers for Obama. He will never see this kind of vote distribution again.

HRC increased her proportion of the vote with both white and black voters compared to Iowa. She increased her % of black votes (rising to 19-20%, see below) over all other states except Michigan, where Obama was not on the ballot. As I've said in a previous post, I think half of those AA votes would have gone to Obama had his name been there.

Edwards also increased his proportion of the white vote in relation to Iowa, from 24% up to 40%. His slice of the black vote plummeted from 8% to 2%. He had pulled 8% AA votes in Nevada, too, so I think this has to be seen as a true decline. So, why did Edwards' white vote go up so sharply? I think that is explained in the next crosstab.

Party Affiliation

There is a lot of movement in this tab compared to the others, and I think it is the most significant, but will be the hardest to interpret.

Obama took a clear majority of AA Democratic votes, 78%. He also took the clear majority of non-Democratic Black votes. 77%. Hillary polled 20% of Democratic black votes and 19% of non-Democratic black votes. Edwards took only 2% of Dem. black votes and 3% of non-Dem black votes. Edwards simply does not attract minority votes.

With non-black Democratic votes, HRC was the winner at 42%, with Edwards at 35% and Obama at 23%. This is the only time Obama has underperformed his usual percentage among white Democratic voters. Does this mean South Carolina white Democrats are racist? I'm sure some of them are, and probably at a higher level than Democrats veiewed nationally. But I think the better explanation is that HRC attracted her usual percentage of Democratic voters for whom there was no explicit race identification. (White normalizing to "non-racial", etc., etc., race/class/gender studies, yadda yadda) . Obama benefitted from a strong racial identification which tipped the balance in his favor among those Democrats. Thus, the "up for grabs" Democratic vote were for non-blacks who were not already voting for HRC.

OK, I want to say the next part veeerrrryyyy carefully. Edwards got higher than his usual chunk of white Democratic voters. I think two things explain this. First and most obviously, he won SC last time, he's from SC and he has been strong all week. Thus, Democrats who are not HRC supporters and did not have a strong race identification reason broke for Edwards. In essence, he pulled away a good part of the Dem. vote that would otherwise have gone to Obama. Among white Democrats, I would venture that it is less racism than it is annoyance with the Obama campaign so overtly playing race politics. In short, I suspect that Obama could have won a bigger percentage of the white SC vote if he hadn't been making up shit about the Clintons being racists. I also think that, to the degree there was white racism involved (unwilling to ever vote for a non-white candidate), those people went with Edwards. Racists don't tend to be feminists, and they would not have gone with Hillary. In short, Edwards is the blameless beneficiary of bad politics.

Among non-Democrats, Republicans are consistently 3-4% of the Democratic party primary voters. Republicans were most likely to go with Edwards and Obama (43% and 37% respectively) than with HRC. However, a full 20% voted for HRC, which is much larger than she has polled before. This says to me that HRC is running more strongly with Republicans (i.e., her ability to attract votes from the other side is increasing), though I want to see a few more primaries before I say there is a trend. Why might this be? The declining economy, I suspect, and the search for someone who they trust to know what to do about it.

Independents voted for Obama in about the same numbers as they alwsy do (low 40s), so he didn't make any gains there. Clinton picked up more Independents compared to Iowa (26% vs. 17%) , but less than she has polled in NH, MI and NV, where she was in the low 30s. Edwards burst up to 32% of Independents. To the degree that white racism played a part in this primary, I think you see it in the disproportionate percentages of Edwards. Again, this is not something Edwards did, nor did he ever make any appeal to that retrograde segment of the population.


The variations in this tab are, as with the rest, skewed by the racial demographics, but still say intersting things. First, the youth vote (under 30) continues to remain low after the initial good showing in Iowa. It is at 14%, comparable to the Nevada 13%. HRC pulled more of this vote than in Iowa, about the same as in New Hampshire, and less than in MI or NV. Edwards clearly is not appealing to this bracket. The big uptick in turn out is in the 30 - 44 group, though this may be an artifact of how the age groups were sliced. This percentage went up from 19% of the voters to 26% of the voters comapred to Nevada. In other contestws, the numbers were also in the teens, but the age group stopped at 40. Since people are more likely to vote as they get older, taking an age slice off the top of a demographic group can greatly reduce the percentages. There were significantly fewer elderly voters in SC than in NV (26% vs. 36%) and a slight majority of them went for Obama. Edwards got a larger than usual slice of them as well, eliminating HRC's usual dominance of this age bracket.


The really fascinating number to me is that women made up 61% of the voters in South Carolina. Clinton won the same percentage of male and female votes in SC as she did in Iowa - 23% and 30% respectively - keeping up her demographic advantage with women even as the total percentage declined. Obama polled the same with both sexes (54/54), just as he did in Iowa (35/35), showing no gender gap. Edwards polled almost the same percentage among men (23%) as he did in Iowa (24%), but did poorly among women (16%). It was better than his NH and NV showings, but this is clearly a weak point and doesn't look like it is turning around.

The rest of the numbers are not worth going into except to note that Iraq becomes less and less important to voters as the economy worsens, which is not good for Obama, and that South Carolina had fewer urban voters compared to the other contests, though the suburban percentage was comparable. HRC polls about the same whether the voter is urban, suburban or rural, Edwards does worst with urban voters, and Obama does best with urban voters. HRC is slightly stronger with rural and suburban voters than Obama, though he won all three categories in SC.

Take away? Edwards simply doesn't appeal to as wide a range of voters as HRC and Obama do. The only demographic he dominates is white non-Democrats, which is not exactly a winning ticket for the Democratic party nominee. Obama won the AA vote, but may have seriously alienated white voters with the vehemence of his race-baiting. He has a lot of repair work to do, much more than HRC will have. HRC chugged along and actually increased her AA support compared to the earlier contests, all of it at the expense of Edwards. In upcoming primaries, look for further declines by Edwards, with a majority of his losses going to HRC, especially women.


1 comment:

Anonymous said...

Yeoman work on the crosstabs. Nice!