All entries for Monday 28 January 2008
January 28, 2008
The South Carolina primary was interesting – Obama, as predicted on this blog, came out strongly ahead of Clinton. What was less interesting was the breakdown in media outlets after the event. The most hotly debated topics of the primary were racial tensions and Bill Clinton’s involvement.
Bill Clinton loves Hillary?
Bill has been playing the role of Hillary’s attack dog leading up to, and throughout the South Carolina campaign. This allows Hillary to viciously lay into the Obama campaign, without seeming unpresidential. When asked in interviews about this, she can simply put it down to his love for her. The reader may insert a Monica Lewinski related cheap-shot of their own preference at this point. After the primary the media have been banging on about an exit polling statistic that shows 6/10 voters thought Bill Clinton was important in the primary and that most of those people voted Obama.
Based on this statistic many people have claimed that Bill hurt Hillary’s campaign. They ignore some of the relevant details of the statistic. Firstly that Hillary’s average amongst people who felt it was important was 10% higher than her average amongst voters overall, and secondly that one’s likelihood to vote for Hillary was directly correlated with how important they felt Bill was. Look at the statistics for more detail, but for example of the 26% who thought he was ‘Very Important’ 46% voted for her, of the 19% who didn’t think he was important at all – only 9% voted for her.
Now having criticised mainstream media for offering easy conclusions based on fallacious reasoning, I’m claiming that this correlation of Bill’s importance causing votes for Hillary over Obama implies a causality. Come on, I’m not that stupid! I think whats important is that Hillary’s voter increase is better correlated to Edward’s voter dropoff than Obama’s. In other words in attacking Obama Bill didn’t stop people from voting Obama, but he made it clear to voters that it was a two horse race. This allowed Hillary to retain 2nd place. Very few people read the daily tracking polls, but Edwards was coming from behind for a while – as the race row burned on, his momentum seemed to drop. As usual with primary related theories it holds a low probability of being correct and its hard to have ‘smoking gun’ evidence – but at least the theory I have outlined above explains the statistics presented by exit polling, rather pretending to explaing them, whilst actually ignoring them.
Obama is a black candidate?
So far we have seen strong demographic biases in these primaries – as you imagine in a battle billed as ‘black guy vs woman’. Again the message thats being put out by the Clinton’s – and being bought by the international media – is that Obama won because most democrats in the states are black. Since most Americans aren’t black we can write off this advantage nation wide. This completely ignores the fact that 60%, yes 60%, of the voters were women! In both the primaries that he has won Obama has beaten Hillary in terms of female votes. My point is that in order to properly analyse trends we need to very detailed statistics on who voted how. For example, amongst black voters in S.C Hillary had 15 times more voters than Edwards. Is that because Hillary has stronger ties to the black community, or were those people voting for her women impressed by her status? Another trend one could identify from CNN’s statistics is that you are more likely to vote for Obama if you are a regular church goer. Lets extrapolate this and, since Americans are more likely to regularly attend church than democrats and conclude that he is the religious candidate. Again easy answers based on one dimensional correlations provde fallacious. Especially when they can’t explain the qualitative, as well as quantitative results.
And now for something not very different…
When I wrote my predictions for South Carolina I observed that all three candidates had excellent shots of taking the state. This was based on their fundamental appeals to the type of people who vote in democratic primaries in S.C. When the race and female factors are taken into account John Edwards goes from first to third. My point here isn’t that demographics are all important in elections, but in situations where all candidates may seem very appealing to voters having something extra to ensure other candidates get perceived as an ‘other’ rather than related to you helps swing the race in your favour. A very banal conclusion, but at least its one that I can believe in.