http://www.usaelectionpolls.com/2008/articles/ron-paul-polling-african-americans.htmlOkay it is obvious that Ron Paul has a sizeable lead here but let's look at how the margin of error factors in because we are dealing with a small subgroup. Keep in mind that I do not have the exact sample size of these subgroups so therefore I have to estimate it based on voter turnout in the 2004 election. You can harp on as much as you want as to how that it is flawed to do so; but I have no choice. I rather get some kind of idea as to whether or not this stuff makes sense than do nothing, right?
The 2004 exit polls showed that 11% of voters were African American. That means the Ron Paul vs. Obama/Clinton poll would have had about 132 samples. This has a margin of error of 8.5%.
All of the other polls, having 800 likely voters polled, we estimate there would be 88 samples. Okay, this too has an insane MoE of 10.5%. MoE's find a range in which you have a 95% confidence rating; meaning that we are 95% confident that Ron Paul is polling between 23.5%-40.5% and 95% confident that Giuliani is polling between 5.5%-26.5%.
We could plot a probability density function and detail out the likelihood of a lot of other petty things but it seems kind of cumbersome to do so when we are working with such small samples. In light of these results, however, I would wish that pollsters would take it upon themselves to do a little bit more analysis and polling of African Americans to see what is going on.
One theory that I am suggesting here is that it seems like the less name recognition a Republican candidate has, the better he will poll among African Americans. Bush only had 11% in 2004. Giuliani is polling at 16% then McCain at 20% then Romney at 23.5% and finally Ron Paul at 32%. The contention would be that the less a Republican is known to a black voter, the less they are "against" them.
Are black voters more likely to pick Ron Paul because he is less known? Maybe not! Ron Paul does have some "liberal views" that sets him apart from the Republican field and may help him gain favor among African Americans, such as: being anti-war, against the war on drugs, against the national ID card, believing labor unions have a right to organize, and being against the Patriot Act. You would assume that if a candidate was unknown to a voter that they would just not choose anyways, right?
And to add to that. There was a subsample of African Americans in a New Hampshire primary poll that showed Ron Paul polling in first place among blacks with 22%. The MoE was 18% though!!! (Link).
It is just too bad that pollsters are too cheap to fund an extensive poll on African American voters to see with whom they jive the best because the reality is we feel that Ron Paul really would outperform his Republican counterparts in this demographic.
Interesting numbers, SJ.
Doesn’t impact my laughing at the notion African Americans are going to vote significantly for the white guy from the GOP...or the Libertarian Party for that matter.