The article eventually points out that pets seem to be a correlation and not the cause of people supporting one candidate or the other. Any kind of survey is hard to run in a way that really allows good statistical analysis. I'd love to see the analysis run in a way that allows them to try to identify how much of the correlation ties only to pets and how much ties to other things. The article suggests (and I agree) that much of the correlation is just that people in a certain place in life tend to have pets and tend to vote for Republicans. That place in life is a cause of both outcomes rather than pet ownership pushing someone to vote for John McCain.
Ultimately, I'll vote for John McCain because the alternative is so horrible, but I don't like John McCain at all politically. However, the fact that he owns a couple of exotic pets makes me like him a little more as a person. I'm not sure that his owning these pets helps me at all in any practical way, but I like him more.
Bill