Ed.Note: Last week we ran an article by Jon Winslow, MapInfo's Director of Social Research about "Who are America's Swing State Voters?" In it, he discussed how demographics would likely play a role in several undecided states (Ohio and Pennsylvania). Jon kindly assented to a post-election interview to talk about demographics, politics and how some of his predictions faired.
Nora Parker (NP): Jon, thanks so much for following up with our readers.Ohio and Florida went with Bush, Pennsylvania with Kerry.Any surprises there?
Jon Winslow (JW): No, not really.It's remarkable how little surprise there was.I'm not a genius, I just did a lot of research and listened to what people were saying, and what we heard over and over was that those three states were the swing states.Which ever candidate got two would win.So Pennsylvania went with Kerry, Florida went with Bush, and there was Ohio sitting there on Wednesday morning. And when it went Bush, that was that.So we feel very vindicated that the analysis was dead on.
NP: In Ohio, Stark County, the county which you discussed in detail in your article, ended up voting for Kerry (92,295 to 89,859 for Bush), according to figures published on USA TODAY's website.Can you discuss how your analysis of three of Stark's six major neighborhood types with a population of over 10,000 households plays into those results? Here's a broad generalization of how you expected those three neighborhood types to vote (and please correct me if I'm wrong!):
Rust Belt Blues: Bush, because although they are "economic Democrats" they are "cultural Repulicans" and that's more important to them right now.
Quiet Streets: Split - a strong pull for Bush for security reasons ('security moms' think he's stronger on terrorism), and a strong pull for Kerry for social reasons (they are more socially liberal in terms of gay right and abortion).
Empty Nest East: Split - a strong pull for Bush for the overall economy, especially as it related to the value of their main asset--their homes, and a strong pull for Kerry for Medicare and Social Security benefits, given that they're at or near retirement and these benefits are critical to them.
Can you comment?
JW: In our article we supposed that certain counties we split because their populations are split. We said the swing states would come down to swing demographic groups. In the election, the idea held up.Certain groups were strongly biased toward one candidate.In very rural areas, for example, Bush won handily.In urban areas, Kerry did well.In areas like Stark County, the difference was negligible.The populations of the key groups could not agree.Ultimately, despite the fact that Stark favored Kerry, enough of the voters in the swing groups went with Bush to give him the victory.In the Quiet Streets neighborhoods, the slight bias toward liberalism was overcome by the focus on security.In the Rust Belt Blues, the historical preference for democratic candidates based on Union issues, was overcome by Bush's stance on cultural issues and his policy in Iraq.
NP: In Pennsylvania, you focused on Lackawanna County as a county that the campaigns should focus on, stating that even though Bush lost it by 25% in the 2000 election, the demographics were favorable for Bush.On Tuesday, it also ended up voting for Kerry by a fairly comfortable margin (59,305 to 44,562 for Bush), according to USATODAY.com.You focused on the 'Village Americana' neighborhood type there, stating it would be difficult to predict, but might trend toward Bush, with a blend of economic dissatisfaction, feverish patriotism, concerns over Social Security and Medicare, and cultural policy.
So what do you think of the results in Lackawanna?
JW: Hard to say on Lackawana - other than my prediction was wrong! And so was Bush's. Ultimately, the urban neighborhoods (Scranton or Wilkes Barre, I forget which) and their leaning toward Kerry overcame the Village America conservatism.I think the point worth making is that Bush and the Republicans know that this bastion for the Democratic Party is vulnerable.Once upon a time, northeast, industrial towns were certain wins for the democrats.Today, victory is not assured.The investment did not pay off in terms of an outright win in the county, but the 44K votes Bush gained helped him make a strong showing in PA as a whole."
NP: Joe Francica's article about "The Media, Mapping and the U.S. Election - Just Plain Boring" generally gave pretty poor marks to the broadcast media for their use of mapping analytics technology as a communication tool while reporting on the results of the election.As you step back and look at the use of demographic data in the 2004 election, can you comment on the level of sophistication in the analysis you observed within the two major presidential campaigns? Are these folks taking advantage of the data in as savvy a manner possible, or do they still have more to learn?
JW: I'll comment on that in two parts.First, I was kind of thrilled on Tuesday night to go online and see all the data being presented in the form of maps.I was a little disappointed at the level of sophistication of those maps - it seems like it would be a no-brainer to go one more level to give more information.USA TODAY's county-level maps offered a little more insight, but even that didn't totally satisfy me.I think a 3D map [showing terrain] would be great, so you could get a sense of where the population is most dense.But maybe our expectations are too high.
Second, I didn't talk to the people within the campaigns but I get the sense that there are people who are trying to use market planning and sales planning techniques based on demographics.It seems like the campaigns themselves "get it" but I think it would help them in the future to use enterprise-wide mapping and web mapping to push the information out. It's the same sort of thing we would tell a yellow pages sales force, for example, that they need to have a person at headquarters preparing the information, but making the information available to the team in the field.