Designing out racism

One of the issues which I wrote about during the US election campaign was the meme, constantly repeated in the BBC's reporting, of the 'Bradley effect'. This is when people say they will but actually don't vote for the black candidate.

One of the odd issues with this is that in the campaign for California Governor in 1982, which this 'effect' is named after, race wasn't a factor (or at least a trivial one). The reason black candidate Tom Bradley lost was because a supposedly 'anti-gun' proposition turned out larger numbers than expected of rural and small town voters.

That sort of 'wedge' issue has been used elsewhere to bring out the vote, most notably the Republicans who used gay marriage propositions in this way in 2000 and 2004.








In a presentation at the TED conference, the baseball statistician cum political pollster guru Nate Silver of fivethirtyeight.com, who most accurately predicted the US election result, elaborated on the question of race in the elections using the presidential outcomes to draw out social and design implications.

He started off by talking about quite how big of a win Obama had.

Electoral maps between 2004 and 2008 show a profound shift towards blue, or liberal, voting. But there’s a block of states - centered on Arkansas, and roughly following the Appalachians - which voted more strongly against Obama than they did against Clinton. And in Louisiana, roughly 1 in 5 white voters told pollsters that race had been a factor in choosing not to vote for Obama - that compares to roughly 4% in states like New York and California.

This made no difference overall because these are less populated states with less national electoral weight.

There's little evidence of race deciding US elections recently and Silver has statistically dismantled its role.

But Silver turns these stats inside out to ask if racism is predictable.

He looked for relationships between independent variables and racism as an electoral factor and found a strong correlation - low education levels correlate closely with racial-based voting. Highly rural states also showed this pattern, though it’s less dramatic than the educational pattern.

The General Social Survey, asks “Does anyone of the opposite race live in your neighborhood?” And, the answers to this are stratified upon density: In the city, yes. In the suburbs, mainly yes. In rural areas, not nearly as much.

He looked at political affiliation - there are more Republicans in monoracial neighborhoods, but it’s not a dramatic difference. Similarly, there’s not much difference in opinion regarding affirmative action. But a question about interracial marriage gets dramatically different results in monoracial neighborhoods - people in these neighborhoods are twice as likely to support a law banning interracial marriage.

What he gleans from this is that if something is predictable then it is designable.

The goal is to facilitate interaction with people of other races. For example a university-based mixing program, sending students from NYU to the University of Arkansas as a form of cultural exchange.

More dramatically Silver suggested that you need to try to create interracial neighborhoods, to reengineer cities.

Cities designed in the 1970s and 80s might actually have helped America become more conservative under Reagan, he suggested.

He thinks that urban design is hugely important to achieving integration: grids vs the windy streets in many parts of suburbia, where grids are better. At the end of the day, he said cul de sacs lead to conservatives.

This idea also relates to a point made by the new black Attorney General, Eric Holder:

In a speech to Justice Department employees marking Black History Month, Holder said the workplace is largely integrated but Americans still self-segregate on the weekends and in their private lives.

Even when people mix at the workplace or afterwork social events, Holder argued, many Americans in their free time are still segregated inside what he called "race-protected cocoons."

"Saturdays and Sundays, America in the year 2009 does not in some ways differ significantly from the country that existed almost 50 years ago. This is truly sad," said Holder.

Although racism in the UK has a very different history I wonder if mapping answers to “does anyone of the opposite race live in your neighborhood?” correlates to BNP voting?