Location Location Location - especially when the polls are at stake
Researchers in the US have found that election results can be swayed significantly by the location of a polling station.
Writing in this week's PNAS, Stanford researcher Christian Wheeler and his colleagues analysed the results of the 2000 general election in the state of Arizona. The election included a proposition to raise sales taxes by 10% to raise money for education. Intriguingly, 53% of voters supported the motion, but amongst voters being polled in schools (which accounted for 26% of the polling venues alongside churches (40%) and community centres (10%)), the number in favour rose to 56%, enough to make the difference between a winner and a loser!
Since proximity to a school might affect the demographic of voters the team controlled for this by comparing the votes of pollsters who lived equal distances from schools but either voted in a school or voted at a polling station located in a different type of building. The effect persisted even after taking this into account.
Next the team ran a voting experiment where 327 people were randomly assigned to vote on an educational tax-related issue in one of two locations. One location contained strong images of school environments including lockers and classrooms, whilst the other location was neutrally-dressed and presented images of offices. Exactly the same result emerged with over 63% of voters supporting the educational tax-initiative when they voted in the school environment, compared with 56% of voters polled in the neutral location.
The results show, say the researchers, that voting in schools undoubtedly biases the electorate in favour of educational issues. But they also raise the tantilising question about what happens at other locations and in relation other issues.
"More research is necessary to determine whether these effects extend to other polling locations and ballot measures. Could voting in a church, for example, influence support for gay marriage or stem cell research?"