There is often a lot of talk in a lot of places during election time about knocking on doors, making phone calls and raising money. But the question people have is, "Do campaigns work?"
This was actually tested by the Service Employees International Union in 2008, and the results are pretty darn impressive.
SEIU held back a number of "control" subjects from the voter contact program. This was done through the use of a voter database that pulled out a group of representative voters from 11 states that SEIU had boots on the ground.
The control and experimental groups were around 2500 each (obviously SEIU contacted more than 2500 people). The control group got no contact from SEIU. They were likely contacted by the Obama campaign, progressive allies, the McCain campaign, conservative groups etc. But they were never contacted by SEIU. By contrast, the experimental group was contacted through mail (an extremely limited part of the SEIU program) and through field contacts. That meant door knocks and phone calls.
Each group was polled on some specific questions, including their view of John McCain, their view of Obama, if they thought Obama was good on jobs and if they thought Obama was good on the issue of health care, and lastly who they voted for.
The experimental group's opinion of McCain was 9 points lower than the control groups. Obama's favorables were 5 points higher. He did better on the issues of jobs and the economy. And most importantly, the experimental group voted for Barack Obama at a rate that was 4% higher. That is a pretty significant change in peoples opinions.
In his explanation of the 2008 race, Mark Blumenthal at Pollster.com called this experiment the largest and most comprehensive ever on the Democratic side.
So the next time someone asks you if campaigns matter, the answer is an unqualified YES!