In our studies we have not found a direct relationship between identification requirements and voter turnout. We've studied the aggregate level and the individual level for individual voters. We found instead—and this is backed up by the majority of political science theory on voting—that there are better indicators of who's going to turn out, and those are the indicators that drive voter turnout.
Number one is interest in politics, interest in the election. Those who are interested in voting are going to show up; they're going to participate. The cost of identification is a concern. The U.S. identification and Canadian identification systems are very different, but when individuals need to get identification, that added cost could stand as an obstacle in the way of voter turnout. But in our research, we found no evidence that, when controlling for interest in the election, a requirement of having identification keeps voters from showing up for the election.