How likely is it that answering 'don't know' to a polling question might be a sign of intelligence and political motivation?... this neat article about 'either/or' polling questions in the US suggests precisely that. More importantly, it suggests that voters are likely to shape their views - at least in opinion polls asking them to choose between a constrained range of options - by following partisan cues.
So far so good. But if these polls tell us anything about real voter behaviour, it suggests that party positions then become important in structuring public opinion. And given that, according to the polls discussed by Greg Marx, voters are evenly split on the option of freezing or cutting federal spending, the fact that parties are not offering increases in federal spending implies that voters don't want it.
Now obviously poll responses are not the only information we need to look at in order to understand voting behaviour, but in many political systems party leaders' decisions to adopt particular policies effectively constrain what people can vote for. Which makes the increasing similarity of partisan positions on economic policy issues all the more troubling. If no-one offers increased spending, you can't vote for it. So much for the popular will.