Two political scientists at the University of California at San Diego, Cheryl Boudreau and Mathew D. McCubbins, wondered if citizens made better decisions, on any number of topics including voting, if they were exposed to polling information on that topic.
They've just published a paper on the issue “The Blind Leading the Blind: Who Gets Polling Information and Does it Improve Decisions?” in The Journal of Politics (Vol. 72, No. 2, April 2010, pp. 513-427)
The short answer to that question, according to Boudreau and McCubbins, is: No. If you are undecided about a given issue, exposure to polls could lead you to make a choice that would actually be not in your best interest. Here's how they phrase it in their abstract:
We find that citizens are more likely to obtain polls when the decisions they must make are difficult and when they are unsophisticated. Ironically, when the decisions are difficult, the pollees are also uninformed and, therefore, do not provide useful information.Wealso find that when polls indicate the welfare-improving choice, citizens are able to improve their decisions. However, when polls indicate a choice that will make citizens worse off, citizens make worse decisions than they would have made on their own. These results hold regardless of whether the majority in favor of one option over the other is small or large.
They expand on what, to me, is a depressing thesis in the article:”… subjects are more likely to obtain polls when the pollees are least likely to help them, and they consistently follow the recommendations of small, uninformed majorities.”
The researchers come to their conclusions after designing an experiment in which one group was given access to poll results while another group was asked to make decisions about a given issue without the benefit of polling information.
… when a majority of pollees recommends the incorrect answer, subjects who receive polls make significantly worse decisions than subjects in the control group. These effects occur regardless of the size of the majority recommending one answer over the other. Thus, the majority’s opinions about the correct choices can cause subjects to make incorrect decisions, even when the majority is not very large. The consequence of these findings is that subjects who receive the polls in each treatment group make decisions that are no better than the decisions of control group subjects.