During the last thirty years, the use of public opinion polling in American politics has exploded. Practically every day, there is a press briefing in Wa s h i n gton on a new poll. In nearly every contested federal campaign, the candidates spend thousands of valuable campaign dollars on their own polls. Most of the c o u n t ry ' s biggest newspapers and television stations conduct polls re g u l a r l y, as do the networks and newsmagazines.
What does all this mean for democracy? For one thing, it starkly re v e a l s two fundamentally differing visions of how re p resentative democracy should work. In one vision, re p resentatives are elected to give direct voice to the p e o p l e ' s pre f e rences. In the other, re p resentatives serve more as delegates than re p resentatives; they are invested with the trust to exercise their own j u d g m e n t .
Some say that, with the proliferation of polling, we are moving more and more toward the first vision of re p resentative democracy. By this analysis, elected officials are functioning increasingly as instruments of a plebiscite, responding directly to what they perceive as public opinion, using the polls to decide what to believe, what to say, and how to say it. At the same time, we know that voters rarely choose their re p resentatives simply on the basis of issue positions. Rather, most voters choose their candidates by combining an inclination toward one political party or the other with an assessment of the individual candidates' character and values. Issues may symbolize values, but few voters arrive at the polls with a checklist of litmus tests.
This raises a question: If voters treat their re p resentatives as delegates, but if polls mean that representatives respond to the public as if they were instruments of a plebiscite, what are the implications for the kinds of decisions that a re made? This is the first question we explore in this essay. The second is how this will change as polling and communications change in the twenty-first century.