inion" (Converse, 1987, pp. S13-14).
By the 1960s, polls became one of the main tools used by candidates to check popular feeling on major issues. In the 1970s and 1980s one of the most significant developments was the entrance of the news media into what was previously a specialized field. "In 1972, no main news organization had a regular poll of its own," said Barry Sussman, former director of polling for The Washington Post. "The chore of sampling public opinion was left mostly to Gallup and Harris . . . fifteen years later almost every elite media had a poll of its own. . . " (Sussman, 1988, p. 238).
The entrance of non specialists into the field of polling created problems of its own namely the use of unscientific methods without adequate sampling and statistical control. For instance, in 1980 ABC News encouraged viewers to call, at a cost of 50 cents, to indicate their opinion on whether Ronald Reagan or Jimmy Carter had won the October 28, 1980, debate. Such random calls are unlikely to be representative and show how polls may be distorted by those asking the questions.
The accuracy of polls is therefore often misleading. Witness, for example, the events surrounding the 1988 election. Even such distinguished journalists as Peter Jennings remarked, "We're not sure what the heck is going on," when reporting the results of an ABC poll that showed Michael Dukakis with a substantial lead over George Bush, even though a few days earlier another ABC poll had shown Bush slightly ahead (Landers, 1988, p. 455).
Polls tell politicians more than who is favored at any given time. They suggest differences in regional attitudes, shifting responses, and demographic preferences. The methods used in collecting polling data are varied, yet even the research design contributes significantly to the interpretation of the results.
...