Tue, Nov 27, 2001 - Page 8 News List

Opinion polls can distort elections

By Lee Ching-hsiung 李慶雄

Opinion polls help political parties understand public opinion and formulate policy platforms and election strategies. There is nothing wrong with the media conducting opinion polls to analyze changes in what the public wants. But political parties are using fake opinion polls to raise their prospects of victory, while candidates are also using polls as propaganda to trigger the "dump-save" effect (棄保效應, strategic voting).

In Taiwan opinion polls have become the most effective tool for influencing election results. When the Election and Recall Law is next amended, legislators should consider whether poll results should be released during election campaigns.

Taiwan learned its public opinion research methods and ways of analyzing elections from the US. But the US has a two-party political culture, and its districts only elect one representative each. Almost every election in the US is a showdown between just two candidates. Because there is no "dump-save effect," most election results match the forecasts. Public opinion polls in the US are fair and scientific. They also have academic research value.

But Taiwan's electoral system is different. Taiwan does not have a two-party political culture. Even in elections for heads of government [president, mayor, county commissioner, etc] where only one person can win, there will be at least three candidates. In legislative elections, candidates from the same party compete with each other for several seats in one district. The resulting "dump-save effect" makes

opinion polls inaccurate and destroys their research value. No wonder opinion polls have become a propaganda tool -- discarded immediately after use.

Multi-member districts, the "dump-save effect" and vote-allocation schemes all explain why opinion polls can influence elections. In the 1995 legislative elections, for example, I led in the polls throughout the campaign in Kaohsiung City's northern constituency but came in seventh in the vote count. It was the biggest shock of the 1995 elections.

Post-election analysis offered three reasons for the disparity between the opinion polls and the election results. First, despite topping the opinion polls, less than 30,000 people, or about 7 percent, supported me -- less than the 35,000 votes necessary to win. When the media released opinion-poll results, they influenced undecided voters, who eventually chose other candidates on voting day.

In this year's legislative elections in Taipei City, candidates will need 4 percent of the votes to ensure victory. A candidate with less support cannot win, even if he or she is on the top 10 "safe list" and suffers no "dump-save" effect.

Second, opinion polls sample more than 1,000 people each. Their margin of error is less than 3.5 percent. In a race with two strong candidates, each candidate might have support of more than 20 percent. The margin of error will therefore have no significant effect on the results. In a multi-member election, however, individual candidates only have single-digit support ratings. If one takes the 3.5 percent margin of error into account, the true picture could be radically different. A 7 percent support rating, for example, could actually mean anything from 3.5 percent to 10.5 percent.

In Taipei City's two legislative constituencies, candidates can not be sure of victory unless they have support ratings of more than 8 percent. Therefore, candidates ranking as low as 15th in polls may still be successful on election day.

Comments will be moderated. Keep comments relevant to the article. Remarks containing abusive and obscene language, personal attacks of any kind or promotion will be removed and the user banned. Final decision will be at the discretion of the Taipei Times.

TOP top