Sun, Jul 14, 2019 - Page 6 News List

EDITORIAL: Polls should be kept private

Hon Hai Precision Industry Co founder Terry Gou (郭台銘), who is vying for the Chinese Nationalist Party’s (KMT) presidential nomination, on Wednesday questioned the accuracy of opinion polls that showed him trailing fellow KMT candidate Kaohsiung Mayor Han Kuo-yu (韓國瑜). The poll was conducted by the Chinese-language Apple Daily and loosely based on the KMT presidential primary polling method. Gou’s concerns are understandable given that the KMT plans to rely solely on landline interviews.

“Everyone knows that the results of the KMT’s primary polls will be inaccurate, and that shows that the polling method is problematic,” Gou said, adding that the KMT rejected his proposal to include mobile phone numbers in the poll.

On July 3, KMT Legislator Wang Jin-pyng (王金平) also expressed concern about the polling method. In response, KMT poll panel convener Huang Teh-fu (黃德福) said that “everything will be conducted based on the principles of fairness and transparency.”

A total of 400,000 respondents would be randomly selected from a pool of 6,446,977 landline numbers in the Chunghwa Telecom database, and one person from each household associated with the chosen phone number would be contacted, he said.

Gou is correct to question the ability of the poll to accurately represent the cross-section of demographics in Taiwan, given that the about 6 million landlines in the sample base pales in comparison with the estimated 20.1 million mobile phone users in the nation, according to data analytics company Statista.

The KMT also plans to poll only one person per phone number, but one landline might represent a household with two or more people over the age of 20 and each might have a different political view.

An article posted on the Web site Communications for Research on April 19 last year said that landlines are becoming obsolete, “and of the people who do still own home phones, many are of older generations, making the data that comes from phone surveys oftentimes less diverse than other sources.” The article also noted other disadvantages of landline polls, such as the inability to read “vital body cues” that reveal a person’s feelings on a subject. Phone surveys are also constrained by time, and people “don’t want to be interrupted during dinner or bath time or while relaxing,” it said.

There is also the question of just how accurate the results from pre-election opinion polls can be.

The Constitutional Rights Foundation Web site says that while polls can help candidates determine where to spend money and exert effort toward the end of their campaigns, the results of the polls sometimes are not reflected in election results.

This is particularly true in Taiwan, where polls are often conducted by organizations that are biased toward a particular party. On May 17, the results of a poll conducted by the KMT’s National Policy Foundation showed that President Tsai Ing-wen’s (蔡英文) disapproval rating outweighed her approval rating at 56 percent to 38.3 percent. However, a poll released on June 24 by the Taiwanese Public Opinion Foundation showed that Tsai would win a three-way presidential race against independent Taipei Mayor Ko Wen-je (柯文哲) and either Han or Gou.

Looking back at poll results from 2015, the majority of pollsters at the time accurately predicted Tsai’s win. However, it would arguably have been hard to make an inaccurate prediction that year after the Democratic Progressive Party’s decisive win in the nine-in-one elections a year earlier. There was also a lot of social media hype about Tsai at the time. This year, voters are more split on social media, and the attention Han and Gou have received has taken many by surprise.

This story has been viewed 1940 times.

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