Kevin Yeh (葉一璋), executive director of Transparency International Chinese Taipei (TICT), found himself in the middle of a controversy when the Berlin-based Transparency International (TI) headquarters released its annual Global Corruption Barometer last month. According to the international survey, Taiwan’s bribery rate is 36 percent, meaning that this rate of Taiwanese respondents said they paid a bribe when dealing with public services and institutions in the past 12 months. The report immediately sparked protest.
The Ministry of Foreign Affairs (MOFA) objected to the findings, citing past data. Looking at the annual Global Corruption Barometers, there is a sharp and inexplicable difference in this year’s bribery rate compared with those of previous years — 3 percent in 2005, 2 percent in 2006 and 7 percent in 2010, said MOFA spokesperson Anna Kao (高安).
The pronounced increase in the bribery rate conflicts with another finding in the same survey: seventy-one percent of the respondents say the level of corruption in Taiwan has decreased a lot or a little over the past two years, she said.
For TICT’s Yeh, also an associate professor at Shih Hsin University’s Department of Public Policy and Management, the truth about the survey may be forever unverifiable.
“As TI didn’t conduct the poll, many questions need to be redirected to the subcontractors, which could possibly embellish their answers,” he told the Taipei Times. But so far, TICT’s requests for information about the poll have yielded nothing very conclusive.
“We don’t have solid evidence to prove that the final Chinese version of the questionnaire we received is the same one [filled out by the respondents],” he said.
Lost in translation?
For Yeh and Kao, the Chinese version of the TI questionnaire that they saw includes serious design errors.
This year’s questionnaire, translated from an English version provided by TI, seems geared more toward the Chinese population than the Taiwanese. For example, the word corruption is translated as fubai (腐敗) instead of tanwu (貪污). In another question, respondents were asked if they knew of an anti-corruption and governance research center in China. This error appeared in a version of the survey that TI headquarters sent to TICT and MOFA. Pressed for an explanation, TI headquarters later said that the questionnaire released was outdated, and that the mistake had been corrected in a different questionnaire that actual respondents saw.
Yeh said that this year’s questionnaire also appeared to include outright translation errors. One question, which in its English original asked respondents whether they or their family members have ever paid a bribe, was translated to ask whether respondents had “witnessed” (目睹) bribery. A second question reverses the structure of a response scale. The original question in English asked respondents to indicate how the level of corruption has changed using a scale of one to five, with one meaning “decreased a lot” and five meaning “increased a lot.” This scale is reversed on the Chinese-language questionnaire. TI headquarters later said they discovered the error and corrected it when analyzing poll responses. According to TI, the survey’s results are correct, even though the question is wrong.
Yeh says that additional errors may have been created by how the survey was conducted — online. “In Taiwan, we usually conduct surveys through telephone interviews because when the elders answer the phone, we can read the questions in Taiwanese,” Yeh explains. “Many people may not have access to or understand online surveys in Chinese. The sample of people can be more representative [of the population] over the telephone.”