Opinion Polls are Not Inherently Evil

Opinion Polls are Not Inherently Evil
United Daily News editorial (Taipei, Taiwan, ROC)
A Translation
April 27, 2011

The two major parties are conducting polls to determine who will represent the party in the upcoming legislative elections. The Democratic Progressive Party's presidential poll has led to rumors of "Mother Tsai." Some have preemptively made public the results of opinion polls, provoking a backlash, and accusations that they were trying to mislead respondents in the ongoing polls. In the KMT legislative primary polls, Luo Shu-lei and John Chiang have been embroiled in a controversy over whether "a 0.00 percentage victory is still a victory." The KMT Taipei City Party Headquarters has already decided against changing the party's nomination criteria.

Opinion polls are not inherently evil. But so many irregularities have occurred critics cannot help wondering whether candidates are abusing the polls. Is it really true that "so many crimes and deceptions are perpetrated in thy name."

Polls are merely instruments to measure public sentiment. Their primary function is to "allow public sentiment to reveal itself," to "let the numbers speak." Now however, the two major parties have adopted opinion polls as the main basis for party primary nominations. Prospective candidates have agreed to abide by the poll results. This mechanism should function as a gauge of public opinion. Instead, each time a poll is held, a storm brews. Some people are even using the polls to make trouble. The "Mother Tsai" rumors and the Chiang vs. Luo struggle within the KMT are merely the tip of the iceberg. As we can see, polls may are not inherently evil, but behold the dust that they have stirred up!

How did this all come about? Polls are in theory merely neutral tools, They must be conducted in accordance with scientific guidelines. If polls are be exploited, their credibility will be undermined. But polls are essentially neutral tools. People who make use of opinion polls do not always hold neutral positions. Politicians pay lip service to "respect for democratic institutions." But once they acquire the power to influence the decision-making process, they resort to every means at their disposal. They look for the best way to promote their own interests. For example, the DPP has engaged in an internal tug of war over the use of "Blue excluded polls," "party member polls," and "all people polls." Was the final decision on which of these polls to use really made on the basis of the high-sounding reasons offered? The party is currently conducting an "all peoples poll." Even before the decision was made, rumors emerged. Given her public image, the rumors said, the "all peoples poll" might favor Tsai Ing-wen. Therefore once the resolution was passed, rumors emerged that some members might abandon the party. In other words, the very moment the polling method was chosen, political considerations already began playing a role.

Leave aside falsified data, sampling biases, and other extreme cases. Assume that when polling organizations conduct their polls, they attempt to remain neutral and adhere to the scientific method. Assume that "numbers don't lie." One must still interpret the numbers correctly. It is not as simple as "a 0,00 percent victory is still a victory." Under direct democracy voters personally cast their ballots. During an election, a one vote victory is still a victory. But polls are based on sampling. Out of the main body of the population, only a sample is polled. This sample is used to infer the results. John Chiang and Luo Shu-lei are embroiled in a dispute. A gap of less than one percentage point separates the two. This difference is extremely significant. One must be cautious when making any statistical interpretations.

But how did the KMT deal with this controversy? First it followed the recommendations of its district party headquarters and suspended Lei Luo-shu's rights as a party member. Later it followed the recommendations of its Public Relations Committee Chairman Su Chun-ping, who urged the party to respect the results of the poll. Was the change in the party's position truly the result of "respect for the democratic process?" Was it truly the result of adherence to the scientific method, and the belief that "numbers don't lie?" Or was it the result of other considerations that cannot be stated openly? Outsiders will always find it difficult to see the whole truth. But skepticism is understandable. They may well be exploiting the polls.

Now consider the even uglier rumors, and attempts to influence voter psychology. They may be conspiracies, or they may be acts of desperation. Either way, they are old electioneering tricks. They are undeniably attempts to ensure that one's opponent loses. Such is the nature of politics. But voters on Taiwan are seasoned veterans. They long ago developed countermeasures. The atmosphere of intrigue has deepened. For example, voters have a tacit understanding that they will "allocate votes on their own initiative." When interviewed by poll takers over the phone, they deliberately engage in disinformation in order to undermine rival parties. Anyone attempting to conduct a scientific survey will find himself lost in a fog. Therefore the DPP's "Mother Tsai" rumors and preemptive publicizing of poll results merely highlight the degree to which polls are subject to exploitation.

Polls have their uses. But they can also be abused. Using opinion polls to deceive voters is tantamount to election fraud. It seriously undermines the credibility of the polling process. The myth that the findings of social science are scientifically true has already been shattered. For anyone exploring social phenomena, polls have their limitations. This is even truer if people are deliberately manipulating the polls. Democracy is better than dictatorship. Using opinion polls to determine party primary nominations, is better than having them determined by the KMT Central Committee. But people may come to feel that "so many crimes and deceptions are perpetrated in thy name." If so, then those who play with fire may find themselvess burned.

民調本無罪 何辜惹塵埃
【聯合報╱社論】 2011.04.27

兩大黨都在進行提名民調。結果民進黨的總統民調不但出現了「蔡馬馬」的耳語流傳，且有人搶先公布民調監聽結果，引發大反彈，被指誤導仍在進行中的民調。國民黨的立委初選民調，則在羅淑蕾和蔣孝嚴之間發生了是否「贏○‧○○一也是贏」的爭議，台北市黨部已決定不變提名考量。

民調本無罪，卻狀況這麼多，不免有評論者質疑，是否有人在「玩民調」。難道果真是，「民調民調，多少罪惡假汝之名而行之」？

民調就是蒐集民意的工具而已，主要目的是「讓民意呈現」、「讓數字說話」。現在兩大黨的提名機制既然多以民調為主要依據，黨內有意參選者也同意行禮如儀，本來可以好好讓此機制發揮展現民意的功能。但結果，每次民調上場，總不可能風平浪靜，甚至有人藉民調而興風作浪。這次的「蔡馬馬」耳語和國民黨的蔣、羅之爭，不過冰山一角而已。亦可見，民調本無罪，何辜惹塵埃！

到底何以致此？民調在理論上是中性的工具，施行上亦有科學準則可依循；如果說民調可以「玩」，未免損傷其公信力。然而，民調本是中性的工具，選擇和運用民調的人卻未必站在「中性」的立場。口中稱「尊重民主機制」的政治人物，在有權力可能影響決策時，莫不用盡政治算計，尋求能為自己贏得最大利益的途徑。例如，民進黨內為使用「排藍民調」、「黨員民調」、或「全民調」，角力不只一回，最終決定的關鍵，難道真如表面理由那般冠冕堂皇？目前正進行中的「全民調」方法，從定案前就傳出，對「全選民」之間形象較佳的蔡英文可能較有利，也因此在決議定案後，還傳出了黨員出走風波。換言之，從選擇民調方法的那一刻開始，即可能有政治考量的手已經伸進來「玩」了，此其一。

如果排除掉數據作假、抽樣偏頗等極端情形，並且相信民調機構進行民調時確保盡可能的中立性與科學性，但即使如此，所謂的「數字會說話」，卻還是需要正確的解讀方法，並非「贏○‧○○一也是贏」那麼單純。直接民主當中，選民親自投票，結果贏一票就贏，是沒錯。但民意調查是抽樣調查，在全體選民的「母體」當中選擇受訪者的「樣本」，用以推論出結果。這次蔣孝嚴和羅淑蕾之爭，兩人差距不到一個百分點，這種差異是否真的顯著，於統計上的意義恐須謹慎判讀。

但國民黨對這次爭議的處理態度，從最初的區黨部建議對羅淑蕾作停權處分，到其後由文傳會主委蘇俊賓表示原則上尊重民調結果，其間的轉折對照，真的百分百是「尊重程序」的民主考量？是相信「數字說話」的科學考量？還是未便明言的其他因素考量？局外人永遠難以窺見全部的真相。這種合理的懷疑，可能正是「玩民調」之立論其二。

至於其他更不堪的耳語流傳，以及種種試圖影響選民心理的手段，不管是陰謀論或告急牌，都是選舉的老招數，也都不必諱言具有「企圖使對手不當選」的意味。政治的本質如此，台灣選民身經百戰，早已自行開發出反制之道，加深了爾虞我詐的氣氛，例如投票時向來有「自動配票」的默契，電話民調時則有「故意擾亂敵情」的自覺，令任何科學調查都可能陷入迷霧。所以，這次民進黨的「蔡馬馬」耳語及搶先公布民調之舉，亦不過再度凸顯了民調的一種「可玩性」而已。

「民調堪用直須用」，但用得過度，或藉民調的「詐術」操弄形同「作票」的不當影響力，則是嚴重毀損了民調的公信機制。社會科學裡本已顛覆了「科學調查結果必然為真」的神話，任何人為進行對社會現象的探查，都有其局限；更不用說如果有人存心「玩」民調。本來民主比獨裁好，以民調為提名依據，總比以黨中央的一言堂為依據要好。但如果走到讓民眾覺得「多少罪惡假民調之名而行之」這一步，恐怕玩火的人要先被火傷。