邇來，我國數個主要政黨陸續實施初選民調制度，嘗試結合黨意與民意，藉以提升勝選的機會，但此舉卻也引起相當的爭議。針對各黨實施初選民調的情形，反對者所持之理由可歸納為：其一，初選民調獲勝的候選人並不必然在大選中具備強勢的競爭能力；其二，迄今唯有我國政黨施行初選民調；其三，政黨喪失「選才」的重要權力；其四，使得選風更加惡化；其五，初選民調將使得在職者與頗具聲望者佔盡優勢；其六，破壞黨內團結，且落敗者仍可能執意參選到底。經由分析檢證相關資料，本文以為這些論點似乎均有其不足之處。反觀，初選民調的主要作用在於確保提名人選產生的正當性，使得落敗者無法將其失敗歸咎於黨務幹部。至於無具體反應者比例偏高的技術性問題，本文以為，提高有效樣本數（譬如增至2,400個）應為改善現行初選民調的適切方式。

　　In the past years, major political parties in Taiwan used the opinion survey method to nominate their candidates, attempting to abridge the gap between party preferences and public opinions. By doing this, they expected to nominate electable candidates. However, this move has been controversial, critics charging it with some flaws: 1) those who won in the poll did not necessarily electable in general elections; 2) only Taiwan's political parties adopted the survey method; 3) the political parties lost the power over candidate selection; 4) it worsened the illegal practices of violence and vote-buying; 5) it favored incumbents and high-publicity as-pirants; and, 6) some unsuccessful aspirants insisted on running as mav-ericks, thereby thwarting party cohesion. By analyzing the relevant data, I argue that these criticisms are unconvincing. The major purpose of the survey method lies in ensuring the legitimacy of nominees and in making unsuccessful aspirants no excuses to blame the failures on the party eche-lons. As for the technical problem on high proportions of "no response," I suggest that increasing valid sample sizes might be appropriate to improve the survey method.