民主政治運作過程中，當主政者無法滿足選民需求時，選民理所當然會以選票表達其自主性，找一個更符合其需求的政黨作為治理國家的代理人（總統）。政黨輪替就成為選民選擇最佳代理人的手段。政黨輪替的本身，就是一種權力交替與權力制衡，特別是在執政黨交不出政績的時候，政黨輪替就是一個民主的制裁機制，更是一個國家走向民主政治的重要指標。 影響選民決定更換治國代理人的投票抉擇因素，主要有政黨因素、候選人因素，抑或是議題因素。無論在美國與其他西方國家，甚至在已開發或開發中國家都已經獲得印證。當選民利用參考標的進行投票，選民的總統選舉投票結果將會匯集成決定最後政權誰屬。把票投給執政黨的候選人，還是把票投給在野政黨的候選人，不僅攸關個別政黨與候選人的選舉成敗，而且還影響到統治權力歸屬，甚至會牽動政黨輪替執政的可能性，其重要性，不言而諭。也正因如此，政治學者亟思理出選舉各黨勢力消長的軌跡，試圖描繪出選民投票抉擇的決定如何影響政治權力的轉移。 本文的目的，是研究選民投票抉擇的一般原則應用到總統選舉結果「政權轉移」或「政黨輪替」這個重要的主題，針對可能造成多數選民由投票支持原有執政黨轉而投票支持原有最大在野黨取得執政權的成因進行討論。台灣與墨西哥這兩個長久以來一黨威權獨大的政權，卻同時在西元二○○○年發生政黨輪替的現象。有關選民選擇支持原有最大在野黨取得執政權的成因，主要從三個面向進行討論： 首先，政黨認同因素，選民政黨認同發生改變所引發的政黨認同重組現象，讓原本獲得多數選民支持的執政黨，因為政黨重組的因素變成僅吸引少數選民支持的在野黨；相反地，原本僅有少數選民支持的在野黨，因為政黨重組的關係變成吸引多數選民支持的執政黨。其次，候選人因素，候選人人格特質、能力與對候選人情感溫度等因素，讓多數選民受候選人這個短期因素所影響而選擇支持最大在野黨執政。最後，議題因素，選民的展望性評價與重大議題中的改革議題，因選民希冀更換原有執政黨，由最大反對黨執政，能夠改變趨於日益不振的經濟現況，並透過大幅改革能改變舊政府的缺失，期望新政府能夠振衰起弊。 / In the democratic operation process, when the director is unable meet the citizens need, the citizen would express their autonomy by votes. Looking for one (or political party) who could conform to the citizens demand takes the government (president) of this country. Party rotation is a democratic sanction mechanism. That is an important indicator for country achieving the democracy. There are some factors must be considered while establishing voting behavior models in the turnover elections. There mainly are the party identification or party preference, the candidate image, and policy orientation. My thesis finds that party realignment is the key variable for 2000 Mexican presidential election, but 2000 Taiwan presidential is not. Candidate image or candidate personality centers on competence, likeability, honesty, reliability, and understanding people’s needs. Candidate evaluation refers to an overall assessment of the candidates: it is measured by way of “feeling thermometers”. My study found that candidate image is an important factor in determining Mexican and Taiwan voters choosing the opposition party candidate for a new president. The results shows that the policy issues, including prospective economic voting and the salience issue of change influence people’s voting choices.

繼「議題設定理論」於1972年問世之後,McCombs整理20年來媒介議題設定 效果的研究結果,提出了議題設定的第二面向:議題屬性的設定。主要論點 認為:每一議題本身都具有某些特定的屬性,而媒介對這些屬性的報導方式 與結論,往往也會模塑了閱聽人對此一議題的認知,使得人們看見這個議題 時,便會想到這些特定的屬性,因而對議題形成一種形象。本研究即以此論 點為理論基礎,意圖了解媒介形象設定理論在台灣選舉中的應用性。研究 結果發現:在「實質面向」(substantive dimension) ,也就是候選人形象 類目的比較中,媒介議題與選民議題間的確存在顯著的相關,也就是說 :「 刺激決定論」(stimulus determined thesis)在此具極大的解釋力;但在 「情感面向」(affective dimension)上 ,也就是對候選人形象評價的比 較中,「感知者決定論」(perceiver determined thesis)便成一項重要的 關鍵指標。故證實了諸多文獻的論點:對候選人形象而言,究竟「刺激決定 」與「感知者決定」何者力量較大,應視情況而定。本研究尚發現:「教育 程度低的選民,較教育程度高之選民更重視候選人的個人形象」、 「女性 選民與男性選民在認知候選人形象時,並無明顯差異存在」、 「當選民愈 支持某一黨派,則對該黨派候選人的整體形象評價便愈高」、 「政治涉入 程度愈高者,愈容易受報紙形象設定效果的影響」、「年輕人較容易受媒 介形象設定效果之影響」等。最後在研究建議方面,本研究建議日後的同 類研究,應採用小樣本連續訪問法 (panel study) 進行研究,以便獲得時 間架構因素的因果關係。此外, 日後研究的問卷設計中,可考慮加入人際 傳播因素的中介,以便實際了解候選人形象設定的來源, 究竟是人際傳播 的影響、或是媒介、政黨的影響較大。

在關於政黨認同的研究中，除了明白表示出政黨認同的這群人之外，其餘沒有明白表示出政黨認同的選民，通常被歸類為所謂的「中立無反應」者，或是「獨立選民」。然而，這群沒有明白表示出政黨認同或投票對象的受訪者是否真是中立或無反應者，抑或真是獨立選民呢？葉銘元（1994）針對我國的獨立選民曾做過相當詳細的研究，本文接續其研究，選擇政治大學選舉研究中心的面訪資料，針對1989年至1999年間我國獨立選民的發展與變遷情形，分別就「我國獨立選民政黨認同及社會背景的發展與變遷」、「我國獨立選民政治涉入的發展與變遷」、「我國獨立選民投票取向的發展與變遷」及「影響獨立選民發展與變遷的因素」等方面作分析。 研究結果發現：我國獨立選民的比例雖然有下降的趨勢，但是近年來此種趨勢似乎有逆轉的情形，且獨立選民比例的高低的確會受到選舉制度不同的影響。獨立選民的社會背景也和政黨認同者有差異，他們的政治涉入較政黨認同者為低，對候選人及政黨評價較低，對議題保持中立的立場，投票時比較會考量候選人的條件優劣，較不會依政黨界限投票。與此同時，我們也發現獨立選民的組成上是由政治知識及政治涉入較低的「冷漠型獨立選民」，逐漸轉變為政治知識及政治涉入較高的「理想型獨立選民」。 在未來的發展趨勢上，我們發現新世代及新新世代選民中獨立選民所佔的比例是較低的，因此我們可以預期，獨立選民的比例應該還有下降的可能性。在民主化的進程繼續推展，政黨政治日益成熟穩定之後，我們預期有政黨認同的選民會呈現增長的趨勢，與此同時，獨立選民並不會大幅度的減少。基於獨立選民是理想公民的期待，我們相信還是會有一定比例的民眾認為自己是獨立選民，但正如同我們觀察到的，此時的獨立選民將不再是對政治一無所知，或是對政治冷漠的冷漠型獨立選民；而將成為對政治保持高度關注，並能依自己客觀的態度做出政治決定的理想公民。 / In most studies about party identification, people who do not clearly indicate their party identification are usually classified as so「neutral non-respondents」or 「Independents」. Are these respondents really neutral or non-response, or are they really independents ? Yeh Ming-Yuan (1994) has done detailed research on independent voters in Taiwan. This thesis follows his research, using individual level survey data to analyze the developing and changing situation of independent voters in Taiwan from 1989 through 1999. In particular, this thesis examines changing party identification and social background of independent voters, changing political involvement among independent voters, changing vote intentions of independent voters, and the facts which have influenced the develop of independent voters. Our research outcomes have shown that although the percentage of independent voters in Taiwan has declined, in recent years this trend seems to have reversed, The percentage of independent voters has actually been influenced by changes in the electoral system. There are differences in social background between independents and party identifiers. The degree of political involvement of independent voters is lower than that of party identifiers. Their evaluation of candidates and parties is also lower than those of party identifiers. They tend to stay neutral to all political issues when making their voting decisions, they tend to consider the conditions of the candidates, rather than voting by the boundary of parties. However, we also discover that independent voters are gradually evolving from "apathetic independent voters", with lower political knowledge and lower political involvement, to "ideal independent voters", with higher political knowledge and higher political involvement. Looking to the future, we discover that the percentage of independent voters in the next generation and the generation after that are much lower. Hence we can predict that the percentage of independent voters will likely decline. With the continuing development of democracy and the increasing maturity and stability of party politics, we expect that the percentage of party indentifiers will grow. However, this does not necessarily mean that the percentage of independent voters will decline dramatically.Because of the expectation that independent voters are the ideal citizens, we believe that there are still a certain percentage of people who will consider themselves as independent voters. Hence, independent voters increasingly will not be "apathetic independent voters" who know nothing about politics. On the contrary, they will become "ideal citizens" who keep highly concerned about politics and make political decisions according to their own subjective attitudes.