The Chinese Nationalist Party's (KMT) assets referendum and the anti-corruption referendum took place alongside the Legislative election in January 2008. The same year, the other national referendums on joining UN and returning to UN were held with the Presidential election in March. In the Legislative election, while 59 percent of electorate voted, turnout rate for the referendums are surprisingly low to 26 percent. In the Presidential election, in spite of a high turnout rate of 76 percent, turnout rate for the referendums are merely 36 percent. What is the reason for people to vote in elections but not on referendums? Partisanship probably explains people¡¦s turnout on referendum, but I suspect that there are more factors. This research provides a framework and a model of voting participation in referendum. Base on my analysis of Taiwan¡¦s Election and Democratization Study(TEDS), I find the following: First, Partisanship is the only factor which affect the referendums in the Legislative election and the Presidential election. However, people¡¦s partisanship reflects their values and leads to take or not to take referendum ballots. Other important factors of referendum voting participation are political interest, identity of Taiwanese or Chinese, What newspaper is used to read, unification-independence issue and perspectives on the relation between President and Legislative Yuan. Second, with identity of Taiwanese or Chinese and unification-independence issue, we could infer that attitudes about referendum issue affect referendum voting participation behavior. Finally, by the electors that prefer to check and balance between the executive power and the legislative power likely decide to take referendum ballots, we could assume that some voters who realize and support referendum system intend to vote on referendum.

過去立法院常被譏為行政院的橡皮圖章，但隨著國內民主政治的快速發展，立法院已成為我國當前最重要的政治舞台，不管是各政黨或是政治人物，對於在立法院合議制度下，所形成的權力分享與利益分配，都有莫大的關注與興趣。再者，立法院職權之行使攸關人民生活財產的維護，在民主化的過程裡，代表民意的立法院，更背負著民眾的高度期待。因此，研究立法院之監督功能，對立法與行政之間的互動關係做一明確的界定與釐清，將有助於了解我國民主政治的建立與實踐民主化的進程。 本研究旨在探討立法院之監督功能，也因此擬先介紹民主國家國會監督概念，針對「監督」一詞，加以界定並分析其理論基礎。再針對各國憲政發展之比較，來了解何以要研究我國立法院監督功能，並從我國民主發展過程中，探討立法院扮演監督功能的意義與角色，最後試圖從法制規範層次與實務操作層次上，找尋出我國立法院所具有的監督效果，以達到探究其監督功能之目的。 從萬年國會到國會全面改選，乃至目前歷經六次修憲以來，我國國會體制逐漸落實立法院真正成為全國最高立法機關之制度設計，不過立法權如何有效落實其監督制衡行政權的功能，除了從靜態的憲法條文賦予的職權來分析，還必須透過動態的立法權行使過程加以探討。因此，本文先從民主憲政原理中的監督背景及其概念，假設我國在國會全面改選後，雖然行政部門過去的主導力量甚強，但立法權仍能發揮其監督行政權之功能，以立法院所行使的職權是否具有監督功能來分析，並觀察出立法院非制度面部分，具有的實際監督功能之機制，最後做出結論與提出建議。 本文分成六個部分。第一章緒論，介紹本論文研究背景、研究目的、概念界定、研究架構設定、文獻回顧、研究方法及研究範圍與限制；第二章國會監督功能之理論基礎，從形成國會的民主概念破題，來介紹國會監督功能之背景、意義、基礎、技術，以了解民主國家之國會監督功能；第三章我國立法院監督功能之介紹，從我國立法院形成之背景，來介紹立法院監督功能之背景、意義、基礎、技術，以呼應第二章；第四章法制規範層次之監督功能分析，是以監督對象為基礎來分析立法院的監督功能，文中將監督對象分成對人如同意權、彈劾權等；對機關如預算權的監督；對政策如法案權，再放置到四個重要民主政治概念（權力分立、代議政治、有限政府、權力制衡）中來檢驗，以了解其法源依據、規範作用及實質影響；第五章實務操作層次之監督功能分析，將提出具有實際運作而無法在法制層次探討如議事杯葛、政黨協商等非正式制度面之監督功能，從實務操作層次中，去探討其實質影響。第六章結論，提出研究過程中所得到之發現，及希冀提供後進研究者之建議。

本研究旨在探討直轄市山地原住民區改制前後，原為官派制轉換成民選制的區，比較兩種制度在地方上的治理運作與模式，以表現不同制度在相同環境、事務時所產生的運作差異，一方面評估不同制度下之運作；方面評估制度中的民意制衡能力。最後分析民選制與官派制於治理上的優缺點和差異，並對當前制度提出相應的改善建議。 本研究採用的研究方法如下： 1.個案研究法：選定新北市烏來區為個案，觀察制度轉換之差異所造成的影響。 2.文獻分析法：整理分析相關文書資料以比較官派制與民選制的結構、機制之差異。 3.深入訪談法：選定直轄市山地原住民區改制前後，擔任行政與民代職務者，進行職 責與實務的訪談，調查制度差異對行為與結果的影響，以及地方感受跟評價。 / The case study is research to the mountain indigenous districts of special municipalities, the former official system is converted into an elected system of district.By comparing the models and practices of these two local governance systems brought by the institutional reform, which has led our attention to the differences of behavior and reaction the two systems have while put in the same environment and under the same government affairs, this present study evaluates the two systems with special focus on their administrative performance and check-and-balance capability of the public sector. The ultimate purpose of this study is to analyze the merits and limits of elected representation and official appointment and through observing one real case, and finally propose advices to improve the current system. Methodologies adopted in this study include: 1.Case study: Wilai District of New Taipei City is chosen to observe the impact of institutional reform. 2.Literature analysis: Relevant documents and data are collected and analyzed for the comparison of official appointment and elected representation in their organizational and functional differences. 3.In-depth interview: Administrative officials and elected representatives who have experienced the transition are interviewed about their duty change and practical exercise of governance, in hope to investigate behavioral and consequential influences the new system has caused, and collect feedback and evaluation from local government.

隨著民主治理的發展，有效調和民主憲政與文官專業價值，才能夠確保民主政治運作的效能。背景因素使得政務官與事務官間的互動，愈具舉足輕重的角色。從相關研究可以理解，欲直接觀察政務官與事務官細部的互動並不容易，兩者間細部的互動對外界而言始終是一個黑盒子。因此，本研究使用鮮少作為該議題研究主體的影視產品進行分析。選擇英國經典影集「Yes, Minister」，乃因該劇主要以描繪政務官與事務官互動為主，描寫程度相當細膩。藉此旁觀的角度一窺政務官與事務官的互動，確實是個能夠協助深入理解該議題的良好個案。 研究設計的基礎為文本分析法，透過建構分析路徑，分別從兩者本質、選擇偏好、表現作為，以及互動後對政策產出的討論。先將14集研究樣本進行相同路徑的文本化，再將文本化的樣本進行論述與類型化。研究發現事務官的互動技巧，包括行政權力的作為與不作為、可行性評估、影響政務體系與政務官以及納入外部勢力等，目的為提升參與決策的條件，細部列出24項互動技巧；另一方面，政務官則可透過經驗複製、行政學習、媒體與聲望經營與納入外部力量等，目的為提升統治便利性與正當性，細部列出13項互動技巧。兩者的互動呈現出互相制衡的本質，若各自無法有效負責與調和時，易形成雙重價值選擇的壓力。 本研究提出以「鐘擺效果」解釋政務官與事務官的互動，兩者在制衡關係的基礎上，隨著時序性的影響而改變互動模式。基本上的互動如鐘擺式擺盪，進而系統性因素將會因時序性遞減，激發出最適的政策結果。最終提出兩大項建議：首先，從制衡到動態平衡的關係，正視制衡現象的存在、立基於分權制衡上的信任關係、減少彼此錯誤解讀的機會，以及培養持續監督與自省能力；其二，設定政策決策的妥協底線，認知非任務型指標的超然價值，以及試著創造沒有共識的共識。 / In any democratic state’s development, how to effectively reconcile the two values of constitutional democracy and bureaucratic expertise is fundamental to ensure the proper functioning and efficacy of democratic politics. As Taiwan further consolidates its democracy, the interaction between the political appointees and senior civil servants becomes ever more important in achieving a balance of pluralistic values on the one hand, and government’s administrative efficiency, on the other. Yet past research has found that it is not easy to observe directly the details of the interaction between political appointees and civil servants; to the external world, the actual interaction between the two has always been a black box. Therefore, to shed light on that black box, this study analyzes the interaction between appointees and civil servants as depicted in film and television programs. The classic British series, “Yes Minister,” was chosen as the study’s research subject because it portrays the minister-bureaucrat interactive behavior skillfully and poignantly, which allows a deeper understanding of the issue. The study employs textual analysis as the principal research method. Each of the series’ 14 episodes had been textually reinterpreted by using the same analytical paths, which were constructed by first categorizing the nature, choice preferences, display behaviors, and post-interaction reaction to policy outputs for both political appointees and senior civil servants. The research has found that civil servants, in their pursuit for greater participation in decision-making, exhibit as many as 24 distinct interactive behaviors with their political superiors. Some of these behaviors include administrative action and inaction, call for feasibility assessment, manipulation of the political system, and incorporation of external forces. On the other hand, political appointees, in their effort to achieve political expediency and legitimacy, have 13 behaviors of their own through experience replication, administrative learning, media and reputation management, and inclusion of external forces. Both sets of interactive behaviors are essentially parts of a checks-and-balances system. When responsibilities are unclear and actions cannot be coordinated effectively, the agency then becomes vulnerable to role ambiguity and double value selection problems. This research proposes a “pendulum effect” to explain the interaction between political appointees and civil servants. Because the appointees and civil servants have a mutual checks and balances relationship, both parties will alter their interactive behavior depending on the timing of events. Essentially, the interaction between the two sides will initially swing freely like a pendulum; subsequently, system factors will cause the pendulum to progressively decrease its swing, ultimately arriving at the most optimal policy result. In conclusion, this study makes two major recommendations. First, both political appointees and civil servants need to understand that they are in a dynamic equilibrium, in which they check and balance each other’s actions. By acknowledging the check and balance nature of their relationship, both minimize the chance of misinterpreting each other, and may develop the healthy capacity of ongoing oversight and self-introspection. Second, both political appointees and civil servants must establish a baseline in any policy negotiation and compromise, recognize the (sometimes) extraordinary value of non-mission-based indicators, and attempt to create a consensus when no consensus exists.