選舉制度的設計與評估，主要有兩方面的考量，一是可治理性；二是分配的比例性。以台灣的情況為例，新選制的選舉結果，理論上政府的運作將趨向穩定；相較之下，在制度替換的過程中，改革者通常只看見舊制度的缺失，舊制度本身所具有的優點－比例性，在改革過程中往往被忽視。本文的研究目的即是重拾對比例性的視野，透過Loosemore-Hanby D指數測量選制改革前後的比例性偏差程度，檢視1992至2008年間比例性偏差數值的變化，並從五個面向－選區規模、席次分配、法定門檻、席次總數、選票結構－解釋比例性偏差發生的原因，發現所有的選制設計皆不利於比例性，導致新選制的比例性偏差與舊制度產生極大的落差。 此外，透過比較其他可能的比例性狀況可以發現，無論是聯立制的模擬或是票源模擬的結果，皆顯示有助於提升小黨的代表性；另一方面，同樣採用並立制的日本，四次的選舉經驗可以做為我國參考的對象，但是由於選舉制度的各項要素不盡相同，也使得兩國選舉結果的比例性和小黨代表性呈現差異性。 最後，混合制固然同時融合了相對多數決制和比例代表制的特性，但是由於制度上選制要素以及非制度上國情與社會狀況的不同，很難期待一個完美的選舉制度出現，因此，我們只能尋求一個理想的選舉制度，且善用混合制所具有的彈性，做出適度的調整。 / Upon the evaluation of an electoral system, the main concerns are the governability and proportionality. In Taiwan’s case, the electoral system of Legislative Yuan altered from SNTV to mixed-member system; however, the reformers invariably accentuated the stability of governance but ignored the semi-proportionality of SNTV. Hence, the result of election in 2008 got unbalance between governability and proportionality. This article tends to measure the degree of disproportionality by Loosemore-Hanby index from 1992 to 2008 and tries to explain how disproportionality occurred from five dimensions－district magnitude, levels of seat allocation, threshold, number of total seats, and ballot structure. In addition, the results could be quite different when it compared with of other situations. According to this paper, the representation of small parties and the degree of proportionality will be enhanced while adopting MMP and redistributing the votes. Moreover, this paper is going to make a comparative study with Japan which have adopted MMM and have experienced four times elections since 1996. Finally, although mixed systems combine some features of both plurality and PR, there is still no perfect electoral system in the world. However, by making good use of flexibility of mixed systems, we might have an “ideal” system which can be regularly updated to suit changing needs and political conditions.