當台灣一旦進行較大型的公職人員或民意代表選舉時，各個新聞媒體都會報導出其所獲得的民意調查結果，可見得媒體、民眾、政黨以及候選人，都會將民調當成很好的指標工具。 當我們獲知所公佈的民意調查資料時，都會看到報導幾乎抽樣樣本都在1000個左右，來說明其抽樣誤差約在3%。但是本文所探討的是，我們是否需要每次固定在1000個左右的樣本數，或許我們可以更快就可以看出結果，早一點讓決策者多一些時間做出決策，不但省下了多餘的時間，也可以節省相當多的金錢。 在本文中以民國八十七年的由全國意向民意調查中心所做的台北市長選舉意向調查為例，分別在五月和六月的兩次的訪問中，我們可以很清楚明顯的看出，當我們進行到六百個樣本數時，我們所訪問的資料已經出現相當穩定的結果，即使訪問更多的樣本數的樣本，對於結果的差異幾乎沒有影響，所以在此兩個例子中，我們可以在訪問過程中，就有足夠很好的證據去停止再抽樣訪問，可以和訪問完成一千多個樣本時才作決策時比較，我們可以省下三、四百個樣本數的訪問時間來思考做決策，也省下了三、四百個樣本數的成本來做其他的工作；若是並無法能夠獲得明確的結果，也是會做到原先設定的固定樣本數訪問法，一樣無法獲得相當好的結論，也是必須再繼續抽樣，才能夠看出更明顯的結果。 本文所探討的內容，並非每次訪問過程都是只要像例子中，只需六百個樣本左右即可，這是需視各個不同訪問的個案，也許訪問中各問項變動性太大，那麼就需要更多的樣本數來做觀測。

Magister Educationis - MEd / This study investigates the attitudes of students and student leaders towards democracy in terms of their demand for democracy, their perception of the supply of democracy, and their awareness of and participation in politics. Existing literature does not provide any conclusive explanation as to how and to what extent higher education contributes to democracy. Mattes and Mughogho (2010) argue that the contribution of higher education to support for democracy in Africa is limited while other scholars such as Bloom et al. (2006), Hillygus (2005), and Evans and Rose (2007a, 2007b) maintain that higher education impacts positively on support for democracy. The study follows the conceptualisation and methodology of previous studies based on the Afrobarometer public opinion surveys into the political attitudes of African mass publics (Bratton, Mattes and Gyimah-Boadi, 2005; Mattes and Bratton, 2003; 2007), and of students in African universities (Luescher-Mamashela et al., 2011; Mwollo-Ntalimma, 2011). The survey uses a stratified random sample of third-year undergraduate students at the University of Botswana. Furthermore, it isolates the subgroup of student leaders to investigate whether active participation in student politics influences support for democracy

No description available.