政治知識做為選民的認知基礎，以進行有意義與有品質的政治參與。對於希望提升民主品質的有識之士而言，探索哪些政治知識題目較容易答對？與哪些政治知識在一段時間之後較容易學習？的確為一個重要課題。本文的問題意識在三方面：第一、在一段時間的觀察之後，是否有哪些特定的政治知識種類廣為人知？與經過一段時間之後較容易學習？第二、不同教育程度的受訪者之間是否存在著政治知識鴻溝？這些鴻溝在不同的政治知識項目上有何不同樣貌？第三、政治競選活動是否顯著地幫助低度政治知識選民去學習？長期而言，接觸政治競選活動越頻繁，是否能夠縮小低度政治知識選民與高度政治知識選民之間的知識差距？也就是所謂的知溝彌合論。本文的研究發現為：美國總統的名字與臺灣行政院院長的名字皆為觀察知識。只是前者為簡單知識，後者為簡單可以學習的知識。此外，「何者為釋憲機關」因為其教科書知識的特質，而較難學習。此外，如果給予低度政治知識受測者一些時間去學習「時事知識」，則這群受測者會顯現出縮小與高度政治知識受測者之間的知識差距。最後，電視競選廣告與總統電視辯論會提供低教育程度組學習機會，也成為知溝彌合的可能因素之一。本文的研究設計採用了跨區分析與條件式固定效果邏輯斯分析雙軌並陳的方式，企圖獲取代表性與因果性的研究發現。

For ordinary citizens, political knowledge serves as a prerequisite for meaningful political participation. Thus, types of political knowledge that are easy to know and easily learned over time among voters are of particular importance to those who long for an informed citizenry. Three research questions are asked. The first is whether or not a specific type of political knowledge is easily known and easily learned over a period of time among respondents. The second concerns the extent to which a knowledge gap exists between different population segments measured by education, and how this gap exists when it comes to different types of political knowledge. The third question is whether campaign activities have the function of helping the low educational group to improve in political knowledge and ultimately close up the knowledge gap. The research findings are such that both the U.S. president’s name and the Taiwan premier’s name are the surveillance fact, and the former is easy fact and the latter is the easily learned fact. On the other hand, knowledge about constitutional interpretation is relatively hard to pick up given that it has the nature of taught fact. Moreover, if given time to digest, the low educational group is likely to catch up with the high educational group when it comes to the surveillance fact. In this case, the knowledge gap tends to disappear. Last, but not least, campaign news on TV and televised presidential debates help the less educated to learn and serve as another possibility for closing up the knowledge gap. This study is based on one-shot cross-sectional analyses as well as conditional fixed-effect logistic regressions for its empirical examination.