本文主要針對政治知識的測量方式及其測量結果進行分析討論，比較不同題型的測量方式是否影響受訪者答題時的行為，進而影響到該概念測量的信度與效度。本文分析的資料來源包括台灣選舉與民主化調查小組針對2012年總統與立委選舉進行的面訪調查資料（TEDS2012），以及作者以實驗設計方式針對東吳大學學生為實驗對象所獲致的資料進行分析。本文根據TEDS2012的資料分析，初步發現開放題設計的政治知識對選舉參與的影響力，比選擇題設計題型來得佳，再根據實驗設計所獲致的資料進一步分析，發現選擇題型提供受訪者「猜猜看」的機會，不僅提高回答率，也的確增加受訪者政治知識的分數，是以不同方式的測量，確實影響政治知識測量的效度與信度問題，更會影響政治知識在解釋其他政治行為與認知能力的解釋力。然而，受訪者政治知識增長的原因，是內在的知識推動，還是外在的猜題空間增加，則有待未來適合的資料做進一步探討。 This paper aims to examine the measurement of political knowledge and its consequences via different formats of question designs. Would using different formats of questions, elicit different response behaviours from the respondents? We argue that different types of questions affect the reliability and validity of the concept. The data used in this paper was collected by the Taiwan's Election and Democratization Study (TEDS) which was designed to study electoral behaviour in the 2012 presidential and legislative election. Another data was collected by an experimental design survey carried out on students of Soochow University. Based on the preliminary analysis from the TEDS data, open-ended formats of political knowledge exert more influence in explaining electoral participation than closed-ended. The results from the experimental design survey further shows that closed-ended questions provide the respondents with opportunity to ＂guess＂ the answers, thus respondents turn out to have a higher score of political knowledge. This shows that different formats of questions truly affect the validity and reliability of the concept of political knowledge. This issue also plays a role in showing how political knowledge serves as an independent variable in explaining political attitudes and behaviour. Whether a higher score is due to the respondents' ＂true＂ knowledge or the chance to guess, however, needs further data to explore in the near future.