過去二十多??台灣有關政治性議題的大型調查研究，往往忽視對其測?指標作有系統地檢驗。然而，唯有提出高品質的信?與效?的?表，將?才能建構可靠與正確的經驗性政治??。本文主要以二○○三?與二○○四?「台灣選舉與民主化調查」（TEDS）?次大型調查為?，統計檢定在這?次調查中，幾組重要的政治態?與?為?表的信?問題。信?問題的檢定方法，大致可分為再測信?檢定與內部一致性檢定，針對再測信?的檢定，本文主要透過卡帕（Kappa）係?，以及以平方加權法為計算基礎的加權卡帕（weighted Kappa）係??作檢驗分析；至於內部一致性的檢定，最常被使用的是克?巴阿爾發的係?（Cronbach’s α），以及折半信?中的蓋特門（Guttman）的計算公式。透過這四種方法檢驗TEDS (2003) 與TEDS (2004P) ?次大型調查後發現，整體而言這?次調查信?的檢驗結果都?錯，過去被認為非常敏感的政治性題目，?如選民的投票支持對象、族群認同、統獨?場與其政黨認同等問題，都具有穩定性。其他一些政治態??表，也具有高?的內部一致性。當然，目前仍有少?幾個題目信?的檢驗結果?佳，這些測?題目將?應在問卷討?中進一步修正。

In the last few years, empirical researches in political science have often neglected to systematically examine their measurement indices and methods. Concerning such a general neglect, this paper uses the large-scale surveys of TEDS conducted in 2003 and 2004 as the bases for analyzing a number of possible problems related to the reliability of the measurements of political attitudes and behavioral scales. The degree of reliability can be measured by performing a test-retest and then examining the result for the extent of internal consistency. Test-retest is carried out with the Kappa coefficient and the weighted Kappa coefficient with quadratic weighting. Internal consistency is tested with the widely used Cronbach’s α coefficient and the Guttman formula of split-half reliability. Using these four measures to examine the two TEDS surveys in 2003 and 2004, we find that, on the whole, the two surveys pass our tests for reliability. It is found that a number of previously considered very sensitive political topics, such as voter support, ethnic identity, Taiwan independence, and party identification, all relatively reveal stable conditions of results. We also find that the most important measures of political attitude display a high degree of internal consistency. In addition to the stability found in the acquired results, our tests also reveal some unexpected results on a small number of items, which may require further revision in our future research.