問卷題目中常見態度問題之設計，其主要目的是想瞭解受訪者對於特定議題或事件的觀點。在不同態度量表的形式中，以李克特量表(Likert scale)最具代表性，受訪者會被要求對於某一個陳述說明其贊成或同意的程度，這些選項的標示語可隨問題的不同而改變。關於態度量表的方法學議題，大部分研究所關注的焦點在於問卷題目的組成方式、順序及措辭等，對於選項設計的研究相對較少。本研究分析大型學術調查資料中，李式量表之選項標示語設計差異對回答模式與分佈之影響，所使用之資料爲「台灣社會變遷基本調查」2005-2007的三筆樣本資料，其問卷內容均包含人格量表，本研究選取內容相同之題項進行分析。潛藏因素分析模型的結果指出：標示語的內容確實影響受訪者的回答類型分佈；此外，中間選項具備作爲二極端態度測量之中間位置的角色，然而在主動提供的情況下，亦確實影響了受訪者的回答模式與分佈。最後並討論了本研究在分析方法上的創新，以及研究結果對於調查訪問的實務貢獻。

Attitude scales have commonly been used in questionnaire design. Among various scales, different versions of Likert-type scales are often employed. Previous research has indicated the importance of response design in terms of its number and labels. However, their suggestion and implication are based on findings using samples and languages in Western societies. It is not clear which design is more appropriate to be used in Chinese questionnaires, in particular when translated versions are applied. This study attempts to examine the influence of response designs in Likert-type scales on response patterns by using large-scale survey data. Data used for this study were items of Personality Scales in the questionnaires of 2005-2007 Taiwan Social Change Survey. The results of latent class factor model indicated that response patterns differed when response labels in a same 5 point scale varied. In addition, the appearance of mid-point response category had a significant influence on the distribution of response patterns. Conclusion and discussion on the contribution of this study to survey research and practice of questionnaire design are provided.