調査硏究被廣泛的使用於現今社會，而研究中出現頻繁的「無反應」問題卻常爲硏究者帶來統計分析與推論的困擾，基於預防重於治療的前提下，「無反應」因素的探究實有其必要性。本硏究的目的即針對「敏感性」問題於「投票意向無反應」的情況，應用卡方檢定與logit模型驗證現有理論基礎的適切性，並進一步建立新的「無反應」問題解釋模型，以幫助我們暸解此種「無反應」問題在現有台灣調査硏究中可能存在的因素，進而減低其對硏究結果的不利影響。

影響「無反應」因素的驗證部分，大致上與前人的硏究結果相同，只有少數的變項有所差異。另外，在模型配適的解釋部分，造成投票意向「無反應」的因素有：受訪者的性別、省籍、政治參與度、對統獨立場的看法、對族群認同的看法、受訪者的性別與對統獨立場看法的交互作用以及受訪者的省籍對族群認同看法的交互作用。

Conducting survey research has been a common practice in the field of social science. When implementing questionnaires survey, it is quite usual for researchers to encounter the non-response feedback which inevitably has some statistical impacts when analyzing and interpreting the results. This research is directed solely to study the effects of sensitive questions on the voter intentions which may have caused the voters' non-response reactions. The two statistical methods used in this study are Chi Square test and Logit analysis. Furthermore, this paper has put an effort in establishing a model for explaining the non-response problem in order to aid the researchers becoming oriented with the possible impact, be it negatively or positively, on the results of survey research in Taiwan.

The factors influencing the non-respondent problem are quite consistent with the previous findings of other scholars except for some variables which may interact with interviewees' political attitude on vote intention questions.