顧客滿意度之直覺模糊多元屬性衡量方法及比較研究分析

以往顧客滿意度相關之研究，常因調查的方便性，仰賴單一化的李克特尺度衡量受訪者的認知與態度，儘管尺度有其建構明確及應用便利的優點，卻無法充份表達人類思維之複雜性、主觀性以及不明確的喜好程度，使得單一且等距的尺度表達可能產生過度解釋之偏誤。爲改善傳統問卷形式的限制，本研究發展直覺模糊問卷，利用直覺模糊集合中的隸屬度、非隸屬度與游移不定程度擷取受訪者對問項的同意、不同意以及不確定的程度，充分量化人類思維的複雜性與模糊性，並且解決語意膨脹或貶值的問題。透過評估函數、直覺模糊點運算式、平均運算式與計分函數爲基礎建構顧客滿意度的多元屬性模型。爲驗證衡量方法的有效性與可行性，本研究以大臺北地區之電影影城爲實證研究對象，且同時發放李克特尺度與直覺模糊尺度問卷以比較兩者之間的差異。經由分析結果發現，三大衡量方法均優於傳統的統計計算方法，亦能運算於非補償性決策問題上，有效地檢視屬性間的連結性與非連結性關係。實務應用上，本研究所發展的衡量方法可視爲一個通用且具備彈性的衡量工具，當調整公式中的參數及運算式後能適用於不同的產業類別，企業更可根據不同的產業特性找到其最適合的衡量方法。

關鍵字

顧客滿意度；直覺模糊集合；評估函數；點運算式；平均運算式；計分函數

並列摘要

Most research concerning the customer satisfaction often uses the Likert Scale to measure subjects' recognition and attitude due to investigating convenience. In spite of the advantages of explicitness and convenience, the Likert Scale obviously overlooks the important perspective that human thoughts are characterized by complexity, subjectivity, and uncertainty. Hence, the results generated by the interval scale may easily lead to a bias of explanation. To improve the shortcomings, this study develops an intuitionistic fuzzy questionnaire. Based on the membership degree, non-membership degree, and hesitancy degree expressed by the intuitionistic fuzzy set, the complexity and fuzziness of human thoughts can be fully quantified, and meanwhile we can resolve the problems of linguistic expansion and contraction. By means of the evaluation function, intuitionistic point operator, averaging operation, and score function, three measures to deal with a multiple attribute decision-making problem pertaining to the customer satisfaction are constructed in this study. In order to manifest the effectiveness and feasibility of the proposed measures, an empirical study on movie theaters in Taipei was employed. The difference between the Likert Scale and the intuitionistic fuzzy scale is compared. The results show that the three proposed measures to assess the customer satisfaction are all superior to the statistical method; in addition, the measures are also capable of coping with a non-compensatory decision problem to examine the conjunctive and disjunctive relationship between attributes. In practice, the proposed measures can be regarded as a general and flexible tool. The measures can be applied in various industries after the parameters and operations are adjusted, since the enterprises can find out the optimal measure according to different industrial characteristics.