零售業是高度競爭性的產業，業者必須隨時注意著消費者的需求變化與顧
客對於企業經營品質的認知，調適其經營型態。亦即零售商店要確實掌握
顧客內心真正需求，發展並提供正確的產品與服務。因此主要目的在探尋
顧客內心潛在需求的零售服務品質衡量，對零售業經營而言，就顯得格外的重要。
本研究有鑑於，以往大部份有關服務品質衡量的研究，其研究焦點都集中
在Likert scale五或七等第尺度量表，取得語意評估變數，經由人為配分後
的「數值」是用何種方式來計算服務品質。但是這種經由人為配分方式，
會有遺失寶貴資訊的缺失。
本研究針對傳統服務品質在衡量機制方法的缺失，提出二因子模糊多準則
方法來衡量顧客對零售商店服務品質評估，並針對傳統服務品質衡量方法與二
因子模糊多準則方法進行比較。二因子模糊多準則方法的使用，是經由服務品
質、模糊理論與排序等相關文獻探討及服務品質二因子模糊數乘積方程式的推
導而完成理論基礎，經由實務訪談建立數據收集的架構（亦即：評估量表的修
正），最後在台中地區，在便利商店、量販店、百貨公司及超級市四個業態中
抽取29家綜合零售商店進行實證調查，收集初級資料，再編寫電腦程式完
成運算、分析。最後得出以下本研究重要的結論：（1）由初級實證調查資料顯
示，服務績效與服務屬性重要程度資料的分佈確實呈明顯負向偏態。而模糊多
準則方法可藉由歸屬函數的調整，使資料分佈回歸常態分配（2）由服務屬性重
要性調查資料顯示，一般顧客重視的服務屬性，都是一些基本核心的服務，顯
示顧客對零售商店並沒有不合理的期望及不願意承擔購物風險等。（3）6個服
務品質構面重要性在四個業態間呈現顯著差異，其中以量販店業態與其他三種
業態在6個構面上的差異，幾乎都達顯著差異，這足以說明量販店將成未來綜
合零售業的主流。（4）二因子模糊多準則方法比傳統服務品質衡量方法能更正
確衡量顧客對零售商店服務品質的評估。效標關聯效度的相關係數為二種方法
比較的準則。從客觀的效標關聯效度的相關係數大小比較、Wilcoxon符號等級
檢定；及主觀的零售商店服務品質指標的排序評估，有充份證據說明模糊多準
則方法比傳統服務品質衡量方法，能正確衡量顧客對零售服務品質的評估。由
此可知，模糊多準則方法比傳統服務品質衡量方法在本研究實證資料中，顯得更加適用。

Since the retailing business is a highly competitive industry, the management
team has to monitor the shift of customers needs and their
perception of quality
closely. The awareness of this kind of shift is helpful in adjusting the
management style for each retail store, or even the whole industry.
The key point is the necessity for the management team to grasp the real needs
of customers, and then develop as well as provide the right
service or products .
Therefore, how to discover the real expectation of customers and establish a
measuring mechanism are important issues for researchers.
Due to the inadequacy or inappropriateness in using Likert''s scale method with
traditional way to measure service quality in some industry, this research is
intended to find out a new mechanism for exploring more information when we
have survey data on hand.
We adopt concepts from Fuzzy theory, as well as develop a new two- factor
evaluation schema to serve this purpose. In the scenario of general retailing
business, we modify the retail service quality scale designed by Dabholkar,
Thorpe and Rentz, use it as a vehicle to help the data collection.
Totally, we survey 4 types of 29 retail stores in Taichung
district, obtain 1376
valid questionnaire samples, and then make a comparison between the new "two
factor fuzzy method" against the traditional data treatment by Likert''s scale.
Our major findings are : (1) the performance and importance of data of service
attributes surely show a negatively skew phenomena, and fuzzy multi-criteria
may adjust it back to be normally distributed; (2) what the
customers care most on the
importance of service attributes are all core items, there are no exceptional
expectation from ordinary customers; (3) it is obvious that among 4 types of
etail stores there exists significant difference in the
importance rating of some
service quality dimensions; (4)the two-factor multi-criteria
method shows better
interpretation on the surveyed raw data than the traditional
one, this also means
that two-factor method has potential to discriminate the service quality among
different stores better.
Since all the data collected with retail service quality scale
are ordinal data in
nature, we use the Wilcoxon "Signed Rank Sum Test" and the Kruskal-Wallis
H-Test to perform all the necessary statistical analysis.