Dear Mahmoud,

Yes, I believe that fuzzy sets can be well incorporated to such measure of quality. In fact I beleive it is necessary.

Please find enclosed the Frame work on Servqual

Dear Kamal,

Thank you for your interests and participation. As a matter of fact my information on this question is poor. That is why I asked the question.

All I know is that SERVPERF and SERVQUAL are the two most prominent scales forming the genesis for service quality assessment in different service sectors. However, the choice between these metrics for service quality measurement is subjective and the research literature lacks evidence on whether these instruments differ in their outcomes significantly or concur with each other. Here I am interested to apply in combination with fuzzy set theory it for measuring teaching quality within academic institutes.

Dear @Mahmoud, I was not aware of this, but after some search on prominent scales, I have found an article which is attached. My dear @Kamal, it is about medical education! The full article's name is "Reliability and Performance of SEVQUAL Survey in Evaluating Quality of Medical Education Services"!

Here, at RG data base there are many papers available on the application of SEVQUAL in educational services! Not bad metrics at all! Here is a paper!

Conference Paper “Service Quality In STEM Higher Education: A Hierarchical Approach”

Dear @Ljubomir, Thank you for joining the discussion and the interesting links to the applications of the topic.  Certainly these metrics are being used in many places such as banking, airlines, restaurants, etc. as I mentioned them in question. Here I was interested to get some ideas on how we can enhance or improve these metrics by combining them with fuzzy set theory. Briefly, I give below the mathematical expressions (some backgrounds) for these metrics:

(1)     SERVPERFi = Pij              (2)      Weighted SERVPERFi = Iij*Pij

(3)     SERVQUALi = Pij-Eij        (4)      Weighted SERVQUALi = Iij*(Pij-Eij)

with Sigma j=1 to k in front of these expressions, and k is the number of service attributes/items. In Eqns. (1) to (4), Pij is the perception of individual ‘i’ with respect to performance of a service firm on attribute ‘j’, Eij is the service quality expectation for attribute ‘j’ that is the relevant norm for individual ‘i’, and Iij is a weighting factor which gives the importance of attribute ‘j’ to an individual ‘i.’,

In these metrics, customers’ responses to their perceptions (P) and expectations (E) are obtained, for example, on a 7-point Likert scale. The higher (P-E), the perception minus expectation score (more positive), then the higher is perceived to be the level of service quality.

Thanks dear @Mahmoud! I wonder as Likert scale was applicable in different areas, what would happend if we do with more than 7 points scale?

http://www.surveygizmo.com/survey-blog/likert-scale-what-is-it-how-to-analyze-it-and-when-to-use-it/

Sure we can!  A 7-point Likert scale I said was just an example (not definite). .

Please see the following link for the paper entitled, " A customer satisfaction model based on fuzzy TOPSIS and SERVQUAL methods". I think it is a useful link.

Dear Mahfuz. Thank you for answering. You left out the link address, however.

Dear Mahmoud. I am so sorry for that. Here it is:

http://www.tadbir.ca/lnms/archive/v5/lnmsv5p74.pdf

Dear Mahfuz. Thank you.

The paper you provided link for it suggest an integrated approach consists of SERVQUAL and fuzzy TOPSIS methodology for evaluation of service quality of public transportation systems.

The authors argue that SERVQUAL is accepted as the basis for measuring service quality in service organizations, They say that the criterion used for measuring the quality of service are not only limited to quantitative, because service quality dimensions cannot be measured quantitatively, so as MCDM approaches like TOPSIS can be used successfully in this area. However, these are determined and evaluated in a subjective and qualitative way, so they should be described linguistically. To handle with this, fuzzy logic is used as a mathematical way to represent and handle vagueness in decision-making. 

"Various techniques such as SERVPERF, SERVQUAL / weighted SERVQUAL are being used for measuring service quality in banking", but their effectiveness, on my opinion, near zero. For example, this was connected with a banking scandal on Libor (http://blogs.channel4.com/factcheck/factcheck-anatomy-of-a-banking-scandal/10849), (http://darwinbondgraham.wordpress.com/2013/02/24/libor-anatomy-of-the-biggest-financial-crime-in-world-history).

Dear @Sylantyev thank you for your comments and links. My question was not about SERVPERF, SERVQUAL, weighted SERVQUAL metrics solely, but rather how to combine and hybridize these metrics (Eqns 1 to 4 in my earlier post) with fuzzy set theory, so their effectiveness improve.

Dear @Mahmoud, theory of fuzzy sets can not be improved due to the fundamental features of the method associated with the loss of information. Firstly on fuzzification stage (the loss of information), and secondly on the defuzzification stage (the loss of information). The use of any methods to combine and hybridize any metrics can not make up this loss of information.

Dear @Sylantyev, Thank you for your remarks. Quite recently [1] we studied these metrics with ANNs. Comparison of ordinal logistic regression and ANN results revealed that ANN modeling approach is able to model students' attitudes more closely. For evaluating the quality of teaching and learning process, four models were used with ANN; including non-weighted SERVPERF, non-weighted SERVQUAL, weighted SERVPERF,and weighted SERVQUAL. The results of hybridizing these metrics with ANNs showed that weighted SERVQUAL metric with ANN is more accurate to evaluate the quality of teaching and predict satisfactory. I thought maybe we can get a better result by involving fuzzy set theory. Why you think we will loss information if we fuzzify or defuzzify a system or use fuzzy set theory? Fuzzy is ideal approach for qualitative data.  

[1] Rajabiyan Gharib, F., 2013, Quality Assessment of Teaching and learning Process in Agriculture Higher Education Using Artificial Neural Networks, Case Study ; School of Agriculture and Natural Resources of the University of Tehran and Faculty of Agriculture of the Ferdowsi University, MSc thesis, Department of Agricultural Extension and Education, University of Tehran, Iran, 144 page, [In Persian]

Dear @Mahmoud, may be this techniques are very effective for to model students' attitudes, but in common for using the different models with various techniques (for different sciences) I am support the next idea of Barnett , W. A. , W. E. Diewert , A. Zellner «. . . all of applied econometrics depends on economic data, and if they are poorly constructed, no amount of clever econometric technique can overcome the fact that generally, garbage in will imply garbage out. . . .».

Barnett , W. A. , W. E. Diewert , and A. Zellner . 2011. Introduction to “Measurement with Theory.” Journal of Econometrics 161: 1–5.

Barnett W.A. Getting it Wrong : How Faulty Monetary Statistics Undermine the Fed, the Financial System, and the Economy / W.A. Barnett, A. Serletis. — Cambridge ; London : MIT Press, 2012. — 360 p. (see page 20.).

(http://www.amazon.com/Getting-Wrong-Statistics-Undermine-Financial/dp/0262516888 )

Dear  @Sylantyev  thank you for comments and link. Your efforts to help me are really appropriated.