1- Preliminary provisionsProgressively and with the aim of qualifying the gap measurement context of the proposed EP, the document approved a number of recognized scientific techniques. The objective was to exploit these techniques in order to define an acceptable management instrument and to anticipate the concept of flexibility. Personal or targeted interviews with service users would be recommended. And to define scale attributes, a manager can capitalize on previous empirical work cited in the literature and relevant to his or her own service sector. On the other hand, adopting the Delphi technique would help refine the multi-item questionnaire according to the highlighted subscales (factors). In addition, carrying out a pilot study would make it possible to refine the final scenario of the scale. However, the above approaches are inferred from the work of Lam, Zhang and Jensen (2005). Alternatively, the Walk–Thru-Audit practice of Fitzsimmons and Maurer (1991), cited in Olorunniwo and Pennington and Hsu (2002), could also be applied to develop a questionnaire. In their method, a third party familiarizes themselves with the plan for a sequential service before experiencing the service itself, and he or she then conducts the walk-in audit while developing their own chronological questions; however, these questions are ultimately formulated to construct the statements which are grouped into different dimensions. .....the grand mean, , which is the average of the means of the four samples.c) Calculate the mean of the sample ranges, . However, it is important to construct an R chart to detect changes in variability over time since the X bar chart is based on the assumption that process variability is constant over time (Pitt, 1994; cited in Orme and Cox, 2001); see worksheet 3 to establish the charter Rd) Finally, the charter is constructed according to the following formulas which define the upper and lower control limits: (1) UCL = + A2 (2) LCL = - A2 For a size d sample n = 5, A2 = 0.577 (the constant A2 is tabulated for different sample sizes and is normally found in most statistical texts). g) Manually, or by Excel, or by any other sophisticated statistical software, the graphical presentation of the graph can be developed (see Triola 2004: 732 on how to produce the X bar graph with Excel).