Some Problems When Adopting Churchill's Paradigm for the Development of Service Quality Measurement Scales