Although the SERVQUAL instrument has been widely applied in a variety of industries and cross-cultural contexts, there are many criticisms of the approach.
Francis Buttle published one of the most comprehensive criticisms of the model of service quality and the associated SERVQUAL instrument in 1996, in which both operational and theoretical concerns were identified. Some of the more important criticisms include: :
Face validity: The model of service quality has its roots in the expectancy-disconfirmation (EDT) paradigm that informs customer satisfaction. Many researchers have argued that the research instrument actually captures
satisfaction rather than
service quality. Other researchers have questioned the validity of conceptualising service quality as a gap. :
Construct validity: The model's developers tested and retested the SERVQUAL scale for reliability and validity. However, at the same time, the model's developers recommended that the applied use of the instrument should modify or adapt it for specific contexts. Any attempt to adapt or modify the scale will have implications for the validity of items with implications for the validity of the dimensions of reliability, assurance, tangibles, empathy and responsiveness. :
Ambiguity of expectations construct: SERVQUAL is designed to be administered after respondents have experienced a service. They are therefore asked to
recall their pre-experience expectations. However, recall is not always accurate and raises concerns about whether the research design accurately captures true pre-consumption expectations. In addition, studies show that expectations actually change over time. Consumers are continually modifying their expectations as they gain experience with a product category or brand. :
Operational definition of the expectations construct: The way that expectations have been operationalized also represents a concern for theorists investigating the validity of the gaps model. The literature identifies different types of expectations. Of these, there is an argument that only
forecast expectations are true expectations. Yet, the SERVQUAL instrument appears to elicit
ideal expectations. Note the wording in the questionnaire in the preceding figure, which grounds respondents in their expectations of what
excellent companies will do. Subtle use of words can elicit different types of expectations. Capturing true expectations is important because it has implications for service quality scores. When researchers elicit ideal expectations, overall service quality scores are likely to be lower, making it much more difficult for marketers to deliver on those expectations. :
Questionnaire length: The matched pairs design of the questionnaire (total of 22 expectation items plus 22 perception items = 44 total items) makes for a very long questionnaire. If researchers add demographic and other behavioral items such as prior experience with product or category, and the standard battery of demographics, including age, gender, occupation, educational attainment etc., then the average questionnaire will have around 60 items. In practical terms, this means that the questionnaire would take more than one hour per respondent to administer in a face-to-face interview. Lengthy questionnaires are known to induce
respondent fatigue, which may have potential implications for data reliability. In addition, lengthy questionnaires add to the time and cost involved in data collection and data analysis. Coding, collation and interpretation of data is very time-consuming, and in the case of lengthy questionnaires administered across large samples, the findings cannot be used to address urgent quality-related problems. In some cases, it may be necessary to carry out 'quick and dirty' research while waiting for the findings of studies with a superior research design. :
Administration of the questionnaire: Some analysts have pointed out that the SERVPERF instrument, developed by Cronin and Taylor, and which reduced the number of questionnaire items by half (22 perceptions items only), achieves results that correlate well with SERVQUAL, with no reduction in diagnostic power, improved data accuracy through reductions in respondent boredom and fatigue and savings in the form of reduced administration costs. :
Dimensional instability: Many studies have reported that the five dimensions of service quality implicit in the model (reliability, assurance, tangibles, empathy and responsiveness) do not hold up when the research is replicated in different countries, different industries, in different market segments or even at different time periods. Some studies report that the SERVQUAL items do not always load onto the same factors. In some empirical research, the items load onto fewer dimensions, while other studies report that the items load onto more than five dimensions of quality. In statistical terms, the robustness of the factor loadings is known as a model's
dimensional stability. Across a wide range of empirical studies, the factors implicit in the SERVQUAL instrument have been shown to be unstable. Problems associated with the stability of the factor loadings may be attributed, at least in part, to the requirement that each new SERVQUAL investigation needs to make context-sensitive modifications to the instrument in order to accommodate the unique aspects of the focal service setting or problem. However, it has also been hypothesised that the dimensions of service quality represented by the SERVQUAL research instrument fail to capture the true dimensionality of the service quality construct and that there may not be a universal set of service quality dimensions that are relevant across all service industries. In spite of these criticisms, the SERVQUAL instrument, or any one of its variants (i.e. modified forms), dominates current research into service quality. In a review of more than 40 articles that made use of SERVQUAL, a team of researchers found that "few researchers concern themselves with the validation of the measuring tool". SERVQUAL is not only widely covered as a subject of academic papers, but also widely used by industry practitioners. == See also==