Reflections on Perceived Online Service Quality: Structure, Antecedents, Ontology, Theory and Measurement

Online services are becoming increasingly ubiquitous, and this growth has been accompanied by increased business interest in measuring and managing online service quality. This interest is also reflected in a large number of academic studies. Despite this, there is very little consensus about the dimensions and antecedents of perceived online service quality (POLSQ). We consider two possible reasons for this: first that the phenomenon of online service quality is changing as new technology affordances arise, so instability in the dimensionality would be a result of changes in the underlying phenomenon. Second, the theoretical approach and assumptions that studies of online service quality are usually founded on is flawed. My research questions are: 1) what is the structure of perceived online service quality? 2) What are the antecedents of perceived online service quality? 3) What is the ontology of perceived online service quality? 4) What are the most appropriate modelling and measurement methods for measuring online service quality quantitatively, and what insights can be gained from psychometrics?5) What insights does this offer IS researchers for the measurement of user attitudes and perceptions towards technologies? We find that leading models and instruments tend to be based on exploratory factor analysis and have not been informed by advances in measurement theory, particularly co-variance-based structural equation models, and a sub-set of those models known as multi-indicator structural models. We apply recent advances in measurement theory to a dataset. We apply and compare four different modelling methods, exploratory factor analysis, confirmatory factor analysis, formative models, and multi-indicator structural models, and discuss the theoretical foundations of each method. We conclude that POLSQ may not have a separate ontology as a multi-dimensional construct, but overall affect towards the service quality of a website is likely to be the result of the perceived usefulness, perceived ease of use, and perceived trust towards the service. This finding supports the explanatory power of information systems theories over marketing and consumer behaviour theories when studying this phenomenon. We find that user's perceptions of detailed affordances of the service, such as the relevance and timeliness of the information are antecedent to overall affect towards the service, rather than being additional dimensions of POLSQ. We find that widely used techniques such as exploratory factor analysis have serious drawbacks. We find that multi-indicator structural models provide an accurate and nuanced method for modelling the formation of attitudes and perceptions towards technology, which is also well-grounded in theory from social psychology. Finally, we suggest that the approach we take to measuring POLSQ is has potential value for other research which aims to measure customer attitudes and perceptions towards technologies.