Reliability assessment and optimization of marketing measurement

Different purposes of measurement reflect corresponding differences in marketing problems underlying the need for measurement. Reliability assessment in marketing, which is based on classical reliability theory, continues to focus on scaling of individual respondents and ignores other purposes of measurement. Besides the shortcomings in reliability assessment, the academic literature does not make any recommendations for efficiencies in scale usage by marketing firms in applied settings. Generalizability theory has long been identified as an alternative psychometric approach that can be applied to design efficient measurement and can explicitly take into account differences in the purpose of measurement. The authors take a generalizability theory perspective to (1) reiterate the rationale for taking into account the purpose of measurement when assessing reliability, (2) propose a general method to optimize the design of marketing measurement in applied studies, and (3) conduct an empirical study on service quality measurement that illustrates the dual advantages of optimizing measurement and the generalizability approach to marketing measurement.