Toward a continuum of measurement scales in Just-in-Time (JIT) research – an examination of the predictive validity of single-item and multiple-item measures

Abstract

We compare the predictive validity of single-item and multiple-item measures utilized in Just-in-Time (JIT) research. The study examines if single-item measures could be used for some of the JIT practices, especially if the object of inquiry is concrete singular and if the attribute to be researched is concrete. Arguments are developed for the concrete nature of the JIT practice of “set-up time reduction” and we examine the ability of a single-item measure of this variable to predict the criterion variable (delivery performance). In addition, the study also examines the efficacy of using multiple-item measures for variables that are abstract in nature, and thereby attempts to develop a continuum of JIT constructs ranging from concrete to abstract. The results obtained by analyzing two sets of survey data show that multiple-item measures are not necessarily more valid than single-item measures for all constructs. The findings provide evidence that multiple-item measures and single-item measures for scale development should be contingent upon the nature of constructs. For concrete constructs, single-item measures are as valid as multi-item measures. Meanwhile, for abstract constructs it is important to ensure that multiple items are considered to capture the multi-dimensional nature of these constructs. Results also reveal that JIT practices display significant differences in terms of abstract/concrete perceptions. The paper presents theoretical and practical implications of the findings, and offers directions for future research.