Analysis of the Use of Mediating and Moderating Variables in Management Accounting System (MAS): Evidence from the Literature Survey

Abstract

This study examined the nature of relationship between variables studied in the management accounting literature and the types of data analysis techniques used to analyse different types of moderation and mediation effects. To accomplish the stated objectives, the study first begins with the conceptualization of the nature of relationship between the research variables and clearly explained their differences. The study also reviewed 87 management accounting journal articles from 38 Journals that were published in between 2003 and 2015. The Journal articles were also categorized into four (4) groups, namely the ones that focused on management control, cost accounting, intellectual resource management and ‘others' and they were analysed using the descriptive statistics (percentages). The findings revealed that majority of reviewed Journal articles focused on the direct relationship among the research variables. The relatively few studies extensively focused on the mediation effect and moderation effect among variables. In addition, the hierarchical regression analysis, SEM including PLS, multiple regression, and OLS are the widely data analysis techniques used in analyzing different types of mediation and moderation effects with little attention on the measurement error bias. Therefore, the study argues for the future management accounting literature to focus on the mediated-moderation effect and moderated-mediation effect while considering the measurement error bias.