Textual Analysis and Complex Construct Development in Management Accounting : the Case of Value-Based Management Sophistication

Complex constructs in management accounting are often hard to replicate if they rely on surveys or manual content analyses. Such constructs include, for example, the balanced scorecard (Kaplan & Norton, 1992), customer accounting (Roslender & Hart, 2010), corporate sustainability (Joshi & Li, 2016), and value-based management (VBM) (Burkert & Lueg, 2013). This study suggests a dictionary-based natural language processing (NLP) approach as an alternative to measuring complex constructs and explains how to apply NLP in management accounting to customize a dictionary. The construct of VBM sophistication serves as our example. We then illustrate how to test the construct’s convergent and predictive validity. Our illustrative construct demonstrates acceptable convergent and predictive validity. First, this study contributes to construct development and validation in management accounting research by exploiting a computerized measurement approach. Second, this study offers a guide for construct development using NLP methods that is applicable to other management accounting constructs. Third, we provide a VBM-specific dictionary for consequent studies in this field. Fourth, our results add to the body of evidence that claims positive economic effects from increasing VBM sophistication. Overall, our study generates future research opportunities to leverage NLP to improve the measurement of complex constructs