📖 Daily Journal | Nonprofessional Investor Judgments: Linking Dependent Measures to Constructs

✍🏻️ Author: H. Scott Asay, Jeffrey Hales, Cory Hinds, Kathy Rupar

📖 Source: The Accounting Review, Volume 98 Issue 7, November 2023

Introduction

This article explores the construct validity of dependent measures commonly used in research on nonprofessional investor judgments. The authors conducted a survey of 90 financial experimental research articles and used factor analyses to identify three underlying constructs related to the judgments of nonprofessional investors: (1) expectations regarding future firm performance and value, (2) holistic perceptions of the firm, and (3) evaluations of the risk associated with investing in the firm.

Discussion

The authors conducted a survey of 90 financial experimental research articles published between 1990 and the first quarter of 2020 within six top accounting journals. They identified research relating to individual investors and cataloged the dependent measures collected in each study.

To analyze the data, the authors conducted exploratory factor analyses (EFA) and confirmatory factor analyses (CFA) to uncover linkages between dependent measures and the constructs underlying nonprofessional investor judgments. They used the EFA to identify the underlying constructs and the CFA to test the fit of the identified constructs to the data.

The authors also provide recommendations for selecting, analyzing, and reporting dependent measures in future research. They suggest that researchers should draw on theory to specify a dependent construct, create a measurement scale, provide evidence of unidimensionality, report descriptive statistics, and use the arithmetic mean to combine dependent measures to test hypotheses. They argue that this five-step process should improve the consistency in conducting and reporting analyses, increase measurement scale reliability, and constrain selective reporting of dependent measures.

Conclusion

The authors conclude that using a theory-driven construct and data-driven measures of that construct should increase statistical power and improve the consistency in conducting and reporting analyses. They also provide several important directions for future research that could increase our understanding of individual investor judgments and provide additional insight to guide future choices of dependent measures.