Automated text analyses of sustainability & integrated reporting. A literature review of empirical-quantitative research

Abstract

Purpose This study aims to focus on automated text analyses (ATAs) of sustainability and integrated reporting as a recent approach in empirical–quantitative research. Design/methodology/approach Based on legitimacy theory, the author conducts a structured literature review and includes 38 quantitative peer-reviewed empirical (archival) studies on specific determinants and consequences of sustainability and integrated reporting. The paper makes a clear distinction between analyses of reports due to readability, tone, similarity and specific topics. In line with prior studies, it is assumed that more readable reports with less tone and similarity relate to increased reporting quality. Findings In line with legitimacy theory, there are empirical indications that specific corporate governance variables, other firm characteristics and regulatory issues have a main impact on the quality of sustainability and integrated reporting. Furthermore, increased reporting quality leads to positive market reactions in line with the business case argument. Research limitations/implications The author deduces useful recommendations for future research to motivate researchers to include ATA of sustainability and integrated reports. Among others, future research should recognize sustainable and behavioral corporate governance determinants and analyze other stakeholders’ reactions. Practical implications As both stakeholders’ demands on sustainability and integrated reporting have increased since the financial crisis of 2008–2009, firms should increase the quality of reporting processes. Originality/value This analysis makes major contributions to prior research by including both sustainability and integrated reporting, based on ATA. ATAs play a prominent role in recent empirical research to evaluate possible drivers and consequences of sustainability and integrated reports. ATA may contribute to increased validity of empirical–quantitative research in comparison to classical manual content analyses, especially due to future CSR washing analyses.

Exploring visual communication in corporate sustainability reporting: Using image recognition with deep learning

“…The original materials for this analysis were shown in Chapter 14 of Nakao et al (2023) and Table A1 and Table 2. 3 Some accounting and organisational research studies have employed machine learning (Fieberg et al, 2022;Hannigan et al, 2019;Hu & Sun, 2022;Krupa & Minutti-Meza, 2022;Velte, 2023) We then identify the detected faces by gender (men or women), and we categorise the age groups as follows: 0-10 years as 'children', The descriptive statistics for face detection in Amazon Rekognition show more men than women, a predominant 'middle-aged' group, 33% smiles and 10% beards, marking the first detailed study on facial images in sustainability reporting. Table A2 provides further details on the descriptive statistics.…”

Integrated Reporting and Group Decision-Making for Sustainable Development: A Bibliometric Review

“…Recently, there has been a growing practice in organizations of disclosing information that goes beyond the traditional limits of conventional reports. This trend is because traditional financial reports are not able to meet the complex information needs of various stakeholders (Velte, 2023). In addition to financial reports, sustainability reports were introduced as a means of attracting investors.…”