Eye-Tracking Experiments in Social and Environmental Accounting Research

ABSTRACT

In this article, we demonstrate the relevance of eye-tracking experiments in social and environmental accounting (SEA) research. Up to now, this type of design has been used in some areas within accounting research, but SEA has been neglected. If one is to adopt a user perspective [Merkl-Davies, D. M., and N. M. Brennan. 2007. “Discretionary Disclosure Strategies in Corporate Narratives: Incremental Information or Impression Management?” Journal of Accounting Literature 27: 116–196; 2011. “A Conceptual Framework of Impression Management: New Insights from Psychology, Sociology and Critical Perspectives.” Accounting and Business Research 41 (5): 415–437], the investigation and the understanding of the way social and environmental information affects user perceptions and decisions requires, among other tools, the use of eye-tracking setups. We discuss the need for eye-tracking experiments in SEA research and provide some preliminary evidence on their usefulness by conducting an illustrative experiment.

Acknowledgements

The authors wish to thank Matias Laine (handing Co-Editor), Carlos Larrinaga (Co-Editor and discussant at the 37th EAA Annual Congress), two anonymous reviewers, and the participants of the 2nd French Congress on Social and Environmental Accounting Research in Montpellier, the 25th International Congress on Social and Environmental Accounting Research in Saint Andrews, the 2nd Critical Studies in Accounting and Finance Conference in Abu Dhabi, the 2014 AAA Public Interest Section Mid-Year Meeting in San Diego, the 37th European Accounting Association Annual Congress in Tallinn, research seminar at Bangor University and brown bag series at the ESSEC Research Center for Capitalism, Globalization and Governance (C²G²).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. According to Parker (2005), experimental studies represent only about 1% of the SEA literature.

2. Since then, while some new experimental studies have been conducted (e.g. Elijido-Ten 2011; Wynder, Wellner, and Reinhard 2013; Dilla et al. 2014; Alewine and Stone 2016), Alewine’s (2010) analysis remains true – there are still very few experimental studies in the SEA literature.

3. See Appendix 1 for a summary of previous research on eye-tracking experiments.

4. Eye-tracker devices currently sold in the market are becoming gradually reliable due to technological innovations (Wedel and Pieters 2007).

5. Studies on shareholders are primarily based on investment decisions and share price reactions.

6. Merkl-Davies and Brennan (2011, 431) note that ‘there is little evidence on the information acquisition strategies undertaken by different strata of that audience’.

7. Researchers also use mouse clicks in the domain of internet search to proxy for users’ attention for example but they provide much less detailed data than eye-tracking (Cutrell and Guan 2007).

8. For example, Reimsbach and Hahn (2015) report, for a subgroup of experiment participants, an inconsistency between actual and perceived behaviour. They ask participants to report whether the CSR information they read affected their judgments of stock price. A sample of 27.3% of the sample indicated that the additional sustainability information had no effect on their estimates while only 16.8% did not revise their initial assessments.

9. Some studies analyse the influence of specific S&E information release such as greenhouse gases (Matsumura, Prakash, and Vera-Muñoz 2014) but information is extracted from other channels than corporate reports (Carbon Disclosure Project in this example) and is isolated from other S&E information areas.

10. This is compared to running an additional experiment to distinguish between the three areas.

11. Confirmation (or confirmatory) bias is the tendency to seek out information in a way that reinforces one's preexisting hypotheses and ignore information that is not in accordance with prior beliefs.

12. Trusting beliefs, in McKnight, Choudhury, and Kacmar’s (2002) e-commerce framework, are perceptions of favorable website attributes, including ‘competence’, ‘benevolence’, and ‘integrity’ (McKnight, Cummings, and Chervany 1998).

13. Trusting intentions are more closely related to the intention to engage in trust-related behaviors; that is, greater trusting intentions increase the likelihood that an individual will depend upon the other party (McKnight, Choudhury, and Kacmar 2002).

14. A vast stream of literature on legitimacy theory has proposed that these environmental and social disclosures are often centered on organizations’ concerns regarding their legitimacy. Legitimacy theory is rooted in the idea of the social contract, or that a company's continued existence is predicated on being perceived as operating in a ‘legitimate’ or socially acceptable manner, by the rest of society (Deegan 2002).

15. Institutional Review Board approval was obtained from one of the authors’ institution prior running the experiment.

16. Given that we used the same eye-tracking machine as Atalay, Bodur, and Rasolofoarison (2012), we provide here a slightly condensed version of the machine specificities.

17. This experiment was pilot tested with six graduate students under normal experimental conditions. Based on the feedback from these pilot tests, several changes were made to the wording of items on the questionnaire and to the resolution of the CSR report text.

18. Trusting intentions signified participants’ willingness to rely upon information in the CSR report, whereas trusting beliefs signified participants’ integrity trusting belief in the company. Following McKnight, Choudhury, and Kacmar (2002), nine questions assessing these trust dimensions, each on a 7-point scale, were included on the questionnaire (see Appendix 2). Factor analysis indicated that following McKnight, Choudhury, and Kacmar (2002) and Cho et al. (2009), four of these questions loaded onto the Trusting Beliefs variable, whereas five questions loaded onto the Trusting Intentions variable. The dependent variable measures for this study used the summed response to these measures.

19. We follow Atalay, Bodur, and Rasolofoarison (2012) and use 100 milliseconds as the cut-off, since fixations below this threshold are unlikely to add any information acquisition.

20. It is one of the fastest movement a human body can achieve – up to 1000 eye movements/second (Wedel and Pieters 2014).

21. For a more detailed description of the eye functioning, see Wedel and Pieters (2007) for example.

Additional information

Funding

We also thank the HEC Paris GREHEC Laboratory (CNRS UMR 2959) for providing access to the Tobii eye-tracking machine and Charles Cho acknowledges the financial support provided by the ESSEC Research Centre (CERESSEC) and the Global Research Network program through the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea [NRF-2016S1A2A2912421].