Interactive data visualisation for accounting information: a three-fit perspective

ABSTRACT

The volume of freely available accounting information is rapidly becoming overwhelming. To be useful, information needs to be delivered to users in a suitable, relevant, and understandable form. Interactive data visualisation (IDV) can help address this need for useful information by organising accounting information, especially financial reports, into forms with these qualities. Given both their prevalence and their likelihood of being future users of IDV, the purpose of this research is to examine the appropriateness of IDV for non-professional investors’ use when they access accounting information. This research uses a 2 × 2 experimental approach involving 404 participants representing non-professional investors from diverse demographic backgrounds. This research suggests that IDV mitigates non-professional investors’ restricted investment capabilities by presenting information that is more salient, thus reducing non-professional investors’ cognitive effort. This combination allows such investors to better perform both simple and multipart investment tasks. By integrating three information systems’ fit perspectives (i.e. task technology, information quality, and cognitive), this research explains IDV’s suitability and fit within the accounting domain. We also discuss how the findings can inform practice and span interdisciplinary research into data and information visualisation.

KEYWORDS:

• Interactive media
• data visualisation
• cognitive theories
• task-technology fit
• information visualisation
• visual analytics

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Arif Perdana http://orcid.org/0000-0002-1260-240X

Alastair Robb http://orcid.org/0000-0002-3177-7854

Notes

1. Given the exploratory nature of this study, we assessed the effect of IDV as a whole, rather than assessing which components of IDV contributed most to the non-professional investors’ performance, e.g. interactivity, visualisations, etc.

2. Inherent quality reflects the quality of the characteristics or attributes of the data or information in their own right (Wang and Strong 1996). Wang and Strong (Wang and Strong 1996) propose four dimensions the inherent quality of data and information i.e. accuracy, believability, objectivity, and reputation. Jayawardene, Sadiq, and Indulska (2013) suggest that inherent quality reflects the representational nature of the data or information and remains dependent on users’ perceptions.

3. In the analyses, therefore, the cognitive load measurement has to be reverse coded.

4. Relative to those tests, our 404 sample size is sufficient to satisfy the multivariate normality assumption (Tabachnick and Fidell 2007). The correlation coefficient (0.747) between each variable (TTF, CF and perceived IQ) was below the threshold for multicollinearity (<0.80) (Stevens 2009). The result of Box’s test of equality of covariance matrices indicated a violation of the assumptions of the equal variance of covariance (p < .05). The randomised and even participant allocation mitigated this violation thus permitting robust interpretation of the data (Hair et al. 2010; Tabachnick and Fidell 2007). The results of Levene’s test for equality of error variance indicated a violation of the assumption of equality of error variance (p < .05). We mitigated this violation by the use of a more stringent alpha when accepting or rejecting the MANOVA hypotheses (= 0.001) (Tabachnick and Fidell 2007).