Graphs as an Impression Management Tactic: Evidence from JSE Top 40 Integrated Reports Before, During and After a Crisis

Integrated reports rely heavily on graphs, yet the content outside the annual financial statements is not audited, which leaves room for graphs to be used to manage users’ impressions. Crisis conditions are expected to increase the pressure on management to protect legitimacy, but little is known about how graphical impression management changes before, during and after a crisis, and prior studies usually rely on a single distortion index. This study examines whether the top 40 JSE-listed companies use graphs as an impression management tactic, and whether graph frequency, selectivity, and measurement distortion differed across the periods before (2018 to 2019), during (2020 to 2021), and after (2022 to 2025) the COVID-19 crisis. A quantitative content analysis was conducted on 8427 graphs across 34,720 report pages from 200 integrated reports. Distortion was measured by the Graph Discrepancy Index (GDI) and the Relative Graph Discrepancy Index (RGD), and differences between periods were evaluated using chi-square tests. Graph density rose during the crisis and returned to its pre-crisis levels afterwards, while the shares of pie charts and non-financial graphs increased during the crisis and remained higher afterwards. Favourable key financial variable (KFV) trends outnumbered unfavourable ones in every period, and their share rose from 62.5% before the crisis to 87.1% after it, which may partly reflect the recovery in company results. The GDI classified 54.9% of the measurable KFV graphs as materially distorted, compared with 19.1% under the RGD, but across indices, material distortion did not favour management more often than it worked against it. Selectivity appears to be a stronger impression management tactic than measurement distortion. This study extends the graph disclosure literature with insights from the crisis setting in South Africa, showing that the choice of distortion index can materially change conclusions about how common distortion is.