Thinking with data visualisations: cognitive processing and spatial inferences when communicating climate change

Harold, Jordan (2017) Thinking with data visualisations: cognitive processing and spatial inferences when communicating climate change. Doctoral thesis, University of East Anglia.

[thumbnail of Jordan_Harold_PhD_Thesis_Final.pdf]
Preview
PDF
Download (4MB) | Preview

Abstract

Data visualisations can be effective for communicating scientific data, but only if they are understood. Such visualisations (i.e. scientific figures) are used within assessment reports produced by the Intergovernmental Panel on Climate Change (IPCC). However, IPCC figures have been criticised for being inaccessible to non-experts. This thesis presents a thematic analysis of interviews with IPCC authors, finding that a requirement to uphold scientific accuracy results in complex figures that are difficult for non-experts to comprehend, and which therefore require expert explanation. Evidence is subsequently presented showing that figures with greater visual complexity are associated with greater perceived comprehension difficulty among non-experts. Comprehension of complex data visualisations may require readers to make spatial inferences. When interpreting a time-series graph of climate data, it was found that non-experts did not always readily identify the long-term trend. Two experiments then show that linguistic information in the form of warnings can support spatial representations for trends in memory by directing visual attention during encoding (measured using eyetracking). This thesis also considers spatial inferences when forming expectations about future data, finding that expectations were sensitive to patterns in past data. Further, features that act on bottom-up perceptual processes were largely ineffective in supporting spatial inferences. Conversely, replacing spatial inferences by explicitly representing information moderated future expectations. However, replacing spatial inferences might not always be desirable in real-world contexts. The evidence indicates that when information is not explicitly represented in a data visualisation, providing top-down knowledge may be more effective in supporting spatial inferences than providing visual cues acting on bottom-up perceptual processes. This thesis further provides evidence-based guidelines drawn from the cognitive and psychological sciences to support climate change researchers in enhancing the ease of comprehension of their data visualisations, and so enable future IPCC outputs to be more accessible.

Item Type: Thesis (Doctoral)
Faculty \ School: Faculty of Social Sciences > School of Psychology
Depositing User: Chris White
Date Deposited: 19 Jul 2018 15:46
Last Modified: 31 Dec 2019 01:38
URI: https://ueaeprints.uea.ac.uk/id/eprint/67669
DOI:

Actions (login required)

View Item View Item