Background:
Knowledge producers wishing to share research findings with their target audiences can be uncertain as to which format(s) (e.g. written summary, infographic, video, etc.) and which data visualization(s) (e.g. bar chart, line chart, pie chart, table, etc.) may be most appropriate. We searched the literature to identify evidence which could guide decisions about the design of knowledge products for different audiences. An understanding of the methods, outcomes, and outcome measures used in research evaluating information formats and data visualizations can help tailoring of knowledge products for improved understanding of findings.
Methods:
We performed a preliminary search of PubMed and Google Scholar to identify keywords which we then used for a more rigorous search across seven databases (PubMed, PsycINFO, IEEE Xplore, Embase, MEDLINE, Cochrane Library, UBC Library Database; 2011-2021 English language; primary, secondary, and tertiary sources) using 27 unique keywords. We extracted data from the full texts into a purpose-built spreadsheet specifying study method (e.g. qualitative, quantitative), outcomes (e.g. preference, comprehension, satisfaction, etc.), outcome measures (e.g. Likert scale questions, open-ended questions, response accuracy, etc.), type of audience (e.g. lay, healthcare professional, researcher, policymaker, etc.), and findings. A descriptive analysis was subsequently undertaken.
Results:
The literature review yielded findings on commonly used information formats (infographics, videos, policy briefs, lay summaries, plain-language summaries, research summaries, narrative summaries, and critical appraisals) and data visualizations (bar charts, pie charts, donut charts, line charts, scatter charts, tables, network analyses, and pictographs). A total of 36 articles were retrieved: 24 articles on information formats (19 original studies and 5 reviews), and 12 articles on data visualizations (10 original studies, 1 review, and 1 methods paper reporting novel graphic representation of qualitative data).
Conclusions:
Considerable variability in audiences, methods, outcomes, and outcome measures exist within the literature and thus extrapolating findings is challenging. Nevertheless, visual formats, such as videos or infographics, are generally preferred over written formats, except by researchers and policymakers who prefer critical appraisals and policy briefs respectively. In terms of data visualizations, although bar charts and pie charts are commonly preferred, it is important to note that preference does not always correspond well with readers’ comprehension (as tested by response accuracy and response time) and thus preference alone should not dictate which data visualization is selected. There is compelling evidence to select types of data visualizations based on the intended purpose as each type of visualization appears to be better suited for conveying specific types of information such as pie charts for identifying proportions, scatter plots for highlighting anomalies, and line charts for depicting trends. Finally, given the wide spectrum of numerical and graphical literacy within the population, it is likely prudent to provide a combination of text, numerical, and visual representations of findings. This review elucidated recommendations for knowledge producers to consider when selecting information formats and data visualizations to share research findings with different audiences.