Author Reflections on Creating Accessible Academic Papers
Rachel Menzies, Garreth W. Tigwell, Michael Crabb · 2022 · ACM Transactions on Accessible Computing · doi:10.1145/3546195
Summary
This mixed-methods study investigates why academic papers remain inaccessible despite available guidance from SIGACCESS and others. The researchers analyzed 3,866 content elements (tables, charts, images) from 330 ASSETS conference papers spanning 2011-2020, then conducted interviews with 13 academic authors ranging from PhD students to Professor Emeritus who publish in top-tier HCI and accessibility venues. The content analysis classified elements into 21 types across four categories. Images were most common (2,225 total), followed by charts (795), tables (662), and other elements like equations and code (184). The researchers documented accessibility challenges including: complex tables with merged cells and embedded images; charts using only color to distinguish categories; insufficient color contrast; missing or inadequate alt text; and misleading or inconsistent figure numbering. The interview study explored how authors approach content element creation, their motivations for accessibility, implementation challenges, and the balance between visual design and accessibility. Participants represented 33 combined years of accessibility research experience yet still faced significant challenges in producing accessible papers. This finding underscores the gap between knowing accessibility principles and successfully implementing them.
Key findings
Four main themes emerged from the interviews: **1. Inclusion is based on visual design:** Authors choose content elements primarily for visual presentation and storytelling, with accessibility considered secondary. Teaser images are seen as essential for attracting readers, but visual emphasis creates tension with accessibility requirements. **2. Motivation is grounded in awareness:** Authors are motivated by reaching more readers, making papers accessible for reviewers with disabilities, and personal or team experience with disability. However, awareness is often limited to screen reader users, with less consideration of cognitive disabilities or low vision. **3. Implementation is heterogeneous:** Accessibility depends heavily on individual skills, team expertise distribution, and available tools. Authors treat accessibility as "the last step"—similar to waterfall software development—rather than integrating it throughout the writing process. This approach makes retrofitting accessibility difficult and costly. **4. Balance between accessibility and visual design:** Authors face genuine tradeoffs between visual presentation and accessibility. Page limits force compression of images and tables, making them harder to access. Tools like Adobe Acrobat Pro required for PDF tagging are expensive and themselves inaccessible to some authors with disabilities. Critically, the publication pipeline often undoes accessibility work: publishers may strip alt text, add untagged front matter, or require formats that break accessibility features.
Relevance
This research provides actionable recommendations across the academic publishing ecosystem: **For authors:** Integrate accessibility throughout the writing process rather than treating it as a final step. Develop expertise beyond screen reader accessibility to address cognitive and low-vision needs. Ensure content elements don't require proprietary software to access (e.g., animations viewable only in Adobe Acrobat). **For conferences/publishers:** Include accessibility requirements in review criteria. Provide exemplar accessible papers and more granular, practical guidance. Ensure publication workflows preserve author accessibility efforts. **For researchers/industry:** Build better authoring tools that support accessible content creation. Current tools often produce inaccessible output or are themselves inaccessible to authors with disabilities. **For universities/funders:** Mandate accessibility training and require accessible outputs for funded research, similar to open access requirements. The paper's unique "Author Reflections" section models transparency by documenting the authors' own struggles with accessibility while writing this paper—including table design tradeoffs, figure tagging challenges, and reviewer feedback on inaccessible elements in their submission. This reflexive approach demonstrates that accessibility expertise doesn't eliminate implementation challenges.
Tags: document accessibility · PDF accessibility · academic publishing · alt text · data visualization · content elements · accessible authoring
Standards referenced: WCAG 2.1 · PDF/UA · ISO 14289-1