An Uninteresting Tour Through Why Our Research Papers Aren't Accessible
Abstract
Our research is delivered as Portable Document Format (PDF) documents, and very few include basic metadata to make them accessible to people with disabilities. As a result, many people are either unable to read them efficiently or at all. Over the past few years, we have tried everything from writing guidelines and giving accessibility feedback, to enforcing accessibility standards and volunteering to make PDFs accessible ourselves. The problem with making PDFs accessible is in part due to the lack of good tools, but the complexity of the PDF format makes improving tools difficult. Making accessible research papers is as much about our choices as a community: our choice of publication format, and our choice to make accessibility a voluntary task for authors. In this paper, we overview the context in which PDFs became our publication format, the difficulty in making PDF documents accessible given current tools, what we have tried to make our PDFs more accessible, and potential options for doing better in the future.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In HCI, only about 5% of publications have valid alt text, which is even worse in other fields like science, medicine, and engineering.
“Recent work [5, 14, 41] shows that in Human-Computer Interaction (HCI), where accessibility research is advocated, only about 5% of publications have valid alt text, which is even worse in other fields like science, medicine, and engineering [51, 54].”
From How University Disability Services Professionals Write Image Descriptions for HCI Figures Using Generative AI (2026) · cites this paper as [5]
Persistent challenges remain in ensuring that documents prepared in PDF consistently include the metadata necessary to make them accessible to people with disabilities.
“Despite this, persistent challenges remain in ensuring that documents prepared in this format consistently include the metadata necessary to make them accessible to people with disabilities [2, 3].”
From iTagPDF: Towards Finally Automating PDF Accessibility (2026) · cites this paper as [2]
Bigham et al. argued authors lacked sufficient incentives to engage with a complex remediation process, largely due to inadequate tooling.
“They argued authors lacked sufficient incentives to engage with a complex remediation process, largely due to inadequate tooling.”
From iTagPDF: Towards Finally Automating PDF Accessibility (2026) · cites this paper as [2]
Making PDFs automatically accessible is a messy problem requiring substantial engineering effort.
“Though making PDFs automatically accessible seems like a solvable problem, those who have worked on it have noted that it is a messy problem requiring substantial engineering effort [2].”
From iTagPDF: Towards Finally Automating PDF Accessibility (2026) · cites this paper as [2]
Prior works have extensively documented the challenges of using Acrobat, including the high cost of software license, an unintuitive user interface, and the low-quality of automated tags requiring authors to manually review and correct them.
“However, as discussed in the previous subsection, prior works [2, 17] have extensively documented the challenges of using Acrobat, including the high cost of software license, an unintuitive user interface, and the low-quality of automated tags requiring authors to manually review and correct them.”
From iTagPDF: Towards Finally Automating PDF Accessibility (2026) · cites this paper as [2]
Authors have long complained about the tedious process of PDF tagging, the difficulty of applying accessibility practices to complex visual elements, the poor usability of existing tools, and even knowing whether they are doing it correctly.
“They have long complained about the tedious process of PDF tagging, the difficulty of applying accessibility practices to complex visual elements, the poor usability of existing tools, and even knowing whether they are doing it correctly [2, 17].”
From iTagPDF: Towards Finally Automating PDF Accessibility (2026) · cites this paper as [2]
No citation lists for this paper
No citation lists for this paper yet — checked again daily.
References 18 · Cited by 34
- Authors
- Bigham, J. P., Brady, E. L., Gleason, C., Guo, A., & Shamma, D. A.
- Affiliations
- Carnegie Mellon University (US)
- Indiana University – Purdue University Indianapolis (US)
- Yahoo (United States) (US)
- Venue
- Proceedings of the 2016 CHI Conference Extended Abstracts on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2016-05-07
- Pages
- 621-631
- OpenAlex
- W2346371394
- Cited by
- 34
- References
- 18
- Citations by year
- 2026: 3 · 2025: 6 · 2024: 5 · 2023: 5 · 2022: 5 · 2021: 3 · 2020: 5 · 2019: 1 · 2017: 1
- FWCI
- 1.09
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) (publishedVersion)
- Funders
- National Science Foundation
- National Institute on Disability and Rehabilitation Research
- National Institute on Disability, Independent Living, and Rehabilitation Research
- Topics
- Social Sciences › Human Factors and Ergonomics › Digital Accessibility for Disabilities
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Mathematics, Computing, and Information Processing
- Social Sciences › Arts and Humanities › History and Philosophy of Science › Publishing and Scholarly Communication
- Keywords
- PDF accessibility, disability, publication formats, metadata, accessibility tools, scholarly publishing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review