The Transparency of Automatic Web Accessibility Evaluation Tools


Accessibility validation tools, Transparency, Guidelines

1. Problem Addressed

Following the adoption of accessibility laws, many public organizations started paying more attention to accessibility guidelines. However, Web accessibility requires constant monitoring of numerous details across many pages in a given site. Thus, to simplify the monitoring, analysis, detection, and correction of website accessibility problems, several automatic and semi-automatic tools have been proposed. Even though accessibility validation is a process that cannot be fully automated [10], automatic tools [1] still play a crucial role in ensuring the accessibility of websites. They help human operators collect and analyse data about the actual application of accessibility guidelines, detect non-compliance, and provide relevant information about addressing the possible problems

2. Relevant background

Our group has long experience in tools for accessibility evaluation [2, 7, 8]. We have also participated in the Wadcher EU project , with the aim to develop a set of tools to monitor large scale accessibility. While discussing with users of such tools and other tool developers, we often noticed that such tools differ in their coverage of accessibility guidelines, in how they interpret and to what extent they are able to support them, and in the design of how they present the results, including errors (and likely errors). Such differences are perceived in different ways by users, are sometimes misinterpreted, and can generate misunderstandings. Better explaining these differences can help create tools that assist web site designers and developers in making informed decisions, and indicate gaps that could be addressed in future versions of these tools (or in new tools). Unfortunately, this issue has not been sufficiently dealt with in previous studies of accessibility tools such as [3, 5, 10]. We have thus introduced [6] the concept of transparency of such tools, as well as some criteria that could be used to analyse it, and provided an initial comparative analysis of four validation tools according to them.

3. Challenges

By the transparency of an accessibility validation tool we mean its ability to clearly indicate to its users what accessibility aspects it is able to validate and the meaning of the results generated. The various available tools follow different approaches to checking accessibility, and have to keep up with the continuous evolution of Web technologies and their use, which imply the need to continuously update their support for the validation of the associated accessibility guidelines. Users are sometimes not even aware of such differences, and they may become disoriented when they see different results in terms of validation. Thus, it is important to make them more aware of this issue, and provide tool developers with indications for making their accessibility tools more transparent. To some extent, we face similar issues to those that people are encountering with the increasing deployment of Artificial Intelligence (AI) tools, which often generate various problems for their users since they do not explain why they are operating in a certain way. Thus, interest in techniques for explainable AI has been increasing in recent times. In this perspective, some researchers have explored the space of user needs for explanations using a question-driven framework. For example, some authors [9] propose a question bank in which user needs for explainability are represented as prototypical questions users might ask about the AI, such as “Why is this instance given this prediction?”, “What would the system predict if this instance changes to …?” Some of such questions can still be relevant for accessibility validation tools, even when they do not use AI methods at all.

4. Outcomes

As stated, transparent tools should enable users to make fully informed decisions based on a clear understanding of how the automatic validation tools work. In particular, in order to be transparent, an automated validation tool should make explicit the following information:

5. Future perspectives

We are carrying out work on empirical validation of the issues associated with transparency, in order to define more precisely such concept and the associated criteria, provide an analysis of a broader set of accessibility validation tools according to such criteria, with the ultimate goal of providing some recommendations for tools developers to improve the transparency of their tools. In this work we consider people who have used accessibility validation tools from different perspectives: web commissioners (people who mainly decide and manage the content of a Web site), accessibility experts (those who are in charge of actually checking whether an application is accessible), and web developers. We will be happy to share and discuss initial results of this new work at the workshop.


  1. W3C, Web Accessibility Evaluation Tools List,
  2. Giovanna Broccia, Marco Manca, Fabio Paternò, Francesca Pulina, Flexible Automatic Support for Web Accessibility Validation. Proc. ACM Hum. Comput. Interact. 4(EICS): 83:1-83:24 (2020)
  3. Andreas Burkard, Gottfried Zimmermann, and Bettina Schwarzer. 2021. Monitoring Systems for Checking Websites on Accessibility. Frontiers in Computer Science 3 (2021), 2.
  4. 4Siddikjon Gaibullojonovich Abduganiev. 2017. Towards automated web accessibility evaluation: a comparative study. Int. J. Inf. Technol. Comput. Sci.(IJITCS) 9, 9 (2017), 18–44.
  5. Marian Pădure and Costin Pribeanu. 2019. Exploring the differences between five accessibility evaluation tools. (2019).
  6. P. Parvin, V. Palumbo, M. Manca, F. Paternò. 2021. The Transparency of Automatic Accessibility Evaluation Tools. In Proceedings of the 18th International Web for All Conference (W4A ’21), April 19–20, 2021,
  7. Ljubljana, Slovenia. ACM, New York, NY, USA, 5 pages. 1145/3430263.3452436
  8. Paternò F., Schiavone A., The role of tool support in public policies and accessibility. ACM Interactions 22(3): 60-63 (2015)
  9. Schiavone A., Paternò F., An extensible environment for guideline-based accessibility evaluation of dynamic Web applications, Universal Access in the Information Society, Springer Verlag, 14(1): 111-132, 2015.
  10. Q. Vera Liao, Daniel M Gruen, Sarah Miller, Questioning the AI: Informing Design Practices for Explainable AI User Experiences, CHI 2020
  11. Markel Vigo, Justin Brown, and Vivienne Conway. 2013. Benchmarking web accessibility evaluation tools: measuring the harm of sole reliance on automated tests. In Proceedings of the 10th International Cross-Disciplinary Conference on Web Accessibility. 1–10.