We now have another research paper demonstrating that a Web site that meets accessibility guidelines is more usable by nondisabled people. “The Benefit of Accessible Design for Able-Bodied Users of the World Wide Web” lists a number of performance improvements of accessible sites when used in Firefox, Lynx, and Firefox reduced to the size of a handheld device (a rather poor substitute). I find the presentation of the numerical results nearly impossible to read, but the results are there.
The benefit of this study is that it does not confound Web accessibility with cluefulness of the developer. Why are accessible sites usable to everyone? Presumably because, by and large, only the best-informed and most mature developers work on accessible sites. They are pretty much incapable of producing an unusable site. Plus their HTML is way better. In this study, an existing crappy site was recoded, but was otherwise left crappy. This provides evidence that Web standards lead to better usability. Better HTML has an influence on usability.
Of great interest was the fact that identical font sizes led to different complaints. In the original coding, the font size was deemed too small by some people, but the same font size in standards-compliant coding was deemed too big by some. Perplexing to say the least. Really, we have no viable research on people’s true reactions to Web typography, and absolutely none about low-vision people’s reactions, which are especially influenced by the flatly atrocious type rendering of most screen magnifiers.
The other papers that demonstrated a similar effect are the DRC survey of Web accessibility (still not online in HTML at the DRC site) and the oft-reiterated results from Legal & General, which have never been directly reported. Maybe Nielsen has something along the same lines, and Fidelity probably has internal research they won’t publish.