Two weeks ago I wrote about Diaz v. Lobels,* a case I think exemplifies some of the confusion concerning just what an accessible website should be. Today I want to take a harder look at the Court’s exclusion of testimony from the plaintiff’s expert to ask the question: Just what should an ADA expert testify about? The issue has come up in only three reported cases, each of which has its own take on the matter.
The exclusion of the plaintiff’s expert testimony in Diaz v. Lobels was based on two failings. The Court found the expert, Michael McCaffrey, failed to sufficiently describe his methodology and process, and failed to establish the methods were widely accepted or standard in the field. The Court did not reject the assumption underlying the report because the plaintiff and defendant stipulated that compliance with WCAG 2.0 or 2.1 would make the website accessible for ADA purposes and that non-compliance would make it inaccessible. Instead the Court focused on the “high level” description of the methods used by the team of individuals working for the expert, which the Court found insufficient when combined with a lack of testimony concerning the standards used.
This decision contrasts nicely with that in Gomez v. Gen. Nutrition Corp., 323 F. Supp. 3d 1368 (S.D. Fla. 2018) because in that case the Court accepted testimony from Mr. McCaffrey but rejected the testimony of the defendant’s expert, whose specialty was e-commerce and who did not know the details of how the error checking software he used worked. The Court was willing to grant summary judgment on liability not because of a lack of WCAG 2.0 compliance, but because the Mr. McCaffrey testified that the website had flaws that would prevent a handicapped person from selecting a product and completing a purchase.
A third case, Wu v. Jensen-Lewis Co., Inc., 345 F. Supp. 3d 438 (S.D.N.Y. 2018) finds the Court itself examining the website to determine whether the plaintiff’s complaint is plausible. The Court wrote:
Acknowledging that its judgment is no substitute for that of an expert’s, the Court finds that its review of the website is inconclusive as to whether Wu’s allegations are plausible. The Court found that images and image links consistently contained alt-text and did not identify any redundant links. The Court could neither verify nor contradict Wu’s allegations that Jensen-Lewis’s website contained empty links.
Here the Court, presumably using a software tool of some kind, looked at purely objective criteria – alt-text is there or it is not – but even so could not conclude one way or the other whether some supposed problems existed.
The decision in Diaz v. Lobel’s rests on the parties’ agreement that WCAG 2.0 would determine whether the website was or was not accessible. This eliminated the problem of deciding what accessible means, but plunged the court into the intricacies of the coding behind a web page, because a good deal of WCAG 2.0 testing is done by using software or humans to look at code behind the web page instead of the behavior of the web page itself. This problem goes back to the false analogy of websites to physical sites, because the “construction” of a website is several steps removed from the perception of the user. This means identifying certain objective characteristics – the absence of alt-text for example – is not sufficient to determine whether the site fails to conform to WCAG 2.x standards or will deny a user meaningful access. After all, WCAG 2.x does not require alt-text for purely decorative elements, but does require that the text serve the “equivalent purpose” of the image. Software can tell whether there is or is not a text alternative, but a human must determine the purpose of the image to know whether it is pure decoration or whether the text serves the same purpose. Thus, a reliable methodology for examining a web page’s compliance with WCAG 2.x must include a human looking at every image to determine whether it is pure decoration and, if not, whether any alternative text serves the same purpose.** And there is a further subtlety. Suppose the alternative text is not sufficient to serve the same purpose. Does that one failure deny a blind user meaningful access to the website? It might or might not depending on the purpose of the image. Last but not least, to have a reliable methodology requires more than simply delegating the task to humans; it requires giving the humans reliable criteria for making judgments about the website itself. Many websites will very clearly pass or fail, but no credible expert in this area argues that software analysis of a website is adequate, or that merely looking at the code is sufficient.
The General Nutrition Corp. case approaches the problem from the right angle; that is, is the website usable, but it is not at all clear what role an expert has to play when usability is the issue. In General Nutrition Corp. the expert testified that there were problems that would prevent a user from selecting a product and completing a purchase, but surely the plaintiff himself could testify to such problems because he or she would experience them directly. That’s what happened in Lobel’s, and made the exclusion of the expert’s report irrelevant at the summary judgment phase. Here the problem with expert testimony runs into the problem of standing, and in particular standing to seek injunctive relief. The real point of expert testimony in cases like Lobel’s and General Nutrition Corp. is not to show the website cannot be used; after all, the plaintiff can testify to that. The experts are brought in to support an injunction requiring a complete overhaul of the website when the plaintiff’s testimony might only support fixing one or two problems the plaintiff encountered. As I’ve often discussed, the question of what relief an individual plaintiff can seek with respect to problems never encountered has different answers in different Circuits, but the leading cases do not address standing in a website case.† In the Ninth Circuit it would certainly seem an expert whose job is to find usability problems never encountered by the plaintiff has relevant testimony to provide. In the Eighth it isn’t clear an expert has any role at all with respect to the existence of usability problems.
Wu v. Jensen-Lewis, in which the Court did its own investigation, shows the danger that comes from treating website accessibility as a matter of individual WCAG 2.x failures. The plaintiff seems to have pleaded the existence only of various discrete WCAG 2.x failures rather than an inability to meaningfully access the website’s goods and services. The Court followed suit, treating the inability to negate specific alleged failures as evidence that a valid claim had at least been pleaded. The minimum pleading standard for an ADA website case must be a plausible allegation that the plaintiff was denied meaningful access to the goods and services of the website, and whether the allegation of a specific WCAG 2.x failure satisfies Iqbal and Twombly depends on the kind of failure and its context. It also seems implausible that any ordinary plaintiff would be able to generate the kinds of lists of WCAG 2.x failures that frequently accompany ADA website complaints. Even in the most complex cases involving scientific evidence the plaintiff knows they harm they suffered. I would suggest that if the Complaint does not plausibly allege some specific inability of the plaintiff to meaningfully access the website then no matter how many WCAG 2.x failures are listed no ADA claim has been stated.
With the early settlement rate of website cases running at greater than 90%‡ the proper role of experts is going to evolve as courts confront the underlying problems of standing and the meaning of accessibility. Defendants should, however, be aware that there are real opportunities for those cases that must be litigated, and that an expert who can do no more than run a software scan is unlikely to pass muster for the plaintiff or defense.
* See, “ADA website accessibility litigation”
** See my earlier blog “ADA website litigation – is there a regulatory fix?”
† See “ADA standing and pleading – common sense from the 8th Circuit” An interesting and as yet unexplored issue concerns problems a disabled user will never know exist. Screen reader software will simply bypass some forms of pop-up window, leaving the user unaware the they were deprived of useful content. This is clearly an appropriate matter for expert testimony, and illustrates how different disabilities may require different approaches to ADA violation problems. A disabled person in a wheelchair is likely to know whether or not they have encountered a barrier to access. A blind person may never know what they missed if appropriate signs are missing. Because physical access cases are dominated by wheelchair using plaintiffs while website cases are dominated by blind users we will see a new set of standing and proof problems in website cases. This further illustrates the problem of treating websites as the metaphorical equivalent of physical places.
‡ See, Usablenet’s Mid-Year Update Usablenet Mid-Year Update 2019