Today’s blog post features a 2011 article called “Analyzing Written Communication in AAC Contexts: A statistical perspective” by Lorenzo Bernardi and Arjuna Tuzzi. The article was sent to us by a proponent of FC (thank you) as “proof” that FC works. The article is already mentioned on our website in a couple different places (see Systematic Reviews and Misleading Articles). I’ll provide specific references below. In fairness to our commenter, however, I reread the Bernardi and Tuzzi article and critiqued it the same as I would any other article.

I should note that, while this study features touch-based FC, I believe the criticisms contained in this blog post apply to all facilitator-dependent techniques (e.g., Spelling to Communicate, Rapid Prompting Method, Spellers Method).

Just some of the "variants" of FC where facilitators "support" their clients by holding their wrists, elbows, shoulders, etc. and/or hold a letter board or stencil in the air. (Individual images taken from various pro-FC videos and movies).

The purpose of the study, according to the authors, was to compare the written output (a short essay) of six individuals with autism spectrum disorder (ASD) and the written output of six individuals who did not have disabilities to identify the unique linguistic characteristics of individuals with ASD. It appears the authors viewed FC-generated output to be independent and free from facilitator influence and control, although there was (and still is) no reliably controlled evidence to support that assertion. During the writing session, ASD participants were paired with assistants (facilitators) using a touch-based form of FC. For some reason, the non-disabled individuals were also paired with facilitators who touched the participants’ arms while they typed. This makes no sense, given that the non-disabled participants would not need the emotional or physical support FC proponents claim is needed to complete the task independently. As reported, the non-disabled individuals completed the written task in a matter of minutes while the FCed individuals took much longer.

The researchers’ choice of FC seems strange to me, given that, starting around 1995, major organizations publicly opposed its use and rejected it as a form of Augmentative and Alternative Communication (AAC). It seems to me that if the authors legitimately wanted to compare the written output of individuals with ASD to the written output of individuals with neurotypical development, then, perhaps they should have chosen ASD individuals who were proficient in the use of evidence-based AAC and who could, therefore, communicate without interference from a facilitator.

The authors chose FC knowing that there were concerns regarding facilitator cueing and control. They wrote that they were aware of tests revealing that facilitators were “liable to influence (deliberately or subconsciously) the pointing of an individual with ASD.” Their reference list barely included criticisms of FC, but to their credit, Mostert’s 2001 systematic review of FC made the list. In my opinion, that report alone should have given them pause.

Mostert wrote in his review:

In sum, FC proponents must be encouraged to subject their claims to further scientific verification, the claims of anecdotal evidence notwithstanding. If any small part of FC is to ever be found effective or even plausible, it is abundantly clear that only by careful use of controlled experimental methods will this be established. (312)

But, did Bernardi and Tuzzi use carefully controlled experimental methods during the FCed essay-writing sessions? No.

Bernardi and Tuzzi claimed there was a “lack of agreement on a validation method” for determining FC authorship within the scientific community. It appears they didn’t even try to implement controls that would separate facilitator behavior from those of the participants with ASD—a requirement if researches want to determine the independent communication of those being subjected to facilitator-dependent techniques. FC proponents don’t like controlled studies because these studies—using valid, scientific protocols— consistently reveal a fatal flaw in the technique (e.g., facilitator control). That, I believe, is the reason why Bernardi and Tuzzi claimed there was a “lack of agreement” with validation methods.

In addition to ignoring concerns about facilitator control, Bernardi and Tuzzi also chose a research method (what they called a “statistical analysis of textual data” or SATD) to study the written output of FC-generated messages that they knew was “non-standard,” was “less well-known,” and had yet to gain acceptance within the scientific community—their words, not mine. (p. 190). I don’t know why they complicated matters. Blinding facilitators from the content under investigation during an FC session is all it takes to reveal who is controlling letter selection. The procedure needs to be done well (to keep from inadvertently leaking information to facilitators), but the structure of these controlled tests is reasonably simple. (Read about two such “tests” that RPM founder Soma Mukhopadhyay failed here).

In 2018, ASHA renewed their opposition to FC, citing a lack of scientific evidence of the validity of FC. (Image from ASHA's website, 2024)

It is difficult to know if Bernardi and Tuzzi were aware that at least four additional systematic reviews had been conducted between Mostert’s 2001 review and the time their article was published. (See Systematic Reviews). Each study revealed a lack of scientific evidence to support the use of FC. In addition, starting around 1995, major organizations had issued statements publicly opposing FC use, citing lack of scientific evidence, facilitator cueing and control, prompt dependency, and other potential harms, but these, too, were not mentioned in their brief discussion about the validity of FC. (See Opposition Statements). Many of the opposition statements adopted in the 1990s are still in use today.

My guess is that the authors chose to overlook concerns about facilitator influence and failed to implement established reliably controlled test protocols because they were already invested in using FC with their clients in a project called the EASIEST University Project and had come to believe, on faith, that FC “worked.” The FC guidelines urge facilitators to “presume competence” in their clients and not test authorship.

I think it’s possible that the authors (consciously or not) were playing with the definition of “authorship”—just like proponents play with the definition of “independence.” For critics of FC, “independent authorship” means that individuals, with or without the use of AAC, communicate independently—that is, without interference from an assistant. For FC/S2C/RPM proponents, it seems, “independent authorship” means any written output—output that sometimes surprises the facilitator and feels “novel”—regardless of the degree of support (e.g., auditory, physical or visual cueing) provided by the facilitator.

Book cover of Uta Frith’s “Autism: Explaining the Enigma”

It certainly seems like Bernardi and Tuzzi wanted people to believe that the facilitator-dependent written essays were produced “independently” by the individuals with ASD and without interference from the facilitators.

But, as one of my mentors says, “truth will out.”

In this instance, the authors seemed baffled when the FCed individuals (the individuals with ASD) showed a richer vocabulary than the non-disabled group. Bernardi and Tuzzi wrote: “this lexical richness remains difficult to explain because language development is often limited in individuals with ASD, whose difficulties in learning new words have been described in detail. The authors referenced Uta Frith’s 1989 book Autism Explaining the enigma to back up this claim. (p. 187).

But is it difficult to explain?

Given that the authors were correct in their assessment that individuals with ASD generally have difficulty acquiring language and literacy skills, where could this unexpected richness of vocabulary possibly have come from? Influence from the facilitators, perhaps?

The authors also noted that the language used by the individuals with ASD “did not comply with the syntactic rules governing the formation of sentences.” (Example: My steady commitment diverts a train of fears). This “non-compliance,” they noted, was also inconsistent and difficult to explain.

Again, are these anomalies difficult to explain? Or is it the (perhaps inadvertent) influence of the different facilitators that caused facilitated word choices that distorted established syntactic rules during FC-generated messaging.

The authors seemed to believe these deviations somehow proved that the facilitator-dependent essays were “evidence” that individuals with ASD exhibited lexical features in their writing that were distinctive from the non-disabled group.

Arguably, a more plausible explanation is that, while the neurotypical group typed independently and with deliberation, the FCed group’s written output was subjected to the stream-of-consciousness thoughts of facilitators who may or may not have been aware of the extent to which their own behaviors influenced letter selection. The FCed messages in the study shared commonalities with individuals who, especially in the early 1900s, used automatic writing to (purportedly) get in touch with their dead loved ones via the Ouija board and/or, as with the surrealist artists and poets, to tap into their creative unconscious. (See B.F. Skinner’s 1961 article “Has Gertrude Stein a Secret?”).

It is quite possible that any differences the analysis revealed in the written essays was due to the difference between the deliberative writing (of the neurotypical group) and the stream-of-consciousness writing of the facilitators. Unedited, stream-of-consciousness, automatic writing produces a much different result than a deliberative, edit-as-you go style of writing. Because the facilitators were not blinded from test protocols (nor blinded from seeing the letter board), none of the written messages obtained in the study could be solely attributed to the individuals with ASD.

Finally, do I think this study analyzing the written output of FCed individuals proves independent communication? No. In fact, I’m not sure this study proves anything of scientific interest.

Until and unless proponents find a way to separate facilitator behavior from those of their clients or loved ones, they cannot claim that FC-generated messages are independent and free from facilitator control. The individuals with ASD in the study may or may not have had independent language and literacy skills (these were not reported in the study), but as long as a facilitator holds onto them and/or holds onto a letter board held in the air, the question of facilitator cueing and control is always going to overshadow the written output. In my opinion, Bernardi and Tuzzi’s study is both useless and irrelevant when it comes to analyzing the unique linguistic characteristics of non-speaking and minimally speaking `individuals with ASD and certainly does not prove authorship in individuals subjected to FC.

Does Bernardi and Tuzzi’s 2011 Analysis of FC Prove…Anything?

Facilitated Communication: a thoroughly discredited yet persistent technique

Does Bernardi and Tuzzi’s 2011 Analysis of FC Prove…Anything?

Announcing a second new page at FacilitatedCommunication.org: Evidence-Based Practioners Using FC.

Facilitated Communication: a thoroughly discredited yet persistent technique