Facilitated spelling: we need rigorous research, not polemics

The current facilitated spelling practices at the center of these controversies are Rapid Prompting Method (RPM) and Spelling to Communicate (S2C). Both involve a second party serving as a facilitator for an autistic person, usually by holding up a letterboard or device for the autistic person to use to spell words. A premise of these methods is that autistic people often struggle with motor planning and execution, both for speech and for other purposes. Facilitation and practice, the idea goes, help them overcome some aspects of these challenges and reveal expressive language abilities that otherwise go undetected.

Evidence exists that at least some nonspeaking autistic people using these methods move on to communicating independently — without being guided in some way, consciously or unconsciously, by a facilitator. Along with a motor-related physiological rationale that I can't dismiss, examples like these leave me open to the possibilities of such practices, along with my closely held ethos of presuming competence and not rejecting lived experience.

Yet those who cannot bear the idea that 'nonverbal' does not equal 'utterly incapable' refuse to accept that such people exist. These ultraskeptics, often co-occurring with enthusiasts of the 'profound autism' label and applied behavior analysis (ABA), assert that without exception, facilitator cues underlie any apparently coherent communication. They refer to the 'Clever Hans' phenomenon as the explanation.

Paradoxically, the ultraskeptics also claim that autistic people aren't pointing on their letterboards where facilitators say they are. It's unclear how the ultraskeptics reconcile these mutually exclusive claims — you can't argue there's unlikely accuracy because of the facilitator at the same time you argue there's inaccuracy because of the communicator.

An 'empty' review by design

I'm writing about this subject because The Transmitter has just published a piece on facilitated spelling with a headline claiming 'Still no proof for facilitated spelling methods.' The subject of the piece is a weird little 'systematic review' that its own authors self-describe as 'empty.' The Transmitter piece describes it as 'an analysis,' despite the fact that it analyzed precisely nothing — which seems to have been precisely the authors' intent.

My primary concern is that this group of authors — who the public record attests are clearly obsessed with attacking these methods at sometimes a very personal level — pretended to conduct a systematic review at all. They had to have known going in what they would conclude. The only thing systematic about this 'review' is that they systematically set the constraints to produce no studies to review. Indeed, in a pre-registered protocol they noted, 'We expect that there will be no or a low number of studies, not enabling meta-analyses across studies.'

Their main inclusion criterion required a study to use 'a quantitative experimental design involving a priori controlled manipulation of knowledge/stimuli presented to the facilitator and the individual to empirically establish who was authoring the messages.' Their exclusions ruled out any study that relied on linguistic analysis, analysis of eye or finger movements, or observations and interviews. In other words, the only studies they would have allowed were those that used their very narrow pet design.

5,857 records, zero examined in depth

Here's what their search process yielded: more than 7,000 records; with duplicates removed, a total of 5,857. They excluded 5,830 publications based on reading the abstract alone. We don't know what those 5,830 publications said, described, reported, or concluded. We just have to take their word that they weren't relevant within the constraints they imposed.

They did a closer, full-text examination of 25 publications and decided that none of those met their criteria, either. So the 'systematic review' was 'empty.' I, too, would come up empty if I conducted a search for studies that don't exist using a design that I know no one — including myself — has used for this purpose.

The Transmitter chose to cover this 'systematic review' that set aside thousands of publications on the subject, but the piece did not mention these details about it.

What the same article's sources actually said

The Transmitter headline claiming 'still no proof' is belied by comments from sources in the piece such as David Amaral, distinguished professor of psychiatry and behavioral sciences at the UC Davis MIND Institute. Amaral is framed as viewing the entire debate as 'too polarized and too dogmatic' and says that 'a subset of autistic people may benefit from facilitated communication … particularly those who go on to become independent spellers.'

Helen Tager-Flusberg, director of the Center for Autism Research Excellence at Boston University, is author on a study suggesting that a quarter of nonspeaking autistic people have receptive language capacities that outpace their expressive language ability. Is this proportion somehow not substantial enough to warrant further investigation and consideration for supports?

Amaral asked Jaswal and colleagues to write a commentary to accompany the empty review. He wrote that after attending a conference featuring adult autistic individuals using letterboards, iPads, or keyboards, his general impression was 'that there was little or no manipulation and that individuals who could not express themselves verbally had found an alternative method to do so.'

Scholarship by self-reference

Vikram Jaswal's emailed comment to The Transmitter was pointed: 'I was surprised that almost all of the references the authors used to motivate this empty review are to commentaries published in journals they edit, blog posts they wrote, policy statements they authored, or prior empty reviews they have written — one of which was not even peer-reviewed. This scholarship by self-reference to non-empirical, often non-peer-reviewed material does not strike me as a form of serious academic engagement.'

He continued: 'The only studies they considered including involved message passing tests. Other methodologies that are standard in contemporary communication science, psycholinguistics, and neuroscience (e.g., eyetracking, movement analyses, stylistic analyses) were excluded. This would be like conducting a systematic review on object permanence, limiting eligible papers to those between 2016 and 2026 that investigated whether 6-month-old infants reach for hidden objects. There are unlikely to be any studies meeting those criteria. But it would be incorrect to conclude from that empty review that 6-month-olds do not understand object permanence.'

The question that deserves an answer

Amaral ended his explanation with the right question: 'What if there is a sizable subset of autistic individuals who are cognitively capable but cannot communicate their wishes because they are unable to talk? Should not we be devoting additional resources to figuring this out?'

Given that even the ultraskeptics admit that studies they'd consider gold standard to settle this question don't even exist — and they themselves don't seem to have conducted them — well, yes, of course we should. Here's to academic rigor and presuming competence.