Non-speaking and minimally verbal autistic children have been almost entirely absent from the qualitative research that claims to describe their lives. A scoping review published in Child Psychiatry & Human Development screened 2,092 records across eight databases and grey literature and found just 11 studies, published between 2012 and 2025, in which these children aged 5 to 18 served as direct sources of qualitative data rather than as subjects described by parents, teachers and clinicians. The review, led by Sebastian Trew with Douglas H. Russell and Wenn B. Lawson, followed the Arksey and O’Malley framework and PRISMA-ScR reporting guidelines and used thematic synthesis to identify six methodological themes across the included work.
The scale of the gap is striking. An estimated 25 to 30 percent of autistic people are minimally verbal, meaning they use little or no functional speech, a group once described in the literature as the neglected end of the spectrum. Most studies that touch this population collect data from proxy informants who speak on the children’s behalf, and the review’s authors note that research relying only on proxy accounts cannot show the children’s own experiences, views or communicative agency. The 11 studies together involved 21 children and young people, but because several studies shared participants, the number of unique children was 13, rising to 16 when three published memoirs written by minimally verbal autistic youths are counted.
Video observation dominated the field. Nine of the 11 studies used naturalistic video recording in everyday settings, filming children’s communicative acts and analysing them afterwards, though the analytic machinery behind the cameras varied considerably. DeThorne and colleagues combined categorical analysis of 146 communicative offers recorded in ethnographic field notes across five classroom observations with a situated discourse analysis of video-recorded small group activity, using a sequential multimodal transcription system that displays all participants’ verbal and nonverbal resources together. Doak developed a hybridised framework drawing on Ethnography of Communication, Conversation Analysis and Multimodal Interaction Analysis, which the author described as enabling fine-grained analysis of communication.
Other studies pushed the technique further. Gómez-Victoria and colleagues applied a relational ontology and multilevel transcription to analyse one child’s communication as an assemblage of semiotic repertoires configured within multimodal and multisensory ecologies, encompassing bodily movement, vocalisation, engagement with materials and gaze. Gombar and Pejić Bach built a longitudinal record of roughly 90 home sessions and 160 video clips over four years, reviewing footage at 0.25x to 0.5x speed for micro-level analysis of affective sequencing, and identified what they termed affective resistance to device-mediated interaction. In that study the researcher was also the child’s parent, and reflexive memos were written after each analytic cycle. Doak’s playground study stood alone as the only one set outside a classroom, in an outdoor environment where no speech, signing or symbol systems were available to the children at all.
The review identified two fundamentally different positions on how a child’s communication system relates to the research method. Nine studies took the first position: the researchers adapted their analytic frameworks to receive the child’s existing communication in everyday contexts. Children used Picture Exchange Communication System, Makaton, gesture, vocalisation, AAC device use, body posture and movement, and the researchers treated all of these modes as data of equal standing. In the Vidal and DeThorne studies, the child’s AAC device was also used in adapted semi-structured interviews with simplified, closed-ended questions, though the review classed these within the first position because the questions were shaped around the child rather than the reverse.
Only two studies took the second position, in which the child’s communication system itself generates first-person qualitative data. Knox used facilitated communication, holding monthly interviews over six months that produced narratives about students’ education, relationships and emotional lives, making it the only study in which students answered researchers’ questions through FC. Welch and colleagues analysed three published memoirs that minimally verbal youths had written independently with AAC before the study began, sustained first-person accounts analysed retrospectively with no contact between researcher and authors. The review treats FC-derived data with caution: professional bodies have concluded that messages cannot be reliably attributed to the facilitated person under controlled conditions, and a 2026 systematic review of related facilitator-dependent methods found no studies that tested authorship, although some researchers argue that the dismissal of FC rests on ableist assumptions and dated quantitative designs.
Every one of the 11 studies named a theoretical framework that guided how children took part, and the review identifies this as a main finding: children’s existing modes of expression were treated as legitimate data when researchers used frameworks positioning the child as a competent communicator. The frameworks named included multimodal communicative competence theory, the distributed model of communication, the bioecological model, movement differences theory and presumption of competence, relational ontology, insider perspective epistemology and affective theories of communication. The review’s own orientation, a presumption of communicative competence, means it cannot compare these studies against any that used deficit-based frameworks, since no such study met the eligibility criteria.
Consent and assent practices varied widely and were reported inconsistently. Doak combined proxy informed consent from parents and school staff with ongoing inference of children’s assent from non-verbal signals, deliberately avoiding a pre-set checklist of assent indicators as inappropriate for children with atypical non-verbal profiles, while cautioning that a six-week period of fieldwork is not necessarily long enough to establish reliable knowledge of how to interpret a child’s non-verbal indications of distress. Knox reported IRB-approved individual assent forms signed by students through FC, a validity-disputed route. DeThorne and colleagues reported informed consent from adult participants but described no assent procedure for the focal child, whose inclusion was covered by parental consent. Vidal and DeThorne stated that consent and assent procedures were approved by the university research board without describing them. Notably, none of the 11 studies described explaining the research to the children themselves using their own AAC systems or accessible communication supports.
Five studies treated refusal and resistance as communication rather than non-compliance. Doak interpreted a child’s foot stamp and vocalisation at the end of a play episode as expressing frustration and showing the limits of play possibility in a segregated setting. Gombar and Pejić Bach described device rejection, such as pushing the AAC device aside or shifting gaze to the caregiver, as affective resistance, a boundary-setting act rather than a failure of communication. Gómez-Victoria and colleagues read a child’s spontaneous engagement with the environment during structured tasks as part of the child’s semiotic repertoire rather than off-task behaviour. The review notes that such signals may also mean a child wants a session to end, linking directly to consent as an ongoing process, and that researchers need clear ways to pause or stop when a child shows this.
The review’s limitations are explicit. Formal quality appraisal was not conducted, consistent with scoping review guidance, so the review cannot show which methods work best. The 11 studies are not fully independent: three Doak studies share one participant group of five minimally verbal autistic children aged 6 to 8 at one special school in the English Midlands, and three Vidal and DeThorne studies concern the same primary participant, a minimally verbal AAC user aged 9 years 8 months. Only English-language sources were included, children under five were excluded, and one author performed the coding without independent duplication. Yet the small corpus is itself a finding about how rarely these children have been direct qualitative data contributors. The studies do demonstrate that direct participation is possible when researchers record and analyse children’s own communication, and the authors call for methods beyond video observation, a consent and assent framework specific to non-speaking participants, and ethics and funding processes that recognise relationship-building time as part of data quality rather than preliminary overhead.
Subject of Research: Methods used to include non-speaking and minimally verbal autistic children as direct participants in qualitative research
Article Title: How Have Non-speaking and Minimally Verbal Autistic Children Been Included as Direct Participants in Qualitative Research? A Scoping Review of Methods
Article References: Trew, S., Russell, D. H., & Lawson, W. B. (2026). How Have Non-speaking and Minimally Verbal Autistic Children Been Included as Direct Participants in Qualitative Research? A Scoping Review of Methods. Child Psychiatry & Human Development. https://doi.org/10.1007/s10578-026-02117-0
Image Credits: AI Generated
DOI: 10.1007/s10578-026-02117-0
Keywords: autism, non-speaking children, minimally verbal, qualitative research, scoping review, video observation, AAC, facilitated communication, consent and assent, presumed competence, multimodal communication, research methods
News Source: Glenn Wilkins. (November 6, 2026). Scoping Review Maps How Non-speaking Autistic Children Join Qualitative Research. Scienmag.
