Methodology · Brian Nuckols / Hesitation Media · 2026-04-29 · v1.0

Methodology

1. Search architecture

The review combined four Wave-1 reconnaissance agents and three Wave-2/3/4 hybrid agents working from a shared seed list. The Wave-1 agents conducted citation-graph searches; the Wave-2/3/4 agents conducted targeted analytical work on the resulting corpus. Searches ran on Semantic Scholar (semanticscholar.org) and OpenAlex (openalex.org) between 2026-04-26 and 2026-04-29.

Query strings, with timestamps

Query	Source	Date	Agent
"double empathy" autism	Semantic Scholar	2026-04-26	Seed Finder
autistic burnout adult	Semantic Scholar	2026-04-26	Seed Finder
autism camouflaging masking adult	Semantic Scholar	2026-04-26	Seed Finder
neurodiversity paradigm	OpenAlex	2026-04-26	Seed Finder
monotropism autism attention	Semantic Scholar	2026-04-26	Seed Finder
autism affirming therapy adult	Semantic Scholar	2026-04-27	Field Scout
autistic group therapy adult	Semantic Scholar	2026-04-27	Field Scout
adult ADHD CBT meta-analysis	OpenAlex	2026-04-27	Field Scout
neurodivergent group intervention	Semantic Scholar	2026-04-27	Field Scout
"critical autism studies"	OpenAlex	2026-04-27	Archive Scout
"mad studies" neurodiversity	Semantic Scholar	2026-04-27	Archive Scout
community-based participatory autism research	OpenAlex	2026-04-27	Archive Scout
Damian Milton author	Semantic Scholar	2026-04-28	Bibliographic Mapper
Dora Raymaker author	Semantic Scholar	2026-04-28	Bibliographic Mapper
Christina Nicolaidis author	Semantic Scholar	2026-04-28	Bibliographic Mapper
Steven Kapp author	OpenAlex	2026-04-28	Bibliographic Mapper
Robert Chapman neurodiversity author	Semantic Scholar	2026-04-28	Bibliographic Mapper
Yalom group therapeutic factors autistic	Semantic Scholar	2026-04-28	Group-Pedagogy Frontier
PEGASUS Hirvikoski autism group	Semantic Scholar	2026-04-28	Group-Pedagogy Frontier
monotropism trans-diagnostic ADHD	OpenAlex	2026-04-29	Group-Pedagogy Frontier
Linden CBT autism anxiety meta-analysis	Semantic Scholar	2026-04-29	Evidence Grader
Sandbank autism intervention umbrella	Semantic Scholar	2026-04-29	Evidence Grader
Flower Delphi affirming therapy	Semantic Scholar	2026-04-29	Evidence Grader
Higgins autistic burnout systematic review	OpenAlex	2026-04-29	Evidence Grader

Trade-press anchors (Hersey, Taylor, Price, Walker, Chapman 2023, Lorde, hooks, brown, Mingus, Okun) entered through citation tracing and curated reference lists rather than through database queries, since these works do not appear in PubMed or Semantic Scholar's primary indices. The exclusion of trade press from search-driven reviews is a known bias and is named here as a condition of the review's design.

2. Seed selection criteria

Seeds qualified for inclusion under one or more of three criteria:

Citation density. More than one hundred citations within the niche (autistic-adult, ADHD-adult, ND-affinity-group literatures) by 2025.
Foundational theoretical statement. A work whose theoretical claim is load-bearing for the field's current shape, irrespective of citation count. Murray-Lesser-Lawson 2005 qualified under this criterion despite being under-cited relative to its theoretical importance.
Community-essential. A work whose role in autistic-adult, BIPOC-disability, or liberation-pedagogy community life is load-bearing despite low formal citation count. Walker 2021, Hersey 2022, Mingus's blog essays, Lorde 1988, and the AAC-user theorist tradition qualified under this criterion.

3. Centrality scoring approach

Centrality scoring weighted citation degree against four additional dimensions: author position (academic, autistic-led, trade-press, AAC-user theorist), publication venue (affirming-paradigm journal, pathology-paradigm journal, disability-studies journal, trade press, community publication), publication year (with attention to recency in a field undergoing paradigm shift), and integration into the four prior literatures the curriculum's design draws from (autism, adult-ADHD, disability studies, liberation pedagogy). Centrality is not a single number; the review treated it as a multidimensional weighting that informed which seeds received deep-read dossier attention and which remained at citation-only level.

4. Cluster naming convention

The review names clusters as positions rather than topics. The convention reflects the field's actual structure: clusters in this literature disagree with each other rather than describing different sub-domains of a shared inquiry. Position-naming clarifies what the disagreements are about and prevents the topic-naming convention's tendency to read disagreement as gap. The six positions are documented in schools.html.

5. AE-tier definitions

Evidence claims are graded on a five-tier autistic-evidence (AE) scale designed for a field whose primary evidence is community-participatory qualitative work, theoretical statement, and consensus-derivation rather than RCT evidence. The AE scale does not collapse all evidence into RCT-grade-or-not; it grades evidence in the registers the field actually produces.

Tier	Definition
AE1	Strong evidence from RCT, well-conducted meta-analysis, or comparable quantitative gold-standard for the question. The claim is operationally defensible at the level the clinical-evidence movement requires.
AE2	Safe to claim with the field's current evidence: replicated qualitative work, multiple converging quantitative studies, or methodologically rigorous consensus-derivation (Delphi). The claim is defensible without strong hedging.
AE3	Needs hedging. The claim is supported by good-quality work but the empirical base is incomplete (single qualitative study, limited replication, small effect size, generalisability questions). Use associated with rather than causes; preliminary evidence rather than established.
AE4	Theoretical. The claim is well-developed argument rather than empirical finding. Useful for design-rationale and reading-list inclusion; not appropriate for outcome claims or treatment-efficacy claims.
AE5	Avoid. The claim exceeds the available evidence and risks misleading participants, clinicians, or the field. The methodology document records why the claim has been refused.

6. The fourteen-claim AE-tier table

Fourteen claims that the curriculum or the review might be tempted to make are graded against the AE scale below. The table is the operational discipline that pins what the review will and will not claim.

#	Claim	AE tier	Source-anchor
1	Autistic-autistic dyads show higher rapport and more efficient information transfer than autistic-allistic or allistic-autistic dyads.	AE2	Crompton et al. 2020 (diffusion-chain trilogy); replicated.
2	Identity-first language is preferred among Anglophone autistic adults surveyed.	AE2	Kenny et al. 2016; Bury et al. 2020; Taboas et al. 2023; Bottema-Beutel et al. 2021.
3	Masking is associated with mental-health burden in autistic adults.	AE3	Hull et al. 2017, 2018, 2020; Cage & Troxell-Whitman 2019. Use associated with, not causes; longitudinal evidence is limited.
4	Late-diagnosed autistic-adult phenomenology is sufficiently described to support clinical recognition.	AE3	Hull et al. 2020 (narrative review); Cage 2017; Cook et al. 2021. Cohort-honesty hedge applies (Anglophone, predominantly white, predominantly verbally fluent).
5	Monotropism describes a cognitive primitive that organises autistic experience and that runs across the broader ND umbrella.	AE3	Murray, Lesser, & Lawson 2005; Dwyer 2024 (trans-diagnostic). Empirical maturation recent; pan-ND extension younger than autism-specific application.
6	Group cohesion in autistic-only spaces exceeds cohesion in mixed-neurotype groups.	AE3	Crompton 2020 (peer transfer); qualitative literature on autistic-affinity space. Quantitative replication thin.
7	Autistic burnout is a clinically distinct phenomenon requiring formal recognition.	AE3	Raymaker et al. 2020; Higgins 2025 systematic review. Diagnostic-criteria work continues.
8	Rest, framed as clinical intervention, is licensed for autistic burnout.	AE4	Theoretically licensed (Hersey 2022; the broader liberation-pedagogy literature). RCT evidence does not exist.
9	ND-group work is contraindicated for active suicidality in the absence of crisis-protocol scaffolding.	AE3	Standard-of-care extension from group-therapy literature; autism-specific RCT evidence does not exist. The contraindication is conservative rather than empirically derived for this specific population.
10	Neurodivergence is a natural variation rather than a developmental disorder.	AE4	Theoretical. The claim is positional rather than empirical and depends on the position-set the speaker takes.
11	Self-diagnosis of autism is legitimate for community-membership and affinity-group participation.	AE4	Theoretically and ethically defended (Lewis 2016; Sarrett 2016); empirical work on self-diagnosis reliability is thin and has its own methodological complications.
12	Neuronormativity is a regulatory ideal produced by industrial capitalism.	AE4	Chapman 2021, 2023. Theoretical; the political-economic argument is rigorous within its tradition.
13	ND-affirming therapy outperforms standard CBT in head-to-head trials.	AE5	AVOID. Linden 2023 network meta-analysis contains no ND-affirming arm. The claim cannot be made from current evidence.
14	Polyvagal theory grounds the curriculum's somatic and pacing logic.	AE5	AVOID. Porges's polyvagal claims are contested in recent neuroscience (Grossman 2023; Taylor 2022). Curriculum grounds somatic and pacing logic in predictive-processing and allostatic-load literatures instead.
15	Allies-inclusive groups are safer or more effective than ND-only groups.	AE5	AVOID. No head-to-head evidence exists. Allies-inclusive design is a candidate frontier contribution rather than an evidence-based practice.

7. Counter-argument integration

The review treated five counter-characterizations as conditions the curriculum has to answer rather than as objections to be deflected. The full tensions document records the rulings; the abbreviations follow.

1. Cohort honesty

Counter: The design serves a thin slice of the ND population (verbally fluent, late-diagnosed Anglophone adults with sufficient executive resource for a closed cohort).

Review's response: Name the slice in recruitment copy, intake conversation, and methodology. Refuse claims of representativeness for the broader ND population.

2. Non-speaking and AAC-user erasure

Counter: The curriculum cannot serve non-speaking autistic adults or adults whose primary modality is AAC. Mel Baggs, Ido Kedar, Naoki Higashida, and the broader AAC-user theorist tradition are essential but absent from cohort design.

Review's response: Include AAC-user theorists in the reading list as essential context. Name the limit. Point toward separate AAC-affinity-design work the curriculum cannot itself attempt.

3. BIPOC late-diagnosis under-treatment

Counter: The literature on late-diagnosis autistic adults of color documents under-treatment, mis-diagnosis, and exclusion from white-led affirming-paradigm spaces. A single ten-session group cannot resolve this.

Review's response: Include Onaiwu, Brown, Giwa Onaiwu (All The Weight of Our Dreams) in the reading list. Design recruitment, facilitator self-audit, and access architecture to reduce reproduction of the under-treatment pattern. Acknowledge that the structural problem exceeds what the curriculum can address.

4. Hersey secularization risk

Counter: Rest Is Resistance is grounded in Black liberation theology and womanist tradition; secular pan-ND uptake risks flattening the source.

Review's response: Honor Hersey inside her lineage. Include Lorde's A Burst of Light, Brueggemann's Sabbath as Resistance, Howard Thurman's Jesus and the Disinherited, Amos Yong's theology of disability, John Swinton's pastoral-theology corpus, and Robin Wall Kimmerer's Braiding Sweetgrass as parallel and contextual sources. Make the lineage explicit in facilitator preparation.

5. Autism-centric evidence base

Counter: Roughly eighty per cent of the ND-group-work literature is autism literature; adult ADHD is thinner; the rest of the umbrella is thin in adult clinical research.

Review's response: Name the unevenness in recruitment copy and methodology. Use Dwyer 2024 to license the trans-diagnostic monotropism move. Use differentiated-practice elements within the shared-core design to meet within-cohort heterogeneity.

8. Visible Absences

What the review did not adequately cover. Naming visible absences is a mandatory part of the methodology rather than an optional appendix.

Non-Anglophone scholarship. The citation-graph review reproduces the field's English-language dominance. Francophone Quebec disability studies, the Nordic disability-studies tradition (especially Sweden), Japanese autistic self-advocacy, lusophone Brazil, hispanophone Latin America, and the post-2010 South Asian and African disability-studies networks are thinly represented. The review does not claim to have addressed this absence; it names it.

AAC-user theorists. Mel Baggs, Ido Kedar, Naoki Higashida, and the broader non-speaking-theorist tradition are present in the reading list and in the genealogy as an essential suppressed line, but they are not weighted as heavily in the body of the review as their importance to the field warrants. The under-weighting reflects the citation-graph's bias against community-essential works that did not enter the academic publication economy.

Intellectually disabled autistic adults. The literature on intellectually disabled autistic adults is thinner than the literature on non-intellectually-disabled autistic adults across every dimension: diagnostic, phenomenological, intervention-research, and community-organising. The curriculum does not serve this population, and the review does not pretend that the curriculum's design choices have been adequately tested against the needs of intellectually disabled neurodivergent adults.

Behaviour-analytic affirming voices. Hanley's skill-based treatment work and Rajaraman's compassionate-care framework belong to a reformed-behaviourism tradition that the affirming and abolitionist literatures have largely refused to engage. The refusal is principled in places and avoidance in others. A more careful review would distinguish; this review acknowledges the gap and the principled-versus-avoidance question.

Faith-based and theological critics of secularizing Hersey. Walter Brueggemann, Howard Thurman, Amos Yong, John Swinton, and the broader theological-disability-studies tradition are named in the curriculum's reading list and in the tensions document, but they are not engaged in depth in the body of the review. The full theological reckoning with the secularization risk exceeds what the review has done.

Tourette's, OCD-as-ND, dyspraxia and dyslexia adult clinical literature. The pan-ND umbrella the curriculum's design claims to serve includes these conditions; the adult clinical literature for each is thinner than the autism-and-ADHD literature. The differentiated-practice elements of the design attempt to meet the unevenness; the empirical floor under the attempt is uneven.

The citation-graph review's own biases. Semantic Scholar and OpenAlex reproduce English-language dominance, high-impact-factor journal weighting, and citation-cartel effects that the field's own meta-research is beginning to surface. The 2024 origins correction and Botha 2022 are the most important methodological-and-historical pieces on this front. The review's method inherits the biases the field is correcting; the corrections are partial.

9. What the methodology can and cannot license

The methodology can license: a coherent map of the field as it stood on 2026-04-29, a position-organised account of the disagreements active in the field, AE-tier-graded claims that pin what the curriculum will and will not assert, and a counter-characterization-tested defence of the curriculum's design choices. The methodology cannot license: claims about populations the review's evidence does not cover, head-to-head outcome claims about ND-affirming versus other modalities, mechanism claims about why rest functions clinically, or claims about the curriculum's effectiveness in cohorts that have not yet been run. The downstream task — outcome evaluation of the first cohort — is named in frontier.html as the next empirical contribution the design intends to make.