Skip to main content
NCBI home page
Sage Choice logoLink to Sage Choice
. 2026 Aug 25;30(10):2610–2627. doi: 10.1177/13623613261475023

English-Language Autism Terminology Preferences of Autistic People and their Parents in the United States: Clustering Illuminates Diverse Views

Patrick Dwyer 1,✉, Zachary J Williams 2,3,4, Jennifer E Magnuson 2,5, Tiffany G Woynaroski 2,3,6
PMCID: PMC13624356  PMID: 42638503

Abstract

Autism terminology preferences are the subject of considerable debate and various studies, the results of which sometimes conflict. This study examined diversity in terminology preferences in a U.S. sample of 3,181 non-autistic parents and 2,098 autistic people (including 451 autistic parents). The study also examined a particularly wide range of language issues: participants rated terms for autism, possible subgroups within autism, autistic people, autism likelihood, and autistic characteristics. Non-autistic parents’, autistic parents’, and autistic non-parents’ terminology ratings differed, but generally modestly. Using clustering to explore response patterns revealed greater variability. Almost half of the participants were grouped in the Open Cluster, characterised by liking or neutrality towards most terms. Others were grouped in the Identity-First Cluster, characterised by support for identity-first language, strong opposition to terms for possible autism subgroups, and rejection of many pathologising terms. Other participants were in the Person-First Cluster, characterised by support for person-first language, yet still some scepticism towards terms for possible autism subgroups. Autistic participants were overrepresented in the Identity-First Cluster, while non-autistic parents were overrepresented in the Person-First Cluster. These results suggest that there is presently no consensus regarding autism terminology preferences in the United States.

Community Abstract

Why was this study done? It is important to respect how members of any community want to be talked about. However, previous studies about terminology preferences in the autism community sometimes provide different results. These conflicting results suggest that prior studies are not always capturing the diversity of terminology preferences in the autism community.

What did the researchers do? We recruited large groups of autistic and non-autistic people in the United States. We asked them to rate different terms for talking about autism, subgroups within autism, autistic people, characteristics of autism, and the likelihood that somebody will later be diagnosed with autism.

We then clustered the data, which means we let participants’ responses sort themselves into clusters of similar responses. This allowed us to highlight the diversity of people’s views.

What did the researchers find? We found some terms that almost everyone liked, such as “autism,” and other terms or phrases that almost everyone disliked, such as “person suffering from autism.” However, most terms were liked by some people and disliked by others.

We describe an “Open Cluster” of participants who like or have neutral feelings towards most terms.

We describe an “Identity-First Cluster” of participants who like identity-first language and dislike terms for subgroups within autism. Identity-First cluster members also disliked many terms that imply autism is a bad thing, like “risk” for autism.

Finally, we describe a “Person-First Cluster” of participants who like person-first language. Person-First Cluster members disliked some terms for subgroups within autism, but not as strongly as Identity-First Cluster members did.

What are the implications? This study shows that autism community members do not always agree about terminology. This means that it is important to be thoughtful and considerate in how we talk about autism.

Keywords: identity-first language, person-first language, terminology, autism, pathologising language, functioning labels, profound autism, Asperger’s syndrome

Introduction

Considerable controversy has emerged regarding how we should talk about autism (Bottema-Beutel et al., 2021; Dwyer et al., 2022; Singer et al., 2023; Vivanti, 2020). Some terms may reflect or exacerbate stigma (Gernsbacher, 2017). Using terminology that people find offensive, or that dismisses their identities, may harm individuals (Bottema-Beutel et al., 2021; Dwyer et al., 2022). Although excessive focus on terminology issues arguably distracts from other priorities (Dwyer et al., 2021), discourse helps construct and sustain ideologies and institutions (Bazerman, 1990; Bottema-Beutel et al., 2021; Dwyer et al., 2021). Indeed, language is intertwined with how autism itself is conceptualised (Bottema-Beutel et al., 2021; Dwyer, 2022b).

There is general agreement that language questions should be decided by the autism community (e.g., Buijsman et al., 2023; Dwyer, 2022b; Keates et al., 2024; Keating et al., 2023), but determining what the community wants is challenging. Autism terminology preferences often appear inconsistent across studies – even across English-language studies (Archibald et al., 2024; Schuck et al., 2025). Not only could countries/regions systematically differ (Keating et al., 2023), but some variation may reflect recruitment methods (Archibald et al., 2024). For example, autistic advocates recruited from social media might have different views from those recruited elsewhere. Notably, social media samples often include more late-diagnosed individuals (Rødgaard et al., 2022), who tend to be more supportive of identity-first language (IFL) (Bury, Jellett, Haschek, et al., 2023).

Indeed, diverse views are certainly apparent in the ongoing debate between advocates of person-first language (PFL) and IFL (Kapp et al., 2013). PFL (e.g., “person with autism”) emerged in reaction to the pervasive use of deficit and disorder language to describe disabled people – what has been called “pathology-first language” (Dwyer, 2022b). PFL advocates argue that disability should be distanced from people’s identities by placing the term “person” first, thereby humanising them (Blaska, 1993; Kailes, 1985). Contrastingly, advocates of IFL (e.g., “autistic person”) argue that because there is nothing wrong with being neurodivergent/disabled, and because it can be central to one’s identity, the term should go first, as with race, nationality, or other identities (Botha et al., 2022; Sinclair, 2013). IFL supporters may view PFL’s need to distance disability from the person as a reflection of discomfort with disability (Gernsbacher, 2017).

Autism-related PFL and IFL preferences have been intensely studied, but results differ (Archibald et al., 2024; Schuck et al., 2025). For example, three English-language studies published in 1 year from one country – the United States – found anywhere from 22% to 87% of autistic adults prefer IFL (Flowers et al., 2023; Keating et al., 2023; Taboas et al., 2023). The study reporting the least IFL support recruited autistic people from a non-autism database (Flowers et al., 2023), potentially including fewer engaged advocates. Support for IFL is increasing (Bonnello, 2022; Zajic & Gudknecht, 2024) and is highest among those whose autism was more important to their self-concept (Bury, Jellett, Haschek, et al., 2023), so those most engaged with the autism community might be more exposed to IFL. However, as highlighted by Vivanti (2020), participating in terminology surveys poses barriers to autistic people with higher support needs. Notably, some members of the autism community, such as the People First developmental and intellectual disability (ID) self-advocacy movement, champion PFL (Horton, 2020; People First of Canada, n.d.); excluding people with co-occurring ID may underestimate PFL support and impact the external validity of studies on preferred terms.

Furthermore, the diversity of autism shapes a crucial debate: whether terms should be used to divide autistic people. Historically, autistic people have often been described as “high functioning” or “low functioning,” a nomenclature denounced by many advocates (e.g., Gardiner, 2018; Sequenzia, n.d.). In prior studies, few community members endorse functioning labels (Keates et al., 2024; Kenny et al., 2016). Functioning labels are often perceived as harmful. “Low functioning” could be derogatory and encourage low expectations, whereas “high functioning” might dismiss challenges and support needs (Bottema-Beutel et al., 2021; Keating et al., 2023; Sequenzia, n.d.).

“Asperger syndrome” refers to a clinically obsolete subcategory which could not be reliably distinguished from other forms of autism (Lord et al., 2012). Still, the Asperger’s label may be perceived as less stigmatising and less suggestive of disability than autism (Smith & Jones, 2020). However, this may encourage “Aspie supremacism”: the idea, denounced by neurodiversity advocates, that one should value people with Asperger’s because of their abilities, implicitly endorsing ableist discrimination (Baggs, 2010; Murray, 2023). Furthermore, the term’s legitimacy is tarred by debate regarding whether its namesake, Hans Asperger, was complicit in systematic killings of disabled children (see, e.g., Czech, 2018; Falk, 2020; Tatzer & Maleczek, 2024).

Given concerns surrounding functioning labels and the diagnosis of Asperger’s syndrome, alternatives have emerged. Some speak of people “with low/high support needs” (e.g., Archibald et al., 2024; Kim & Gillespie-Lynch, 2023; Mason et al., 2024), although others consider that usage oversimplified and dismissive of contextual influences (Bottema-Beutel et al., 2021). One can also indicate whether someone has a co-occurring ID or not. Some speak of “mild” or “severe” autism (Carbone et al., 2024) or different autism levels (American Psychiatric Association, 2013), but such pathology-rooted “severity” judgements are rejected by neurodiversity movement-aligned guidelines (Bottema-Beutel et al., 2021; Dwyer et al., 2022; Monk et al., 2022). One proposal advances the term “profound autism” (Clarke et al., 2024; Lord et al., 2022), but this too has been criticised (e.g., Des Roches Rosa, 2021; Pukki et al., 2022). Currently, community members’ views regarding these terms remain unclear.

Views on the heterogeneity of autism may be influenced by how group-based advocacy movements often strategically minimise internal diversity to strengthen coalitions (Ellis, 2023). This need to emphasise common ground may be especially crucial to the neurodiversity movement (Chapman, 2020; Dwyer, Gurba, et al., 2025), as its opponents argue that autistic people’s needs are too diverse for neurodiversity advocacy to apply to all autistic people (Donvan & Zucker, 2016; Jaarsma & Welin, 2012). Neurodiversity advocates fear such a division could leave those with higher needs to face unchallenged ableism (Baggs, 2010; Murray, 2023). However, while many neurodiversity advocates acknowledge the diversity of autistic people’s disability-related challenges and support needs (e.g., Ballou, 2018; Des Roches Rosa, 2021; Thinking Person’s Guide to Autism, n.d.), “neurodiversity lite” rhetoric (Neumeier, 2018) can celebrate abilities while neglecting practical access needs and societal reforms, making it misaligned with genuine neurodiversity advocacy objectives (Dwyer, Gurba, et al., 2025). Some neurodiversity advocates even question lived realities of people with high support needs and their families (Vivanti, 2024), perhaps reflecting “neurodiversity lite” ideas. Frustration with such gaslighting may fuel concern over whether neurodivergent advocates should speak for all autistic people. Subgroup terms can strategically advance (presumptively non-autistic) parents as representatives of autistic people with the highest support needs (Lutz, 2020).

Both proponents and opponents of various terms appeal to science. While advocates of medicalising autism terminology and terms for autism subgroups assert their preferred nomenclature is scientifically accurate (e.g., Singer et al., 2023; Wachtel et al., 2024), others emphasise how real-world functioning is shaped by varied individual and contextual factors (e.g., Alvares et al., 2020; Bottema-Beutel et al., 2021; Gardiner, 2018). Critics of “profound autism” note the concept relates to the degree of co-occurring intellectual and language disability, not autism, and could duplicate already existing classifications and service systems (Kapp, 2023; Pukki et al., 2022; Woods et al., 2023).

Scientific accuracy concerns also arise around traditional pathologising terms, negatively valenced terms which categorise certain human variations as deviant, defective, or inferior (Chapman, 2023; Dwyer, 2022b). “Disorder” locates disability entirely within the person, neglecting how context shapes disability (Dwyer et al., 2022). Some research suggests the terms “autism spectrum disorder” and “disorder” are endorsed by some autistic people but less frequently than “autism” (Keating et al., 2023; Kenny et al., 2016); however, other research suggests most oppose “disorder” language (Dwyer, Hersh, et al., 2025). The term “risk” for autism, which makes subjective, stigmatising value judgements (Bottema-Beutel et al., 2021), was strongly disliked in one study, with alternatives like “higher likelihood” or “showing early signs” of autism being preferred (Bent et al., 2024). Furthermore, the term “symptoms” has often been used to pathologise autistic people’s traits or characteristics (Monk et al., 2022). Some guidelines recommend neutral alternatives like characteristics, traits, or features (Bottema-Beutel et al., 2021; Dwyer et al., 2022; Monk et al., 2022), but community views on these terms are unclear.

Thus, many complex factors could shape views on autism terminology, potentially accounting for mixed research findings. Attempting to describe a single current consensus regarding autism terminology might be premature. A comprehensive examination of diversity in preferences is likely required. Furthermore, prior literature focuses heavily on PFL/IFL debates, arguably leaving other important terminology questions – including views on pathologising language and heterogeneity and subgroups within autism – comparatively understudied. Analyses often focus on group differences, rather than exploring the full range of variability in terminology views using techniques like clustering.

The present study begins to address these gaps by investigating views regarding a wide variety of autism terms in a large sample of participants from one country, the United States. The study not only characterises the views of autistic people and non-autistic parents but also describes clusters defined by differing terminology preferences.

Methods

Participants were recruited through SPARK Research Match (RM0191Woynaroski), a large American database of autistic people and their family members (Feliciano et al., 2018). Responses were solicited from autistic adults, autistic children, and parents of autistic people. In total, 5,413 participants responded to our REDCap survey in June to September 2022.

Generally, SPARK Research Match recruits only autistic people with established diagnoses. However, because many adults realise they are autistic only after a child’s diagnosis (Lilley et al., 2023), we anticipated that some autistic parents would be self-identified (n = 210) or would report autism diagnoses not yet in SPARK’s database (n = 72). These participants were included. We eliminated participants who did not report being autistic or having autistic children (n = 58) and people who contradicted themselves within our survey about whether they were autistic or not (n = 76).

This left 5,279 participants (Table 1): 3,181 non-autistic parents of autistic people (referred to as “non-autistic parents”), 451 autistic parents of autistic people (“autistic parents”), and 1,647 autistic people who have no autistic children (called “autistic non-parents,” although some may have had non-autistic children).

Table 1.

Study Participant Demographics.

Autistic non-parent Autistic parent Non-autistic parent
Age (years) M = 29.65 (SD = 14.31) M = 44.49 (SD = 9.43) M = 46.62 (SD = 9.12)
Gender Female 562 (34.12%) 308 (68.29%) 2,587 (81.33%)
Male 897 (54.46%) 108 (23.95%) 586 (18.42%)
Non-binary/non-conforming 162 (9.84%) 31 (6.87%) 7 (0.22%)
Other 26 (1.58%) 4 (0.89%) 1 (0.03%)
Race/ethnicity Non-Hispanic a White 1,292 (79.26%) 372 (83.22%) 2,506 (79.00%)
Hispanic, Latina/e/o/x, or Spanish a White 93 (5.71%) 15 (3.36%) 215 (6.78%)
Mixed 137 (8.40%) 38 (8.50%) 132 (4.16%)
Black or African American 48 (2.94%) 7 (1.57%) 157 (4.94%)
Asian 37 (2.27%) 7 (1.57%) 93 (2.93%)
Hispanic, Latina/e/o/x, or Spanish a with other/unspecified race 13 (0.80%) 1 (0.22%) 43 (1.36%)
American Indian or Alaska Native 9 (0.55%) 4 (0.89%) 15 (0.47%)
Middle Eastern or North African 2 (0.12%) 2 (0.45%) 7 (0.22%)
Native Hawaiian or Other Pacific Islander 1 (0.22%) 3 (0.09%)
Other race/ethnicity b 1 (0.06%) 1 (0.03%)
Education (categorical) 1. Some high school or less 391 (23.74%)c 10 (2.22%) 23 (0.72%)
2. High school diploma/GED 298 (18.09%) 41 (9.09%) 240 (7.54%)
3. Trade school, vocational/technical 43 (2.61%) 21 (4.66%) 152 (4.78%)
4. Some college 289 (17.55%) 77 (17.07%) 459 (14.43%)
5. Associate degree 120 (7.29%) 56 (12.42%) 335 (10.53%)
6. Bachelor’s 244 (14.81%) 93 (20.62%) 806 (25.34%)
7. Some graduate/professional school 56 (3.40%) 39 (8.65%) 141 (4.43%)
8. Master’s 143 (8.68%) 78 (17.29%) 754 (23.70%)
9. Doctorate 57 (3.46%) 36 (7.98%) 268 (8.43%)
Education (numeric values as given above) M = 3.89 (SD = 2.48) M = 5.59 (SD = 2.14) M = 5.90 (SD = 2.06)
Open in a new tab
a

Original item asked about identification as “Hispanic, Latinx, or Spanish.” bIf open-ended descriptions of other racial identities did not match another category. cRecall that 337 (20.46% of) autistic non-parents were <18 years old.

Participants came from 4,641 families (4,041 with one respondent, 563 with two, and 37 with three or more). A subset – 338 autistic participants, including one autistic parent and 337 autistic non-parents – were under age 18 (M = 15.53, range = 13.33–17.92). Furthermore, 331 autistic adults (an autistic parent and 330 non-parents) had been recruited into SPARK through their parents.

Autism and Heterogeneity Terms

Liking and Offensiveness Ratings

Similar to some prior research (Bury, Jellett, Haschek, et al., 2023; Bury, Jellett, Spoor, et al., 2023), we collected both liking and offensiveness ratings. Participants indicated how much they liked or disliked various terms “used to describe autism and/or particular subtypes of autism”; the terms are shown in Figure 1 and Supplementary Tables A1 to A8. The rating scale ranged from “Strongly dislike” (1) to “Strongly like” (7).

Figure 1.

The violin plot compares liking and offensiveness of autism-related terms by autistic individuals versus non-autistic parents.

Open in a new tab

Violin plot depicting liking and offensiveness rating density, by group, for terms for describing “autism and/or particular subtypes of autism.”

Note. Although autism and heterogeneity terms are analysed separately, they are presented together here to reflect that participants answered the questions together. Terms are ordered by mean liking among autistic participants. Liking of 1 means “strongly dislike,” while 7 is “strongly like.” Offensiveness of 1 means “not at all offensive,” but 2 is “a little bit offensive,” so even relatively low offensiveness scores may indicate a term is problematic. 5 means “extremely offensive.”

Participants were also asked to rate the terms on a 5-point offensiveness scale from “Not at all offensive” (1) to “Extremely offensive” (5).

Preferred Terms

Participants identified which of the same terms they “prefer to use.” Participants could select as many terms as they wished, or none.

Person Terms

Liking and Offensiveness Ratings

Participants rated terms for “referring to a person who has a diagnosis of autism” (options displayed in Figure 2 and Supplementary Tables A9–A12) on the aforementioned 7-point liking scale and the 5-point offensiveness scale.

Figure 2.

Violin plot showing rating density for autism-related terms by autistic and non-autistic parents, including data for person with A S D.

Open in a new tab

Violin plot depicting liking and offensiveness rating density, by group, for terms for referring to people diagnosed with autism.

Note. Terms are ordered by mean liking among autistic participants.

Preferred Terms

Similarly, participants selected which of the terms they “prefer to use.”

Likelihood Terms

Participants provided 7-point liking ratings for terms describing “a person who has a higher than average chance of being diagnosed with autism later in life (e.g., an infant with an autistic parent or older sibling).” The terms are shown in Figure 3 (upper portion) and Supplementary Tables A13 and A14.

Figure 3.

Violin plot showing mean liking ratings for autism-related terms by group. Autistic Parent, Non-Autistic Parent, Autistic Non-Parent.

Open in a new tab

Violin plot depicting liking rating density by group. Above are terms for referring to “a person who has a higher-than-average chance of being diagnosed with autism later in life.” Below are terms for “the various behaviours and traits associated with autism.” Terms are ordered based on mean liking among autistic participants.

Characteristics Terms

On the same 7-point liking scale, participants rated terms for “the various behaviours and traits associated with autism (such as taking things very literally, having difficulty with change, “stimming,” or not making eye contact).” The terms are shown in Figure 3 (lower portion) and Supplementary Tables A15 and A16.

Comparisons of Groups and Items

Liking and offensiveness ratings for autism, heterogeneity/subgroup, and person terms, as well as liking of likelihood and characteristics terms, were analysed using Bayesian multilevel cumulative link probit models in brms (Bürkner, 2017). Fixed effects of group (autistic non-parent, autistic parent, and non-autistic parent) were estimated, along with varying effects of participant ID and family ID (“random intercepts”) accounting for nesting of participants within families, as well as whether effects of group varied across different items (“random slopes”). After models were fitted, the marginaleffects package (Arel-Bundock et al., 2024) was used to derive average/marginal estimates for effects of group at each term/item and for differences between items (Supplementary Appendix A, Tables A1–A16). 95% credible intervals (CrIs) were extracted as an indicator of effects’ reliability.

Clustering Analysis

Liking and offensiveness ratings for autism, heterogeneity/subgroup, and person terms, plus liking of likelihood and characteristics terms, were also included in a k-means clustering analysis. This organises data points into clusters around centroids, which are iteratively moved to minimise the distance between data points within each cluster. Before clustering, data were inverse-normalised (i.e., converted to ranks, then mapped onto a normal distribution) to focus on relative values.

We used clustering here to reduce and illuminate continuous variability, rather than to “recover” categorically distinct groups. Where data are continuous, multiple schemes for sorting participants into differing numbers of clusters can all have value (Dwyer, 2022a). Nevertheless, techniques for optimising within- versus between-cluster variability can help researchers select an appropriately descriptive number of clusters. Therefore, we used NbClust (Charrad et al., 2014) to calculate 27 of 30 indices 1 aimed at finding an optimal number of clusters (between two and 15 clusters).

As with group comparisons, we used Bayesian cumulative link probit models to examine effects of cluster on (raw, not inverse-normalised) ratings while accounting for nesting within families. We extracted average/marginal estimates for effects of cluster on ratings of each term (Supplementary Appendix B, Tables B1–B8).

We used Bayesian multilevel logit models to examine fixed effects of group on cluster membership. Due to modelling difficulties, we accounted for varying effects of family ID but did not include a participant ID term.

Participatory Methods

This study did not use participatory methods.

Results

Liking and Offensiveness Ratings

Autism Terms

Liking and offensiveness of autism terms are displayed in Figure 1 and Supplementary Tables A1 to A4. “Autism” was the most-liked term, with a median rating of “like”; groups’ ratings of it did not statistically differ. “Autism spectrum disorder (ASD)” was less liked, with a median rating of “somewhat like”; the autistic groups liked it slightly less and found it more offensive than did non-autistic parents. Autism spectrum “condition” and “development” were more poorly regarded, with respective median ratings of “neither like nor dislike” and “somewhat dislike.” However, there was considerable variability, reflected by wide distributions of responses in Figure 1.

Heterogeneity Terms

Liking and offensiveness of autism heterogeneity/subgroup terms are also displayed in Figure 1, as well as Supplementary Tables A5 to A8. These terms were generally not well-liked and were sometimes seen as offensive, but again with considerable variability. The most-liked terms – “Asperger’s syndrome” and “autism with high/low support needs” – had median ratings of “neither like nor dislike.” The most disliked was “profound autism,” with median ratings of “somewhat dislike” and “a little bit offensive.” Compared to non-autistic parents, autistic non-parents rated all heterogeneity/subgroup terms as being somewhat less liked and more offensive, although Figure 1 shows response distributions were overall similar across groups. Autistic parents also liked most heterogeneity/subgroup terms less than non-autistic parents, and even relative to autistic non-parents they found most of these terms to be more offensive; the exceptions were ID and support need specifiers.

Person Terms

Liking and offensiveness of terms for referring to autistic people are displayed in Figure 2 and Supplementary Tables A9 to A12. Overall, the most-liked person terms were “person on the autism spectrum” and “person with autism,” with median ratings of “somewhat like.” “Person on the autism spectrum” was the least (and very seldom considered) offensive. However, the autistic groups liked these terms less and tended to consider them more offensive than did non-autistic parents.

While the IFL term “autistic person” was less liked overall, with a median rating of “neither like nor dislike,” both autistic groups rated it somewhat more positively than non-autistic parents. Albeit by a thin margin, “autistic person” captured the top position in Figure 2, which is arranged by autistic people’s mean liking.

Although the PFL/IFL hybrid “person who is autistic” had a median rating of “somewhat like,” it was similarly well-liked to and slightly less offensive than “autistic person.” On average, non-autistic parents appeared slightly less favourable to this PFL/IFL hybrid than the autistic groups.

Other PFL-aligned and IFL-aligned terms showed broadly similar patterns, with autistic groups on average rating IFL terms slightly more highly and non-autistic parents slightly preferring PFL.

The terms “Aspie” and “autism spectrum person” had mean ratings around “somewhat dislike,” with “Aspie” being more often seen as offensive than “autism spectrum person.” The terms “autist and “autie” were even less liked. “Person suffering from autism” was most disliked, with median ratings of “strongly dislike” and “moderately offensive.”

Likelihood Terms

Liking of likelihood terms is displayed in Figure 3 (upper portion) and Supplementary Tables A13 and A14. Although views varied considerably, the term “increased likelihood of autism” was best liked, with a median rating of “somewhat like.” Autistic parents gave “likelihood,” “probability,” and “predisposition” language higher ratings than did other groups. “High risk for autism” was least liked, with a median rating of “neither like nor dislike.” The autistic groups liked “risk” language less than non-autistic parents.

Characteristics Terms

Liking of terms for autistic characteristics is displayed in Figure 3 (lower portion) and Supplementary Tables A15 and A16. The best-liked were “autism . . .” and “autistic characteristics,” with median ratings of “somewhat like.” Although responses were widely distributed, language referencing “symptoms” and “features” was generally less liked, with median ratings of “neither like nor dislike.” “Dispositions” terms were least liked, although the median rating was still “neither like nor dislike.” Ratings of “autism . . .” and “autistic . . .” forms of the terms were generally similar at the overall average level.

Preferences

Autism Terms

A decisive majority of participants preferred to use “autism” (Figure 4). A smaller majority used “autism spectrum disorder (ASD).” Only small fractions of participants endorsed other autism terms.

Figure 4.

Bar plot showing proportional preferences for autism terminology among parents and non-parents; higher prevalence in autistic non-parents.

Open in a new tab

Bar plot showing proportion of participants in each group expressing preferences to use different terms.

Heterogeneity Terms

No term for possible autism subgroups had majority usage preferences in any group (Figure 4). However, at least one-fifth of participants preferred to use Asperger’s syndrome (33.64%), support need specifiers (25.57%), functioning labels (33.55%), and severity labels (21.80%).

Person Terms

Person terms were also controversial (Figure 4). A thin majority of non-autistic parents preferred to use “person on the autism spectrum” (51.27%) and “person with autism” (54.57%), but autistic people did not (41.37% and 45.57%, respectively). Autistic people showed a majority preference for “autistic person” (53.19%), which fewer (33.61%) non-autistic parents used.

Clustering

NbClust indices provided varied recommendations; the optimal number of clusters was two (nine indices), three (eight indices), or five or more (nine indices). We felt the three-cluster solution best balanced simplicity and informational value. We labelled the three clusters “Open,” “Identity-First,” and “Person-First” based on salient features of the clusters. Two- to five-cluster solutions are presented in Supplementary Appendix C for reference.

Visually, plots depict considerably greater between-cluster differences in terminology preferences (Figures 5–7) than between-group differences (Figures 1–3).

Figure 5.

A violin plot shows ratings for autistic terms, clustered by liking and offensiveness among autistics; terms ordered by liking scores.

Open in a new tab

Violin plot depicting liking and offensiveness rating density, by cluster, for terms for describing “autism and/or particular subtypes of autism.”

Note. Terms are ordered based on liking ratings among autistic participants.

Figure 6.

Violin plot shows density of liking and offensiveness ratings for various autism-related terms. Terms ordered by autistic participants' liking.

Open in a new tab

Violin plot depicting liking and offensiveness rating density, by cluster, for terms for referring to people diagnosed with autism.

Note. Terms are ordered based on liking ratings among autistic participants.

Figure 7.

Visualizes liking for autism-related terms and behaviors across three clusters, Open, Identity-First, and Person-First.

Open in a new tab

Violin plot depicting liking rating density by cluster. Above are terms for referring to “a person who has a higher-than-average chance of being diagnosed with autism later in life.” Below are terms for “the various behaviours and traits associated with autism.” Terms are ordered based on liking ratings among autistic participants.

Open Cluster participants generally had open-minded views towards almost all terms. Open participants’ median ratings only dropped to “somewhat dislike” or below for three terms – “autist,” “autie,” and “person suffering from autism” (Figures 5–7). Interestingly, in the four-cluster solution, the Open Cluster essentially divides into two groups – people liking numerous terms, and people rating most terms neutrally (Supplementary Figures C5–C8).

The Identity-First Cluster was strongly characterised by preference for IFL over PFL (Figure 6; Supplementary Tables B5 and B6). However, the cluster had other noteworthy features. Strikingly, its participants found heterogeneity/subgroup terms less likable and more offensive than did participants clustered elsewhere, with especially strong effects for functioning and severity labels (Figure 5; Supplementary Table B3). They were also less supportive of “autism spectrum disorder,” giving it a median rating of “neither like nor dislike,” compared to other clusters’ medians of “like” (Figure 5; Supplementary Table B1). Identity-First Cluster participants more clearly rejected other pathologising terms, including “person suffering from autism” (“strongly dislike”), “high risk for autism” (“dislike”), and “autism symptoms” (“somewhat dislike”) (Figures 6 and 7; Supplementary Tables B5–B8). Thus, while participants in the Identity-First Cluster may not have been universally opposed to pathologising language, they often disliked it, and they tended to rate subgroup terms unfavourably.

In contrast, Person-First Cluster participants clearly supported PFL over IFL (Figure 6; Supplementary Tables B5 and B6) and liked “autistic” forms of characteristics terms (e.g., “autistic characteristics”) less than did Identity-First Cluster members (Figure 7; Supplementary Table B8). They rated many heterogeneity/subgroup terms unfavourably but still rated them more favourably than Identity-First Cluster members (Figure 5; Supplementary Tables B3 and B4). Person-First Cluster members did not rate some pathologising terms favourably, joining the Identity-First Cluster members in strongly disliking “person suffering from autism” (Figure 6; Supplementary Table B5), but again, they generally rated pathologising terms more favourably than Identity-First Cluster members.

Usage preference data paralleled liking and offensiveness data, with Identity-First Cluster members mostly using IFL and avoiding heterogeneity/subgroup terms, while Person-First Cluster members mostly used PFL (Supplementary Figure B1).

The Open Cluster was the largest, containing nearly half the sample, split relatively evenly across groups (Table 2). With autistic non-parents and the Open Cluster as a baseline, non-autistic parents were more likely to be part of the Person-First Cluster, b = 0.55, 95% CrI = [0.37, 0.74]. Non-autistic parents were less likely to be in the Identity-First Cluster, b = −1.24, 95% CrI = [−1.53, −0.97]. Autistic parents and non-parents did not differ in probability of cluster membership.

Table 2.

Counts and Proportions of Participants in Each Cluster From the Three-Cluster Solution as a Function of Group Membership.

Autistic non-parent Autistic parent Non-autistic parent
Cluster 1: Open 767 (46.57%) 215 (47.67%) 1,530 (48.10%)
Cluster 2: Identity-First 472 (28.66%) 141 (31.26%) 430 (13.52%)
Cluster 3: Person-First 408 (24.77%) 95 (21.06%) 1,221 (38.38%)
Open in a new tab

Discussion

Examination of terminology preferences in this large dataset suggested only modest overall differences in liking and offensiveness ratings between autistic parents, autistic non-parents, and non-autistic parents. Interestingly, autistic parents – not autistic non-parents – generally showed the largest differences from non-autistic parents, perhaps reflecting the independence and high education of autistic parents, but these effects were still modest. More strikingly, individual participants’ ratings ranged widely, and when we used clustering to illuminate some of this variability, we described three clusters, each associated with distinct views.

Before discussing study results in greater detail, one could ask what their practical implications should be. In general, it is customary for outsiders to defer to the terminology preferences of members of marginalised communities (Aspinall, 2008; Retzlaff, 2005; Sharif et al., 2022). This obviously becomes more complex if those communities are internally divided. It is common for mass opinion regarding terminology in marginalised communities to lag behind community leaders’ thinking, which does not invalidate the importance of the leaders’ activism (Martin, 1991). It is also important to consider that not all community members may be familiar with all terms and relevant arguments (Dwyer, 2022b). In addition, autism terminology issues are complicated by representational questions. For example, some suggest parents of autistic people should be involved in deciding on autism terminology (e.g., Buijsman et al., 2023), whereas others emphasise the preeminence of autistic people’s own views (e.g., Dwyer, 2022b; Schuck et al., 2025). In particular, debate continues regarding who should represent autistic people whose disabilities may prevent them from representing themselves (McCoy et al., 2020). Given this, simply deferring to majority opinion is not straightforward. However, ignoring community members’ identities and terminology preferences could be frustrating and disempowering. Thus, the study’s findings have important implications for guiding autism terminology use.

The largest group – just under half of study participants – was the “Open” Cluster; its participants had positive or neutral views towards almost all terms. Thus, any suggestion that most autistic people or autism community members in the United States embrace a particular ideological stance on language seems premature. Instead, many community members seemingly lack strong language preferences (Flowers et al., 2023). However, the existence of this subgroup seems broadly consistent with the history of mass opinion lagging behind community leaders’ terminology activism in other marginalised communities (e.g., Martin, 1991); this must not be used to dismiss more engaged advocates’ opinions.

The “Identity-First” Cluster contained the next-largest group of autistic participants, with considerably fewer non-autistic parents being clustered there. The cluster was strongly characterised by a preference for IFL terms. In contrast, members of the “Person-First” Cluster – which was dominated by non-autistic parents but still included nearly a quarter of the study’s autistic participants – tended to prefer PFL.

However, Identity-First and Person-First Cluster members also seemed generally more likely than Open Cluster members to have strong views on terminology issues beyond PFL/IFL. Indeed, few terms in this article attracted consensus support. A rare but important exception might be “autism” itself: almost all participants liked “autism” or viewed it neutrally, virtually nobody found it offensive, and a clear majority preferred to use it. “Autism characteristics” was also well-liked. Otherwise, most terms – even terms like “person on the autism spectrum” that can appeal to people with diverse views (Bury, Jellett, Spoor, et al., 2023) – were apparently disliked by at least some participants.

Heterogeneity Terms

Many participants, especially in the Identity-First Cluster, were critical of terms for separating autism into subgroups. It is possible that this could partly reflect ableist discomfort towards acknowledging people with generally higher support needs (Vivanti, 2024). Notably, many Identity-First and Person-First Cluster members disliked speaking of autism “with/without intellectual disability,” some even finding this offensive, which might imply discomfort with ID. However, Identity-First Cluster members could instead have disliked the PFL towards ID inherent in “with/without.”

Moreover, Identity-First Cluster members were also critical of “Asperger’s syndrome,” which is usually associated with lower support needs. While this could reflect Asperger’s Nazi connections, another interpretation could be that Identity-First Cluster members’ aversion to subgrouping terms reflects opposition to dividing the autistic community and concern about leaving those with generally elevated support needs to face unchallenged ableism (Baggs, 2010; Murray, 2023). Contrastingly, in the five-cluster solution, members of a cluster drawn from both the main text’s Person-First and Identity-First Clusters were selectively hostile to high-support-need subgroup terms but not “Asperger’s syndrome” (Supplementary Appendix C).

Notably, “profound autism” was especially strongly disliked in this sample, and numerous Person-First and Identity-First Cluster members found it offensive. However, these data were collected in 2022, soon after the term was proposed (Lord et al., 2022). As we did not measure familiarity, it is unclear how much this aversion stemmed from unfamiliarity versus informed opposition of this specific phrase.

Pathologising Language

The pathologising term “autism spectrum disorder” was disliked by relatively few participants, even in the Identity-First Cluster. This surprised us, as in the first author’s prior research, almost all autistic participants disliked “disorder” (Dwyer, Hersh, et al., 2025). This may suggest that this study’s clusters, diverse as they are, exclude a further subgroup of community members.

Other pathologising terms in this dataset were viewed more harshly, especially in the Identity-First Cluster. While opinions varied, participants invariably preferred relatively neutral terms such as “characteristics” over the pathologising “symptoms,” and – consistent with Bent et al. (2024) – terms such as “likelihood” over “risk.” “Person suffering from autism” was the most disliked term in this study, with almost no one in the sample reportedly liking this phrase and a substantial proportion of participants finding it offensive. Thus, it is reasonable to conclude that such word choices should be avoided going forward (Bottema-Beutel et al., 2021).

Limitations and Future Directions

Although this study sample is much larger than prior autism terminology studies, permitting in-depth exploration of diversity in participants’ views, the sample is not representative of autism community members in the United States, let alone globally. Non-Hispanic White participants were overrepresented, while less-educated participants – especially when one considers some participants were children – were underrepresented. Recruitment through SPARK Research Match requires formal diagnosis and excludes self-identified participants. More subtly, the Simons Foundation’s autism research priorities emphasise basic science and incorporate pathologising language (SFARI, 2025), and neurodiversity advocates can find the organisation controversial (e.g., Crosman, 2019; Des Roches Rosa, 2021). While SPARK recruitment includes broad community and social media outreach (Feliciano et al., 2018; Fombonne et al., 2022), these messages could potentially be less successful in penetrating neurodiversity-aligned parts of the autism community, which might lead to overestimation of support for PFL and pathologising language. Furthermore, our survey was not designed to be accessible for people with autism and ID, and only 92 participated. Given the key role of self-advocates with ID in championing PFL, it seems possible this could have led to underestimation of support for PFL. Moreover, the study sample is self-selected and potentially therefore unusually interested in terminology issues; it is thus possible that these data underestimate how many autism community members have the more flexible terminology preferences associated with the “Open” cluster. Thus, further research is still needed to fully map community preferences, which are likely changing over time. Future research should also measure familiarity to contextualise liking/offensiveness; such ratings may be more meaningful if raters are familiar with relevant terms and arguments for and against them. Moreover, research needs to explore factors that might shape or relate to different profiles of terminology preferences.

Conclusions and Suggestions

This large-sample exploration of autism terminology preferences suggests that almost all autism community members in the United States like some terms, particularly “autism” and “autism characteristics.” However, most terms were controversial. While a large cluster appeared to be relatively open to various options, other participants had clear views on questions like whether PFL or IFL should be used, how (not) to describe autism’s heterogeneity, and whether pathologising language should be used. Given this diversity of responses, we suggest that non-autistic readers should, where possible and not burdensome, ask about specific autistic people’s opinions.

Otherwise (e.g., for speaking to larger audiences), our data – and the desirability of avoiding potentially harmful, non-parsimonious value judgements – suggest that pathologising terms such as “risk” and “symptoms” should be avoided in favour of terms like “likelihood” and “characteristics.” Similarly, while ratings of “autism spectrum disorder” were reasonably positive, they were inferior to “autism,” and prior research suggests harsher views towards “disorder” could exist elsewhere within the autism community. However, until service systems are reformed, we recognise that clinicians may need to use pathologising language to some extent in reports to justify clients’ support needs.

Although autistic participants in this study overall slightly preferred IFL to PFL, this margin was insufficient to constitute consensus. Given this study’s shortage of respondents with intellectual disabilities, and considering that People First movement leaders support PFL, we suggest it may be advisable to use PFL for co-occurring ID. Otherwise, as both PFL and IFL attracted support in our data and are theoretically intended to be non-pathologising, one could justifiably use either, or compromise terms like “person on the autism spectrum” or “person who is autistic.” However, given autistic/neurodiversity advocates’ pro-IFL writings, it seems quite possible that support for IFL will increase over time.

Heterogeneity/subgroup terms were most controversial of all, and these data suggest none are preferred; in any case, our sample excluded many people impacted by such nomenclatures. For legitimate decisions to be made, all relevant interested parties, including impacted autistic people who might be able to self-advocate and express relevant preferences, as well as impacted autistic and non-autistic family members and caregivers, would need to be represented in discussions about whether and how to subdivide autism.

Supplemental Material

sj-docx-1-aut-10.1177_13623613261475023 – Supplemental material for English-Language Autism Terminology Preferences of Autistic People and their Parents in the United States: Clustering Illuminates Diverse Views
sj-docx-1-aut-10.1177_13623613261475023.docx (5.2MB, docx)

Supplemental material, sj-docx-1-aut-10.1177_13623613261475023 for English-Language Autism Terminology Preferences of Autistic People and their Parents in the United States: Clustering Illuminates Diverse Views by Patrick Dwyer, Zachary J. Williams, Jennifer E. Magnuson and Tiffany G. Woynaroski in Autism

Acknowledgments

The authors thank participants for their time and Evan Suzman for assistance in this project. We thank SPARK Research Match for providing database access and administrative/technical support at no cost.

1.

Indices based on quantifying how often points in the same cluster were more distant than points not clustered together – tau, G+, and gamma (see Charrad et al., 2014; Milligan & Cooper, 1985) – were omitted due to computational demands.

Ethics Considerations: All procedures were approved by the Vanderbilt Institutional Review Board.

Author Contributions: Patrick Dwyer: Formal analysis; Visualisation; Writing – original draft.

Zachary J. Williams: Conceptualisation; Formal analysis; Investigation; Writing – review & editing.

Jennifer E. Magnuson: Conceptualisation; Writing – review & editing.

Tiffany G. Woynaroski: Conceptualisation; Investigation; Supervision; Writing – review & editing.

Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Vanderbilt University provided a Family S Scholarship to support this project. Authors’ time was supported by the United States National Institute on Deafness and Other Communication Disorders (NIDCD) grants (F30DC019510 to Z.J.W.; F31DC022173 to J.E.M.; R01DC020186 to T.G.W.).

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: P.D. is an autism advisory editor.

Positionality: The authors include both autistic and non-autistic researchers from multiple disciplines, including psychology, speech-language pathology, and psychiatry. All authors support the neurodiversity movement. Authors generally support identity-first language today, although some have used person-first language in the past. One author has an autistic child, who has alternatively preferred person-first versus identity-first language. Some authors reject proposals to subdivide autism into subcategories and prefer to list co-occurring diagnoses/identities or support needs to characterise autism’s diversity.

Authors are connected to several active or former groups offering autism terminology guidance, including the Autism Intervention Research Network for Physical Health (P.D., Z.J.W.) and the journals Neurodiversity (P.D., Z.J.W.), Autism in Adulthood (P.D.), and Autism (P.D., Z.J.W.). This article’s terminology usage largely reflects the first author’s preferences, except when listing terms studied or referring to subpopulations authors are not part of (e.g., people with intellectual disability).

Supplemental Material: Supplemental material for this article is available online.

Data Availability Statement

Data from this project will be available to approved researchers via an application on SFARI Base (https://base.sfari.org; RM0191Woynaroski).

References

  1. Alvares G. A., Bebbington K., Cleary D., Evans K., Glasson E. J., Maybery M. T., Pillar S., Uljarević M., Varcin K., Wray J., Whitehouse A. J. (2020). The misnomer of ‘high functioning autism’: Intelligence is an imprecise predictor of functional abilities at diagnosis. Autism, 24(1), 221–232. 10.1177/1362361319852831 [DOI] [PubMed] [Google Scholar]
  2. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). 10.1176/appi.books.9780890425596 [DOI]
  3. Archibald D., Dwyer P., Bury S. M. (2024). Current debates on autism language preferences: An overview. In Bennett G., Goodall E. (Eds.), The Palgrave encyclopedia of disability (pp. 1–13). Springer Nature. 10.1007/978-3-031-40858-8_88-1 [DOI] [Google Scholar]
  4. Arel-Bundock V., Greifer N., Heiss A. (2024). How to interpret statistical models using marginal effects for R and Python. Journal of Statistical Software, 111(9), Article 9. 10.18637/jss.v111.i09 [DOI] [Google Scholar]
  5. Aspinall P. J. (2008). The categorization of African descent populations in Europe and the USA: Should lexicons of recommended terminology be evidence-based? Public Health, 122(1), 61–69. 10.1016/j.puhe.2007.02.018 [DOI] [PubMed] [Google Scholar]
  6. Baggs M. (2010, March 7). Aspie Supremacy can kill. Ballastexistenz. https://ballastexistenz.wordpress.com/2010/03/07/aspie-supremacy-can-kill/
  7. Ballou E. P. (2018, February 6). What the neurodiversity movement does – And doesn’t: Offer. Thinking Person’s Guide to Autism. http://www.thinkingautismguide.com/2018/02/what-neurodiversity-movement-doesand.html
  8. Bazerman C. (1990). Discourse analysis and social construction. Annual Review of Applied Linguistics, 11, 77–83. 10.1017/S0267190500001963 [DOI] [Google Scholar]
  9. Bent C. A., Aulich A., Constantine C., Fidock E., Dwyer P., Green C., Smith J., Gurba A. N., Harrington L. T., Gore K. E., Rabba A. S., Ayton L. N., Fordyce K., Green J., Jellett R., Kennedy L. J., MacDuffie K. E., Meera S. S., Watson L. R., Hudry K. (2024). Autistic and autism community perspectives on infant and family support in the first two years of life: Findings from a community consultation survey. Autism, 29(9), 2282–2296. 10.1177/13623613241262077 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Blaska J. (1993). The power of language: Speak and write using “person first”. In Nagler M. (Ed.), Perspectives on disability (pp. 25–32). Health Markets Research. [Google Scholar]
  11. Bonnello C. (2022, March 23). Results and analysis of the autistic not weird autism survey: Autistic not. Weird. https://autisticnotweird.com/autismsurvey/
  12. Botha M., Dibb B., Frost D. M. (2022). “Autism is me”: An investigation of how autistic individuals make sense of autism and stigma. Disability & Society, 37(3), 427–453. 10.1080/09687599.2020.1822782 [DOI] [Google Scholar]
  13. Bottema-Beutel K., Kapp S. K., Lester J. N., Sasson N. J., Hand B. N. (2021). Avoiding ableist language: Suggestions for autism researchers. Autism in Adulthood, 3(1), 18–29. 10.1089/aut.2020.0014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Buijsman R., Begeer S., Scheeren A. M. (2023). ‘Autistic person’ or ‘person with autism’? Person-first language preference in Dutch adults with autism and parents. Autism, 27(3), 788–795. 10.1177/13623613221117914 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Bürkner P.-C. (2017). brms: An R package for Bayesian multilevel models using Stan. Journal of Statistical Software, 80(1), Article i01. 10.18637/jss.v080.i01 [DOI] [Google Scholar]
  16. Bury S. M., Jellett R., Haschek A., Wenzel M., Hedley D., Spoor J. R. (2023). Understanding language preference: Autism knowledge, experience of stigma and autism identity. Autism, 27(6), 1588–1600. 10.1177/13623613221142383 [DOI] [PubMed] [Google Scholar]
  17. Bury S. M., Jellett R., Spoor J. R., Hedley D. (2023). “It defines who I am” or “it’s something I have”: What language do [autistic] Australian adults [on the autism spectrum] prefer? Journal of Autism and Developmental Disorders, 53, 677–687. 10.1007/s10803-020-04425-3 [DOI] [PubMed] [Google Scholar]
  18. Carbone P. S., Stipelman C. H., Villalobos M. E., Ellzey A., Stuart A., Stoddard G. J., Campbell K. (2024). A comparison of parent-reported severe autism with mild/moderate autism among US children. Journal of Developmental & Behavioral Pediatrics, 45(5), e422–e430. 10.1097/DBP.0000000000001306 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Chapman R. (2020). Defining neurodiversity for research and practice. In Rosqvist H. B., Chown N., Stenning A. (Eds.), Neurodiversity studies: A new critical paradigm (pp. 218–220). Routledge. 10.4324/9780429322297 [DOI] [Google Scholar]
  20. Chapman R. (2023). Empire of normality: Neurodiversity and capitalism (1st ed.). Pluto Press. 10.2307/jj.8501594 [DOI] [Google Scholar]
  21. Charrad M., Ghazzali N., Boiteau V., Niknafs A. (2014). NbClust: An R package for determining the relevant number of clusters in a data set. Journal of Statistical Software, 61(6), Article 6. 10.18637/jss.v061.i06 [DOI] [Google Scholar]
  22. Clarke E. B., McCauley J. B., Lutz A., Gotelli M., Sheinkopf S. J., Lord C. (2024). Understanding profound autism: Implications for stigma and supports. Frontiers in Psychiatry, 15, Article 1287096. 10.3389/fpsyt.2024.1287096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Crosman C. (2019, November 28). Good autistic advocacy organizations vs. bad autism “charities”. In the Loop about Neurodiversity. https://intheloopaboutneurodiversity.wordpress.com/2019/11/28/good-autistic-advocacy-organizations-vs-bad-autism-charities/
  24. Czech H. (2018). Hans Asperger, national socialism, and “race hygiene” in Nazi-era Vienna. Molecular Autism, 9, Article 29. 10.1186/s13229-018-0208-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Des Roches Rosa S. (2021, December 8). Creating a “profound autism” category is segregation, not progress. Thinking Person’s Guide to Autism. https://thinkingautismguide.com/2021/12/creating-profound-autism-category-is.html
  26. Donvan J., Zucker C. (2016). In a different key: The story of autism. Crown. [Google Scholar]
  27. Dwyer P. (2022. a). Heterogeneity and clustering in autism: An introduction for critical scholars. In Milton D., Ryan S. (Eds.), The Routledge international handbook of critical autism studies (pp. 42–55). Routledge. 10.4324/9781003056577-6 [DOI] [Google Scholar]
  28. Dwyer P. (2022. b). Stigma, incommensurability, or both? Pathology-first, person-first, and identity-first language and the challenges of discourse in divided autism communities. Journal of Developmental & Behavioral Pediatrics, 43(2), 111–113. 10.1097/DBP.0000000000001054 [DOI] [PubMed] [Google Scholar]
  29. Dwyer P., Acevedo S. M., Brown H. M., Grapel J., Jones S. C., Nachman B. R., Raymaker D. M., Williams Z. J. (2021). An expert roundtable discussion on experiences of autistic autism researchers. Autism in Adulthood, 3(3), 209–220. 10.1089/aut.2021.29019.rtb [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Dwyer P., Gurba A. N., Kapp S. K., Kilgallon E., Hersh L. H., Chang D. S., Rivera S. M., Gillespie-Lynch K. (2025). Community views of neurodiversity, models of disability and autism intervention: Mixed methods reveal shared goals and key tensions. Autism, 29(9), 2297–2314. 10.1177/13623613241273029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Dwyer P., Hersh L. H., Kapp S. K., Rivera S. M., Gillespie-Lynch K. (2025). Neurodiversity movement identification and perceived appropriateness of terms used to describe autism. Autism in Adulthood. Advance online publication. 10.1089/aut.2024.0297 [DOI]
  32. Dwyer P., Ryan J. G., Williams Z. J., Gassner D. L. (2022). First do no harm: Suggestions regarding respectful autism language. Pediatrics, 149(s4), Article e2020049437. 10.1542/peds.2020-049437N [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ellis J. (2023). Imagining neurodivergent futures from the belly of the identity machine: Neurodiversity, biosociality, and strategic essentialism. Autism in Adulthood, 5(3), 225–235. 10.1089/aut.2021.0075 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Falk D. (2020). Non-complicit: Revisiting Hans Asperger’s career in Nazi-era Vienna. Journal of Autism and Developmental Disorders, 50, 2573–2584. 10.1007/s10803-019-03981-7 [DOI] [PubMed] [Google Scholar]
  35. Feliciano P., Daniels A. M., Green Snyder L., Beaumont A., Camba A., Esler A., Gulsrud A. G., Mason A., Gutierrez A., Nicholson A., Paolicelli A. M., McKenzie A. P., Rachubinski A. L., Stephens A. N., Simon A. R., Stedman A., Shocklee A. D., Swanson A., Finucane B., Chung W. K. (2018). SPARK: A US cohort of 50,000 families to accelerate autism research. Neuron, 97(3), 488–493. 10.1016/j.neuron.2018.01.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Flowers J., Dawes J., McCleary D., Marzolf H. (2023). Words matter: Language preferences in a sample of autistic adults. Neurodiversity, 1, Article 1216548. 10.1177/27546330231216548 [DOI] [Google Scholar]
  37. Fombonne E., Coppola L., Mastel S., O’Roak B. J. (2022). Validation of autism diagnosis and clinical data in the SPARK cohort. Journal of Autism and Developmental Disorders, 52(8), 3383–3398. 10.1007/s10803-021-05218-y [DOI] [PubMed] [Google Scholar]
  38. Gardiner F. (2018, March 23). The problems with functioning labels. Thinking Person’s Guide to Autism. https://thinkingautismguide.com/2018/03/the-problems-with-functioning-labels.html
  39. Gernsbacher M. A. (2017). Editorial perspective: The use of person-first language in scholarly writing may accentuate stigma. Journal of Child Psychology and Psychiatry, 58(7), 859–861. 10.1111/jcpp.12706 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Horton K. (Director). (2020, December 16). People first—In the shadow of fairview (15(1)) [Broadcast]. In Oregon experience: Oregon public broadcasting. https://www.opb.org/news/series/move-to-include/people-first-in-the-shadow-of-fairview/
  41. Jaarsma P., Welin S. (2012). Autism as a natural human variation: Reflections on the claims of the neurodiversity movement. Health Care Analysis, 20(1), 20–30. 10.1007/s10728-011-0169-9 [DOI] [PubMed] [Google Scholar]
  42. Kailes J. I. (1985). Watch your language, please!. Journal of Rehabilitation, 51(1), 68–69. [Google Scholar]
  43. Kapp S. K. (2023). Profound concerns about “profound autism”: Dangers of severity scales and functioning labels for support needs. Education Sciences, 13(2), Article 106. 10.3390/educsci13020106 [DOI] [Google Scholar]
  44. Kapp S. K., Gillespie-Lynch K., Sherman L. E., Hutman T. (2013). Deficit, difference, or both? Autism and neurodiversity. Developmental Psychology, 49(1), 59–71. 10.1037/a0028353 [DOI] [PubMed] [Google Scholar]
  45. Keates N., Martin F., Waldock K. E. (2024). Autistic people’s perspectives on functioning labels and associated reasons, and community connectedness. Journal of Autism and Developmental Disorders, 55, 1318–1328. 10.1007/s10803-024-06316-3 [DOI] [PubMed] [Google Scholar]
  46. Keating C. T., Hickman L., Leung J., Monk R., Montgomery A., Heath H., Sowden S. (2023). Autism-related language preferences across the globe: A mixed methods investigation. Autism Research, 16(2), 406–428. 10.1002/aur.2864 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Kenny L., Hattersley C., Molins B., Buckley C., Povey C., Pellicano E. (2016). Which terms should be used to describe autism? Perspectives from the UK autism community. Autism, 20(4), 442–462. 10.1177/1362361315588200 [DOI] [PubMed] [Google Scholar]
  48. Kim S. Y., Gillespie-Lynch K. (2023). Do autistic people’s support needs and non-autistic people’s support for the neurodiversity movement contribute to heightened autism stigma in South Korea vs. the US? Journal of Autism and Developmental Disorders, 53, 4199–4213. 10.1007/s10803-022-05739-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Lilley R., Lawson W., Hall G., Mahony J., Clapham H., Heyworth M., Arnold S., Trollor J., Yudell M., Pellicano E. (2023). “Peas in a pod”: Oral history reflections on autistic identity in family and community by late-diagnosed adults. Journal of Autism and Developmental Disorders, 53, 1146–1161. 10.1007/s10803-022-05667-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Lord C., Charman T., Havdahl A., Carbone P., Anagnostou E., Boyd B., Carr T., de Vries P. J., Dissanayake C., Divan G., Freitag C. M., Gotelli M. M., Kasari C., Knapp M., Mundy P., Plank A., Scahill L., Servili C., Shattuck P., McCauley J. B. (2022). The Lancet Commission on the future of care and clinical research in autism. The Lancet, 399(10321), 271–334. 10.1016/S0140-6736(21)01541-5 [DOI] [PubMed] [Google Scholar]
  51. Lord C., Petkova E., Hus V., Gan W., Lu F., Martin D. M., Ousley O., Guy L., Bernier R., Gerdts J., Algermissen M., Whitaker A., Sutcliffe J. S., Warren Z., Klin A., Saulnier C., Hanson E., Hundley R., Piggot J., Risi S. (2012). A multisite study of the clinical diagnosis of different autism spectrum disorders. Archives of General Psychiatry, 69(3), 306–313. 10.1001/archgenpsychiatry.2011.148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Lutz A. S. F. (2020, June 23). When autism advocacy is “partial.”. Psychology Today. https://www.psychologytoday.com/intl/blog/inspectrum/202006/when-autism-advocacy-is-partial
  53. Martin B. L. (1991). From Negro to Black to African American: The power of names and naming. Political Science Quarterly, 106(1), 83–107. 10.2307/2152175 [DOI] [Google Scholar]
  54. Mason R. A., Lory C., Gregori E., Gerow S., Rispoli M. J., David M., Kim S. Y., Wang D. (2024). Evidence-based practice in inclusive settings for students with autism: A best-evidence synthesis. Review Journal of Autism and Developmental Disorders, 11(3), 564–580. 10.1007/s40489-022-00352-4 [DOI] [Google Scholar]
  55. McCoy M. S., Liu E. Y., Lutz A. S. F., Sisti D. (2020). Ethical advocacy across the autism spectrum: Beyond partial representation. The American Journal of Bioethics, 20(4), 13–24. 10.1080/15265161.2020.1730482 [DOI] [PubMed] [Google Scholar]
  56. Milligan G. W., Cooper M. C. (1985). An examination of procedures for determining the number of clusters in a data set. Psychometrika, 50(2), 159–179. 10.1007/BF02294245 [DOI] [Google Scholar]
  57. Monk R., Whitehouse A. J. O., Waddington H. (2022). The use of language in autism research. Trends in Neurosciences, 45(11), 791–793. 10.1016/j.tins.2022.08.009 [DOI] [PubMed] [Google Scholar]
  58. Murray F. (2023, February 28). We need to talk about Aspie supremacists. Thinking Person’s Guide to Autism. https://thinkingautismguide.com/2023/02/we-need-to-talk-about-aspie-supremacists.html
  59. Neumeier S. (2018, February 9). ‘To Siri with love’ and the problem with neurodiversity lite. Rewire News Group. https://rewirenewsgroup.com/2018/02/09/siri-love-problem-neurodiversity-lite/ [Google Scholar]
  60. People First of Canada. (n.d.). Some history. https://www.peoplefirstofcanada.ca/about-us/history/
  61. Pukki H., Bettin J., Outlaw A. G., Hennessy J., Brook K., Dekker M., Doherty M., Shaw S. C. K., Bervoets J., Rudolph S., Corneloup T., Derwent K., Lee O., Rojas Y. G., Lawson W., Gutierrez M. V., Petek K., Tsiakkirou M., Suoninen A., Yoon W. (2022). Autistic perspectives on the future of clinical autism research. Autism in Adulthood, 4(2), 93–101. 10.31234/osf.io/qw5h2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Rødgaard E.-M., Jensen K., Miskowiak K. W., Mottron L. (2022). Representativeness of autistic samples in studies recruiting through social media. Autism Research, 15(8), 1447–1456. 10.1002/aur.2777 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Retzlaff S. (2005). What’s in a name? The politics of labelling and native identity constructions. The Canadian Journal of Native Studies, 25(2), 609–626. [Google Scholar]
  64. Schuck R. K., Chetcuti L., Dwyer P., Milosavljevic K., Bury S. M., Hedley D., Begeer S., Vivanti G., Uljarevic M. (2025). Preferences for identity-first and person-first language: A systematic review of research with autistic adults/adults with autism. Journal of Autism and Developmental Disorders. Advance online publication. 10.1007/s10803-025-07174-3 [DOI] [PubMed]
  65. Sequenzia A. (n.d.). More problems with functioning labels. https://ollibean.org/problems-functioning-labels/
  66. SFARI. (2025, May 8). SFARI scientific perspectives. https://www.sfari.org/funding-opportunities/sfari-scientific-perspectives/
  67. Sharif A., McCall A. L., Bolante K. R. (2022). Should I say “disabled people” or “people with disabilities”? Language preferences of disabled people between identity- and person-first language. In Proceedings of the 24th international ACM SIGACCESS conference on computers and accessibility (pp. 1–18). ACM. 10.1145/3517428.3544813 [DOI] [Google Scholar]
  68. Sinclair J. (2013). Why I dislike “person first” language. Autonomy, the Critical Journal of Interdisciplinary Autism Studies, 1(2), 2–3. [Google Scholar]
  69. Singer A., Lutz A., Escher J., Halladay A. (2023). A full semantic toolbox is essential for autism research and practice to thrive. Autism Research, 16(3), 497–501. 10.1002/aur.2876 [DOI] [PubMed] [Google Scholar]
  70. Smith O., Jones S. C. (2020). ‘Coming out’ with autism: Identity in people with an Asperger’s diagnosis after DSM-5. Journal of Autism and Developmental Disorders, 50, 592–602. 10.1007/s10803-019-04294-5 [DOI] [PubMed] [Google Scholar]
  71. Taboas A., Doepke K., Zimmerman C. (2023). Short report: Preferences for identity-first versus person-first language in a US sample of autism stakeholders. Autism, 27(2), 565–570. 10.1177/13623613221130845 [DOI] [PubMed] [Google Scholar]
  72. Tatzer E., Maleczek W. (2024). Was Hans Asperger complicit in the Nazi child Euthanasia by participating in the Gugging Commission?: An assessment. Zeitschrift für Kinder- und Jugendpsychiatrie und Psychotherapie, 53, Article a001001. 10.1024/1422-4917/a001001 [DOI] [PubMed]
  73. Thinking Person’s Guide to Autism. (n.d.). Neurodiversity FAQ: A neurodiversity facts and myths primer. https://thinkingautismguide.com/so-youre-doing-story-about
  74. Vivanti G. (2020). Ask the editor: What is the most appropriate way to talk about individuals with a diagnosis of autism? Journal of Autism and Developmental Disorders, 50, 691–693. 10.1007/s10803-019-04280-x [DOI] [PubMed] [Google Scholar]
  75. Vivanti G. (2024). Autism early intervention: Progress, steps backward, and the reconciliation of conflicting narratives. Current Psychiatry Reports, 26, 753–760. 10.1007/s11920-024-01552-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Wachtel L. E., Escher J., Halladay A., Lutz A., Satriale G. M., Westover A., Lopez-Arvizu C. (2024). Profound autism: An imperative diagnosis. Pediatric Clinics of North America, 71(2), 301–313. 10.1016/j.pcl.2023.12.005 [DOI] [PubMed] [Google Scholar]
  77. Woods R., Williams K., Watts C. (2023). “Profound autism”: The dire consequences of diagnostic overshadowing. Autism Research, 16(9), 1656–1657. 10.1002/aur.2985 [DOI] [PubMed] [Google Scholar]
  78. Zajic M. C., Gudknecht J. (2024). Person- and identity-first language in autism research: A systematic analysis of abstracts from 11 autism journals. Autism, 28(10), 2445–2461. 10.1177/13623613241241202 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

sj-docx-1-aut-10.1177_13623613261475023 – Supplemental material for English-Language Autism Terminology Preferences of Autistic People and their Parents in the United States: Clustering Illuminates Diverse Views
sj-docx-1-aut-10.1177_13623613261475023.docx (5.2MB, docx)

Supplemental material, sj-docx-1-aut-10.1177_13623613261475023 for English-Language Autism Terminology Preferences of Autistic People and their Parents in the United States: Clustering Illuminates Diverse Views by Patrick Dwyer, Zachary J. Williams, Jennifer E. Magnuson and Tiffany G. Woynaroski in Autism

Data Availability Statement

Data from this project will be available to approved researchers via an application on SFARI Base (https://base.sfari.org; RM0191Woynaroski).

Articles from Autism are provided here courtesy of SAGE Publications

RESOURCES