Internal Consistency of Self-report Anxiety Measures for Autistic Adults Without Intellectual Disabilities: A Systematic Review and Meta-analysis.
Everyday anxiety questionnaires stay reliable when verbally fluent autistic adults answer them.
01Research in Context
What this study did
The team looked at 27 studies that used self-report anxiety scales with verbally fluent autistic adults.
They checked how well the questions hang together inside each test. This is called internal consistency.
Only adults who could read and answer questions on their own were included.
What they found
All the big-name anxiety tools passed the math check. Cronbach’s alpha was 0.87 or higher for every one.
In plain words, the GAD-7, LSAS, and DASS-21 give steady, trustworthy scores with this group.
How this fits with other research
Garrison et al. (2025) widened the door. They showed that even autistic teens with mild intellectual disability can complete the same scales if their verbal skills are strong enough.
Gotham et al. (2015) sounded a warning for depression. They found only modest validity when the same adult group filled out depression forms. Laura et al. now show anxiety forms behave better than depression forms in a similar sample.
Schiltz et al. (2017) tracked autistic youth for months and saw stable anxiety scores over time. That longitudinal stability plus the new high internal consistency tells us the tools are solid across age and time.
Why it matters
You can treat the GAD-7, LSAS, and DASS-21 like you do with neurotypical adults. Give them to verbally fluent clients, score with confidence, and use the numbers to guide treatment goals. No need to hunt for autism-specific anxiety forms unless the person has limited reading or language skills.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Hand your verbal adult client the GAD-7 and score it without extra adjustments.
02At a glance
03Original abstract
PURPOSE: Anxiety prevalence is higher for autistic adults than the general population, yet there is no consensus on whether self-report questionnaires provide internally consistent measurements. This review aimed to identify and narratively compare the internal consistency of all available self-report anxiety measures for autistic adults, and meta-analyse the measures with enough studies. METHODS: Six databases were searched to identify relevant literature. After eligibility criteria were applied, 27 articles were retained. Included papers were evaluated for risk of bias using bespoke criteria. RESULTS: The Generalised Anxiety Disorder seven-item questionnaire, Liebowitz Social Anxiety Scale, and Depression Anxiety Stress Scale (21-item - anxiety subscale [DASS-21]) were meta-analysed as they were examined in sufficient numbers of studies. Pooled Cronbach's of 0.91 (CI 0.91-0.92), 0.97 (0.96-0.97), and 0.87 (0.83-0.92), respectively, were identified. Acceptable heterogeneity was found for all measures except the DASS-21 (I2 = 95.4%). Pooled internal consistency was comparable to equivalent measures used with general population samples. A complementary narrative review revealed that internal consistency across all 27 studies ranged from good to excellent. CONCLUSION: Published data suggest self-report anxiety measures have excellent internal consistency when completed by autistic adults. Future studies should explore complementary psychometric properties and include autistic samples with intellectual disabilities to replicate the findings.
Journal of autism and developmental disorders, 2026 · doi:10.1080/19315864.2021.1992549