Creating Accessible Survey Instruments for Use with Autistic Adults and People with Intellectual Disability: Lessons Learned and Recommendations.

UNLABELLED

Despite growing appreciation of the need for research on autism in adulthood, few survey instruments have been validated for use with autistic adults. We conducted an institutional ethnography of two related partnerships that used participatory approaches to conduct research in collaboration with autistic people and people with intellectual disability. In this article, we focus on lessons learned from adapting survey instruments for use in six separate studies. Community partners identified several common problems that made original instruments inaccessible. Examples included: (1) the use of difficult vocabulary, confusing terms, or figures of speech; (2) complex sentence structure, confusing grammar, or incomplete phrases; (3) imprecise response options; (4) variation in item response based on different contexts; (5) anxiety related to not being able to answer with full accuracy; (6) lack of items to fully capture the autism-specific aspects of a construct; and (7) ableist language or concepts. Common adaptations included: (1) adding prefaces to increase precision or explain context; (2) modifying items to simplify sentence structure; (3) substituting difficult vocabulary words, confusing terms, or figures of speech with more straightforward terms; (4) adding hotlinks that define problematic terms or offer examples or clarifications; (5) adding graphics to increase clarity of response options; and (6) adding new items related to autism-specific aspects of the construct. We caution against using instruments developed for other populations unless instruments are carefully tested with autistic adults, and we describe one possible approach to ensure that instruments are accessible to a wide range of autistic participants.

LAY SUMMARY

To understand what can improve the lives of autistic adults, researchers need to collect survey data directly from autistic adults. However, most survey instruments were made for the general population and may or may not work well for autistic adults. To use lessons learned from our experience adapting surveys-in partnership with autistic adults-to create a set of recommendations for how researchers may adapt instruments to be accessible to autistic adults. Between 2006 and 2019, the Academic Autism Spectrum Partnership in Research and Education (AASPIRE) and the Partnering with People with Developmental Disabilities to Address Violence Consortium used a participatory research approach to adapt many survey instruments for use in six separate studies. We reviewed records from these partnerships and identified important lessons. The adaptation process includes the following:  (1) Co-creating collaboration guidelines and providing community partners with necessary background about terminology and processes used in survey research; (2) Collaboratively selecting which constructs to measure; (3) Discussing each construct so that we can have a shared understanding of what it means; (4) Identifying existing instruments for each construct; (5) Selecting among available instruments (or deciding that none are acceptable and that we need to create a new measure); (6) Assessing the necessary adaptations for each instrument; (7) Collaboratively modifying prefaces, items, or response options, as needed; (8) Adding "hotlink" definitions where necessary to clarify or provide examples of terms and constructs; (9) Creating new measures, when needed, in partnership with autistic adults;(10)Considering the appropriateness of creating proxy report versions of each adapted measure; and(11)Assessing the adapted instruments' psychometric properties. Partners often said that, if taking a survey that used the original instruments, they would experience confusion, frustration, anxiety, or anger. They repeatedly stated that, faced with such measures, they would offer unreliable answers, leave items blank, or just stop participating in the study. Common concerns included the use of difficult vocabulary, confusing terms, complex sentence structure, convoluted phrasings, figures of speech, or imprecise language. Partners struggled with response options that used vague terms. They also felt anxious if their answer might not be completely accurate or if their responses could vary in different situations. Often the surveys did not completely capture the intended idea. Sometimes, instruments used offensive language or ideas. And in some cases, there just were not any instruments to measure what they thought was important. Common adaptations included: (1) adding prefaces to increase precision or explain context; (2) modifying items to simplify sentence structure; (3) substituting difficult vocabulary words, confusing terms, or figures of speech with more straightforward terms; (4) adding hotlinks that define problematic terms or offer examples or clarifications; (5) adding graphics to increase clarity of response options; and (6) adding new items related to autism-specific aspects of the construct. We hope that this article encourages researchers to collaborate with autistic adults to create better survey instruments. That way, when researchers evaluate interventions and services, they can have the right tools to see if they are effective.