A Proposed Measurement Solution to Psychometric Concerns with Existing ASD Knowledge Assessment Tools

Background:

ASD knowledge deficits contribute to current disparities in the timing and quality of ASD services throughout the United States and globally (Khan et al., 2012; Magaña, Lopez, Aguinaga, & Morton, 2013). We will present a systematic review of the literature to examine measures used to assess ASD knowledge (Harrison et al., in press). This review examined the psychometric strength of 44 unique ASD knowledge measures across 67 studies conducted in 21 countries. Of the 67 studies reviewed, only 7% were rated as using a measure with strong psychometric support compared to 45% that were rated as using a measure with no reported psychometric support. Additionally, we will describe a tendency to examine a unitary assessment of ASD knowledge rather than thoroughly assessing subdomains of ASD knowledge content overlap and subdomains of ASD knowledge assessed (e.g., etiology, symptoms) and a failure to design measures compatible for cross-cultural research.

Objectives: We will present a subsequent study that arose following this review, in which we developed the Autism Stigma and Knowledge Questionnaire (ASK-Q) with the goal of providing a measure with strong psychometric support, cross-cultural utility, and that comprehensively assesses the multiple subdomains of ASD knowledge.

Methods:

We will describe the measurement development process including the collection of validity data. The ASK-Q has a proposed 4 factor structure including three subscales assessing specific components of ASD knowledge (diagnosis, etiology, and treatment), and a fourth domain assessing the endorsement of stigma. 149 items derived from derived from a pool of pre-existing items drawn from previously published peer-reviewed research were evaluated by a group of 16 international researchers representing 11 countries (US, UK, Iran, India, Saudi Arabia, Malaysia, Tanzania, Senegal, Cape Verde, Zambia, and Burkina Faso). Researchers rated the face, construct, and cross-cultural validity of each item using a Likert scale.

Results:

Of the total 47 items selected from the item pool for inclusion in the ASK-Q, 17 items were categorized into the symptom subscale (item mean range: 3.6 - 4.0), 16 in the etiology subscale (item mean range 3.31 - 3.63), and 14 items in the treatment subscale (item mean range 3.38 - 4.0). Ratings revealed that 7 of these knowledge items would also load on to a factor assessing the endorsement of ASD stigmas (item mean range from 3.62 - 3.89). Three researchers reviewed the remaining items to ensure that each had a clear true or false answer. Additionally, item content was reviewed to minimize repetition and to ensure that each subdomain covered a range of topics. Reponses on the ASK-Q collected from a large sample of college students and the general public (n = 617) allowed for an examination of the proposed four factor structure. Using a newer form of psychometric analysis called Diagnostic Classification Modeling we confirmed the factor structure and evaluated the statistical validity of each item among a lay sample of n = 617.

Conclusions:

The resulting measure will allow for a more valid and reliable measure of multiple domains of ASD knowledge and stigma endorsement across cross-cultural settings.