The Importance of Using an Optimal Cut-Off Value for the 10-Item Autism Spectrum Quotient (AQ10)

The 10-item Autism-Spectrum Quotient (AQ10) is frequently used to screen adults for high autistic traits in clinical practice and research. For the past decade, however, the National Institute for Health and Care Excellence has recommended the use of a suboptimal ≥7 cutoff value, instead of the optimal ≥6 value specified during the AQ10’s development. A comprehensive review into the use and reporting of the AQ10 cutoff suggests that this discrepancy has proliferated across the literature, with over 58% of reports citing a suboptimal (27.15%) or ambiguous (31.13%) cutoff value. After examining the use of the AQ10 cutoff in previous research, we drew on 25 published data sets (N = 13,692) to test how applying different AQ10 cutoffs can influence research. Our analyses suggest that a striking 36.80% of the participants classified as having high autistic traits using the ≥6 cutoff would be classified as having low autistic traits using the ≥7 cutoff. The ≥6 cutoff was also found to provide a better balance between the sensitivity and specificity of the AQ10 with respect to a clinical autism diagnosis. Most critically, our analyses showed that even a 1-point difference in the AQ10 cutoff—the error made in the National Institute for Health and Care Excellence guidelines—can meaningfully change both the statistical significance and the magnitude of autism-related effects. These findings demonstrate that the suboptimal use of the AQ10 cutoff can be consequential for research, and we discuss the urgent need to establish and apply appropriate autism screening cutoff values in the future.

Public Significance Statement

Many studies have used a suboptimal cutoff value for a popular screening measure of autistic traits, leading them to poorly identify which participants have high autistic traits in their samples. We demonstrate the statistical consequences of using this suboptimal cutoff value and the implications for the interpretation of results. The importance of using autism trait measures more appropriately in the future is discussed.