However, many autism researchers consider the high-quality, state-level data collected by the California Department of Developmental Services (DDS) to be the "most reliable long-term record of autism prevalence trends in the United States" for cases on the more severe end of the spectrum.

The ability to break down California's data by county makes it possible to conduct more granular analyses of ASD prevalence than statewide or national data permit. A study published in the Journal of Autism and Developmental Disorders does just that, "unmask[ing] patterns that are hidden in aggregated . . . data" to show that autism rates are evolving in strikingly different ways for different groups. Specifically, the county-level analysis shows that racial/ethnic minorities and lower-income families are increasingly bearing the brunt of severe autism, while conversely, ASD prevalence is flattening and even declining among whites in California's wealthier counties.

The lead author of the new paper—University of Colorado researcher Cynthia Nevison—and her coauthors have used California's DDS data several times to great effect, sometimes in tandem with other datasets. This previous work has challenged the unconvincing narrative that rising autism rates are simply due to better diagnosis and has also uncovered concerning racial/ethnic disparities.