Estimation of the prevalence of autism spectrum disorder in South Korea, revisited.
Big autism surveys can look precise when they are not—expect confidence intervals twice as wide as the paper claims.
01Research in Context
What this study did
Bouck et al. (2016) went back to the 2011 South Korea autism survey. They checked the math behind the headline number: 2.64 % of kids have ASD.
The team re-ran the statistics. They asked, 'How sure can we be about that 2.64 %?'
What they found
The old confidence band was too skinny. The true 95 % interval is about twice as wide.
In plain words: the real rate could be a lot lower or higher than 2.64 %. Plan for more wiggle room next time you screen.
How this fits with other research
Tureck et al. (2013) showed new toddler ADI-R rules work well. C et al. remind us that even good tools give fuzzy numbers when you scale them to a whole country.
Mulder et al. (2020) tweaked SCQ and SRS-2 cut-offs to fix false positives in fragile X. C et al. make the same point for big epidemiology: small math choices snowball into huge swings.
Kim et al. (2025) found Korean adults often dehumanize autistic people. If the true prevalence is murkier than we thought, stigma may rest on shaky ground too.
Why it matters
When you read a prevalence headline, double the margin of error in your head. If your district says '1 in 50,' think '1 in 35 to 1 in 70.' Build budgets, staffing, and parent talks around that wider band so you are not caught short. Share the uncertainty with stakeholders; it protects your credibility when later counts shift.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Add a footnote to your next grant or IEP report: 'Prevalence estimates carry ~2× wider uncertainty than stated.'
02At a glance
03Original abstract
Two-phase designs in epidemiological studies of autism prevalence introduce methodological complications that can severely limit the precision of resulting estimates. If the assumptions used to derive the prevalence estimate are invalid or if the uncertainty surrounding these assumptions is not properly accounted for in the statistical inference procedure, then the point estimate may be inaccurate and the confidence interval may not be a true reflection of the precision of the estimate. We examine these potential pitfalls in the context of a recent high-profile finding by Kim et al. (2011, Prevalence of autism spectrum disorders in a total population sample. American Journal of Psychiatry 168: 904-912), who estimated that autism spectrum disorder affects 2.64% of children in a South Korean community. We reconstructed the study's methodology and used Monte Carlo simulations to analyze whether their point estimate and 95% confidence interval (1.91%, 3.37%) were reasonable, given what was known about their screening instrument and sample. We find the original point estimate to be highly assumption-dependent, and after accounting for sources of uncertainty unaccounted for in the original article, we demonstrate that a more reasonable confidence interval would be approximately twice as large as originally reported. We argue that future studies should give serious consideration to the additional sources of uncertainty introduced by a two-phase design, which may easily outstrip any expected gains in efficiency.
Autism : the international journal of research and practice, 2016 · doi:10.1177/1362361315592378