Table 1.

Between-clinics variation in incidence rate for ADHD diagnosis, high levels of ADHD symptoms, and unexplained variation in ADHD diagnosis after controlling for ADHD symptom levels

Model	(1)	(2)	(3)	(4)	(5)
	ADHD diagnosis, unconditional	ADHD symptoms ≥ 95%, unconditional	ADHD symptoms ≥ 90%, unconditional	Residuals:ADHD diagnosis, conditional on symptoms ≥ 95%	Residuals:ADHD diagnosis, conditional on symptoms ≥ 90%
Coefficient of variation (CV)
Observed CV	.46	.18	.14	.45	.44
Mean CV under ${\text{H}}_0$	.06	.17	.12	.17	.09
[Min, Max]	[.04–.08]	[.12–.24]	[.08–.16]	[.08–.54]	[.05–.18]
Test statistic: Percent deviation between observed CV and mean CV under H0	713.9	7.6	19.4	192.3	414.0
p-value	0	.23	.06	0	0
R2	–	–	–	.13	.04

The coefficient of variation (CV) shows how much variation there is relative to the mean and is calculated as the variable’s standard deviation divided by its mean value. ${\text{H}}_0$ for CV is that variation does not exceed chance variation. p value is proportion of expected values under ${\text{H}}_0$ with values equal to, or above, observed value from 10,000 trials. Models 4 and 5 are weighted by participants in MoBa. R² from fractional regression models with diagnosis as response and symptom levels as explanatory variable