Table 2.

Results from the multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) distinguishing between measures of association (ORs) and measures of variance and discriminatory accuracy.

	Without mental health issues		With mental health issues
	Model 1	Model 2	Model 1	Model 2
Measures of association
Age
12–17 years		Reference		Reference
18–23 years		1.78 (1.36–2.42)		1.57 (1.38–1.76)
24–30 years		2.09 (1.65–2.70)		2.66 (2.36–3.00)
Income
High income		Reference		Reference
Medium income		1.05 (0.78–1.37)		0.87 (0.77–0.98)
Low income		1.10 (0.81–1.41)		0.87 (0.77–0.98)
Immigrant background
None		Reference		Reference
Yes		0.63 (0.49–0.79)		0.55 (0.49–0.61)
Hormonal contraception
No		Reference		Reference
Yes		1.62 (1.34–2.06)		1.19 (1.08–1.31)
Measures of variance and discriminatory accuracy*
Variance	0.30 (0.18–0.50)	0.10 (0.06–0.18)	0.29 (0.18–0.49)	0.02 (0.01–0.03)
VPC	8.45%	3.02%	8.18%	0.49%
PCV		66.29%		94.48%
AUC	0.62 (0.62–0.62)	0.62 (0.62–0.62)	0.64 (0.64–0.64)	0.64 (0.64–0.64)

The analyses are stratified by the existence of previous mental issues.

Values are point estimations (with 95% credible intervals) or percentages where indicated.

*Between-strata variance.

AUC, area under the curve; PCV, proportional change of the variance; VPC, variance partition coefficient.