Table 4. Lifetime and 12-month DSM-IV disorders by sexual attraction, controlling for bullying-victimization (N = 6392).

	Lifetime prevalence of DSM-IV disorders			12-month prevalence of DSM-IV disorders
Predictor: Same/both-sex attraction (1), other-sex attraction (0)	aOR [95%CI] controlling for bullying-victimization^a	Indirect effect	% of effect explained	aOR [95%CI] controlling for bullying-victimization^a	Indirect effect	% of effect explained
Any mood disorder	2.33 [1.53, 3.54]	.001	14.5%	2.99 [1.59, 5.61]	.004	10.2%
Major depression	1.94 [1.24, 3.05]	.002	15.2%	2.75 [1.43, 5.31]	.025	8.5%
Bipolar disorder	2.97 [0.61, 14.44]	.013	14.1%
Dysthymia	3.19 [1.68, 6.07]	.019	9.3%
Any anxiety disorder	2.07 [1.26, 3.40]	<.001	18.5%	2.29 [1.30, 4.05]	<.001	15.4%
Social phobia	2.55 [1.43, 4.56]	.001	14.7%	3.61 [1.49, 8.77]	.001	8.7%
Panic disorder				2.41 [1.03, 5.67]	.165
Any substance	1.95 [1.23, 3.09]	.045	6.6%
Substance abuse	1.52 [0.99, 2.33]	.264
Substance dependence				4.30 [1.03, 17.97]	.313
Any Axis-1 disorder	1.42 [0.98, 2.04]	.001	26.8%	2.37 [1.54, 3.64]	.001	12.4%
	B (SE) controlling for bullying-victimization^a			B (SE) controlling for bullying-victimization^a
Number of Axis-1 disorders	0.50 (0.14)**	<.001	15.4%	0.28 (0.11)*	.001	13.7%

Note. Controlling for gender (0 = men, 1 = women), age, and education level.

Bullying-victimization was added into the models of DSM-IV disorders for which there was a significant difference by sexual attraction. Indirect associations are only assessed when differences in DSM-IV disorders by sexual attraction were significant, see Table 3a b. Bold adjusted odds ratios and regression coefficients represent significance. Percentage mediated is calculated using the following formula: $\frac{a . b}{a b + c^{'}}$