Table 4.

Summary of ordinal logistic regression analyses for risk of suicidality (n=1676).^a

Outcome: risk of suicidalityb		Model 1					Model 2c
		Estimate	SE	P value	Crude odds ratio (95% CI)	Estimate		SE	P value	Adjusted odds ratio (95% CI)
Predictor 1: Frequency of Close Friends use (Reference: never)
	Often	0.47	0.21	.03	1.60 (1.06-2.41)	0.40		0.22	.07	1.49 (0.98-2.28)
	Sometimes	0.44	0.21	.03	1.56 (1.04-2.34)	0.42		0.21	.047	1.53 (1.01-2.31)
	Seldom	0.19	0.19	.34	1.21 (0.82-1.76)	0.23		0.20	.25	1.26 (0.85-1.85)
Predictor 2: Online help-seeking willingness (Reference: no)
	Yes	0.46	0.18	.01	1.59 (1.11-2.27)	0.40		0.19	.03	1.50 (1.04-2.15)
Predictor 3: Peer-oriented help-seeking willingness (Reference: no)
	Yes	–0.58	0.16	<.001	0.56 (0.41-0.77)	–0.61		0.17	<.001	0.55 (0.39-0.75)

^aOrdinal logistic regression analyses were used to analyze the data.

^bRisk of suicidality had 3 outcome levels: “no ideation,” “having SI,” and “having attempted suicide.” The response of “no ideation” was chosen as the reference category.

^cModel 2 adjusted for gender and school banding.