Results

Logistic Regression

Data from Lacourse et al. (2001). This example uses data from a study about suicide risk in women and heavy metal music.

Model Summary - suicide_risk
Model	Deviance	AIC	BIC	df	ΔΧ²	p	McFadden R²	Nagelkerke R²	Tjur R²	Cox & Snell R²
M₀	135.533	137.533	140.329	120			0.000		0.000
M₁	85.116	111.116	147.461	108	50.417	< .001	0.372	0.506	0.412	0.341

Note. M₁ includes age, marital_status, mother_negligence, father_negligence, self_estrangement, isolation, normlessness, meaninglessness, drug_use, metal, worshipping, vicarious

Coefficients
						Wald Test			95% Confidence interval
Model		Estimate	Standard Error	Odds Ratio	z	Wald Statistic	df	p	Lower bound	Upper bound
M₀	(Intercept)	-1.110	0.211	0.330	-5.271	27.782	1	< .001	-1.522	-0.697
M₁	(Intercept)	-19.012	6.209	5.538×10^-9	-3.062	9.376	1	0.002	-31.180	-6.843
	age	0.693	0.323	1.999	2.142	4.589	1	0.032	0.059	1.327
	marital_status (Separated or Divorced)	0.183	0.677	1.201	0.271	0.073	1	0.786	-1.144	1.511
	mother_negligence	-0.020	0.053	0.981	-0.368	0.136	1	0.713	-0.124	0.085
	father_negligence	0.085	0.048	1.088	1.768	3.127	1	0.077	-0.009	0.178
	self_estrangement	0.155	0.065	1.168	2.393	5.727	1	0.017	0.028	0.282
	isolation	-0.006	0.076	0.994	-0.076	0.006	1	0.939	-0.155	0.143
	normlessness	0.191	0.109	1.211	1.758	3.089	1	0.079	-0.022	0.405
	meaninglessness	-0.067	0.061	0.936	-1.092	1.191	1	0.275	-0.186	0.053
	drug_use	0.317	0.103	1.373	3.073	9.446	1	0.002	0.115	0.519
	metal	0.136	0.092	1.145	1.478	2.184	1	0.139	-0.044	0.315
	worshipping	0.159	0.129	1.172	1.227	1.506	1	0.220	-0.095	0.413
	vicarious	-0.342	0.196	0.710	-1.741	3.033	1	0.082	-0.727	0.043

Note. suicide_risk level 'Suicidal' coded as class 1.

We can conclude that listening to heavy metal did not significantly predict suicide risk in women (of course not; anyone I’ve ever met who likes metal does not conform to the stereotype). However, in case you’re interested, listening to country music apparently does (Stack & Gundlach, 1992). The factors that did significantly ( $ $ = 0.05) predict suicide risk were age, self-estrangement, and drug use. For these three predictors, the regression estimates are positive, indicating that higher scores on these variables are associated with higher suicide risk. It's important to note here that these results should be treated like any other correlation, and so should not be taken as proof of a causal relationship (the original article is observational in nature, not experimental).

The predictor with the highest odds ratio is drug use. So, this shows you that, for girls, listening to metal was not a risk factor for suicide, but drug use was. To find out what happens for boys, you’ll just have to read the article! This is scientific proof that metal isn’t bad for your health, so download some Deathspell Omega and enjoy!