Question

Omnibus Tests of Model Coefficients
		Chi-square	df	Sig.
Step 1	Step	53.959	5	.000
	Block	53.959	5	.000
	Model	53.959	5	.000

Model Summary
Step	-2 Log likelihood	Cox & Snell R Square	Nagelkerke R Square
1	93.346a	.375	.519
a. Estimation terminated at iteration number 6 because parameter estimates changed by less than .001.

Classification Tablea
	Observed		Predicted
			EverPot		Percentage Correct
			.00	1.00
Step 1	EverPot	.00	25	14	64.1
		1.00	5	71	93.4
	Overall Percentage				83.5
a. The cut value is .500

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)
Step 1a	TimeStudy	.275	.269	1.045	1	.307	1.316
	TimeExtracurrs	-.479	.241	3.963	1	.047	.619
	ReligImport	-.382	.183	4.339	1	.037	.683
	FriendsUse	1.258	.349	13.015	1	.000	3.517
	TimeFriends	1.140	.297	14.760	1	.000	3.127
	Constant	-6.823	2.007	11.559	1	.001	.001
a. Variable(s) entered on step 1: TimeStudy, TimeExtracurrs, ReligImport, FriendsUse, TimeFriends.

11. How much of the change in the dependent variable is explained by the model as a whole? How do you know?

12. Which variables significantly predict marijuana use? How do you know?

13. Are the variables that significantly predict marijuana use are in the expected direction? How do you know?

14. Which variable best predicts marijuana use? How do you know?

15. How many more times is someone whose friends use drugs and alcohol likely to use marijuana that someone whose friends do not use drugs and alcohol?

16. How much less likely is someone who spends time in extracurriculars to use marijuana than someone who does not spend time in extracurriculars?

Answer 1

11. How much of the change in the dependent variable is explained by the model as a whole? How do you know?

Nagelkerke R Square is 0.519, By this measure, 51.9% of change in the dependent variable is explained by the model.

12. Which variables significantly predict marijuana use? How do you know?

The significant variables are those for which p-value (Sig) is less than 0.05. Thus, the significant variables are TimeExtracurrs, ReligImport, FriendsUse and TimeFriends

13. Are the variables that significantly predict marijuana use are in the expected direction? How do you know?

Obviously, marijuana use will decrease with increase in TimeExtracurrs and ReligImport. And the estimated slope coefficients of these variable is negative.

Obviously, marijuana use will increase with increase in FriendsUse and TimeFriends. And the estimated slope coefficients of these variable is positive.

Thus, the variables that significantly predict marijuana use are in the expected direction.

14. Which variable best predicts marijuana use? How do you know?

The best variable that predicts marijuana use is TimeFriends because the Wald test statistic of this variable is maximum.

15. How many more times is someone whose friends use drugs and alcohol likely to use marijuana that someone whose friends do not use drugs and alcohol?

The coefficient of FriendsUse is 1.258. The odds of to use marijuana will increase by exp(1.258) = 3.518378

16. How much less likely is someone who spends time in extracurriculars to use marijuana than someone who does not spend time in extracurriculars?

The coefficient of TimeExtracurrs is -0.479. The odds of to use marijuana will decrease by exp(-0.479) = 0.6194025