Entering high school students make program choices among general program, vocational program and academic program. Their choice might be modeled using their writing score and their social economic status. For our data analysis example, we will expand the third example using the hsb demo data set. Let’s first read in the data. The data set contains variables on 200 students. The outcome variable is program type. The predictor variables are social economic status, , a three-level categorical variable and writing score, write, a continuous variable. Let’s start with getting some descriptive statistics of the variables of interest.

Question

Interpret the output

Snapshot of data

	female	ses	schtyp	prog	read	write	math	science	socst	honors	awards
12	male	high	public	1	39	33	38	47	41	not enrolled	0
19	male	middle	public	1	34	46	45	39	36	not enrolled	0
25	female	low	public	1	47	37	43	42	46	not enrolled	0
26	male	high	private	1	44	38	49	39	46	not enrolled	0
29	female	low	public	1	47	41	46	40	41	not enrolled	0

Table 1.1Case Processing Summary
	N	Marginal Percentage
prog	1	105	52.5%
2	45	22.5%
3	50	25.0%
female	female	109	54.5%
male	91	45.5%
Valid	200	100.0%
Missing	0
Total	200
Subpopulation	200a
a. The dependent variable has only one value observed in 200 (100.0%) subpopulations.

Table 1.2 Model Fitting Information
Model	Model Fitting Criteria	Likelihood Ratio Tests
-2 Log Likelihood	Chi-Square	df	Sig.
Intercept Only	408.193
Final	329.362	78.831	12	.000

Table 1.3 Goodness-of-Fit
	Chi-Square	df	Sig.
Pearson	422.540	386	.097
Deviance	329.362	386	.983

Table 1.4Pseudo R-Square
Cox and Snell	.326
Nagelkerke	.374
McFadden	.193

Table 1.5 Likelihood Ratio Tests
Effect	Model Fitting Criteria	Likelihood Ratio Tests
-2 Log Likelihood of Reduced Model	Chi-Square	df	Sig.
Intercept	329.362a	.000	0	.
read	331.819	2.457	2	.293
write	331.559	2.197	2	.333
math	343.486	14.124	2	.001
science	340.723	11.361	2	.003
socst	336.158	6.796	2	.033
female	329.951	.589	2	.745
The chi-square statistic is the difference in -2 log-likelihoods between the final model and a reduced model. The reduced model is formed by omitting an effect from the final model. The null hypothesis is that all parameters of that effect are 0.
a. This reduced model is equivalent to the final model because omitting the effect does not increase the degrees of freedom.

Table1.6Parameter Estimates
proga	B	Std. Error	Wald	df	Sig.	Exp(B)	95% Confidence Interval for Exp(B)
Lower Bound	Upper Bound
1	Intercept	-9.294	1.626	32.684	1	.000
read	.033	.032	1.076	1	.300	1.034	.971	1.101
write	.048	.033	2.086	1	.149	1.049	.983	1.119
math	.114	.037	9.755	1	.002	1.121	1.044	1.205
science	-.064	.031	4.215	1	.040	.938	.883	.997
socst	.067	.026	6.458	1	.011	1.070	1.016	1.127
[female=female]	-.344	.470	.535	1	.465	.709	.282	1.781
[female=male]	0b	.	.	0	.	.	.	.
2	Intercept	-4.572	1.603	8.133	1	.004
read	-.011	.033	.113	1	.737	.989	.928	1.054
write	.019	.035	.303	1	.582	1.019	.953	1.090
math	.016	.036	.207	1	.649	1.017	.947	1.091
science	.033	.031	1.141	1	.286	1.034	.973	1.098
socst	.035	.025	1.917	1	.166	1.036	.985	1.089
[female=female]	-.129	.500	.067	1	.796	.879	.330	2.340
[female=male]	0b	.	.	0	.	.	.	.
a. The reference category is: 3.
b. This parameter is set to zero because it is redundant.

Table.1.7 Classification
Observed	Predicted
1	2	3	Percent Correct
1	90	5	10	85.7%
2	27	3	15	6.7%
3	18	6	26	52.0%
Overall Percentage	67.5%	7.0%	25.5%	59.5%