Binary Logistic Regression in SPSS & 2 pg Article Critique

Logistic regression is usually applied when the dependent variable is binary (dichotomous). It is normally used to determine the relationship between the binary dependent variable and the independent variable (Orme & Combs-Orme, 2009). In our case the dependent variable is gender, while the independent variable is the Class Type. We would like to investigate whether class type can influence gender of and individual.

Research question: Based on the class type, does class Type1 or class Type2 have higher odd ratio of determine the gender of an individual?

Analysis

Model Summary
Step	-2 Log likelihood	Cox & Snell R Square	Nagelkerke R Square
1	427.751a	.002	.002
a. Estimation terminated at iteration number 3 because parameter estimates changed by less than .001.

Classification Tablea
	Observed	Predicted
		Gender	Percentage Correct
		0	1
Step 1	Gender	0	0	113	.0
		1	0	222	100.0
	Overall Percentage			66.3
a. The cut value is .500

Variables in the Equation
		B	S.E.	Wald	df	Sig.	Exp(B)	95.0% C.I.for EXP(B)
								Lower	Upper
Step 1a	ClassType			.539	2	.764
	ClassType(1)	-.222	.347	.408	1	.523	.801	.406	1.582
	ClassType(2)	-.137	.253	.294	1	.588	.872	.531	1.431
	Constant	.756	.165	20.999	1	.000	2.130
a. Variable(s) entered on step 1: ClassType.

Based on the Model Summary table, it can be deduced that 2 % of variation in the dependent variable in the model is explained by the independent variable with regard to Cox & Snell R Square. The Variables in the Equation table above indicates the contribution made by each independent variable to that particular model and whether it is statistically significant or not. The column named Wald test indicates the statistical significance of the independent variable. From the result above it can be clearly deduced that the independent variable did not add some significance to the prediction since all of them have a p-value which is more than 5% level of significant. In other words, they were insignificance to the model.

Nevertheless, it can be concluded that based on the Class Type, Class Type 2 has 0.872 times higher odds (95%CI = 0.531 to 1.431), compared to Class Type 1 with 0.801 times odds (95%CI=0.406 to 1.582) of determining the gender of the student (Logistic Regression, 2018). Since the odd ratios are less than 1, it can be related or associated with lower odd. Therefore, the measure of the relationship between outcome and exposure, based on our analysis, is very low, and therefore insignificant.