1. These data are hospital discharge data from the state of Maryland. These data are being used to predict mortality for patients over 17 years of age with septicemia. The sample size is 105 patients. A binary logistic regression model was fit to the data with Mortality as the dependent variable, with predictor variables of age (Age), sex (Sex), number of diagnoses (NDX), and number of procedures (NPR). The dependent variable is coded as 0 = did not die and 1 = died. The independent variables are all continuous, with the exception of sex, which is dichotomous. The data analysis output for this study is presented in a publication type table below. (10 points)

Variables that may predict mortality

	B	SE (B)	Wald	Odds Ratio (OR)	p value	95% CI for OR
Lower	Upper
Constant	8.797	2.927	9.032	0.000	0.003
Age	0.079	0.034	5.328	1.082	-	1.012	1.156
Sex	0.554	0.639	.751	0.575	0.386	0.164	2.011
NDX NPR	0.207 0.071	0.104 0.226	3.922 .097	1.230 0.932	0.048 0.755	1.002 0.598	1.509 1.451
	N = 105, Nagelkerke R2 = 0.256 Hosmer-Lemeshow chi-square test = 7.118, df= 8, p value= .524

What would be a good research question using the dependent and independent variables. Is the logistic regression model a good fit? Provide statistical evidence. How would you interpret the estimated Nagelkerke R² value (please include both the independent and the dependent variables in your statement)?