Question:
The Hospital database associated with this text and found in Wiley PLUS contains a dichotomous variable, Service, that represents two types of hospitals, general medical and psychiatric. In the database, general medical hospitals are coded as 1 and psychiatric hospitals as 2.However, to run a logistic regression model to predict service, the data were recorded so that general medical hospitals are 0 and psychiatric hospitals are 1.Shown below is Minitab output from a logistic regression attempt to predict service by payroll expenditures. Study the output. What is the model? How good is the overall fit of the model? Comment on the strength of the predictor. Thinking about hospitals, does the model make sense to you? Use the model to estimate the probability that a hospital is a psychiatric hospital if the payroll expenditures are$80,000.
Transcribed Image Text:
Value Count Variable Service 32 (Event) 168 200 Total Logistic Regression Table Odds 95% CI Predictor Constant Payroll Exp. 0.0000323 0.0000123 2.64 0.008 1.00 1.00 1.00 Log-Likelihood = -82.346 Test that all slopes are zero: G = 11.175, DF = 1, P-value = 0. O01 Coef SE Coef P Ratio Lower Upper -0.932546 0.289749 -3.22 0.001