Warner (2007, see Case Study 13.1) categorized neighborhoods as having either High, Medium, or Low Faith in police. She then studied records of noncommercial robberies in these neighborhoods,...

Warner (2007, see Case Study 13.1) categorized neighborhoods as having either High, Medium, or Low Faith in police. She then studied records of noncommercial robberies in these neighborhoods, classifying them as “Involving Gun” or “Not Involving Gun.” The data are given in Table 13.14.

(a) Using the methods of Chapter 12, test for a difference in the probability a robbery will involve a gun, among neighborhoods with different levels of Faith in police.

(b) Create a system of dummy variables that would make it convenient to compare High and Medium neighborhoods to the Low Faith-in-police neighborhoods using a logistic regression. Is there a relationship between level of Faith in police and the odds a robbery will involve a gun? If so, describe where the differences lie. Are these results consistent with those of part (a)?