To reduce the use of drugs and other harmful substances, some public schools have started to use dogs to locate undesirable substances. Many arguments have been directed against this practice, including the allegations that (1) the dogs too often point at suspects (or their lockers or cars) where there are no contraband substances and that (2) there is too much difference in the abilities of different dogs.

In this experiment, four different dogs were randomly assigned to different schools such that each dog visited each school the same number of times. The dogs pointed to cars in which they smelled a contraband substance. Permission was then obtained from the owners of these cars, and they were then searched. A

“success” was deemed to consist of a car that contained, or was admitted by the owner to have recently contained, a contraband substance.

Cars that for some reason could not be searched have been deleted from the study. The resulting data are given in Table 12.14.

(a) Give a 0.99 confidence interval for the proportion of success for the set of dogs (see Chapter 4).

(b) Test the hypothesis that the dogs all have the same proportion of success.

Transcribed Image Text:

Table 12.14 Data for Exercise 1. RESULT Dog Fail Success Total A 51 103 154 G 43 103 146 K 79 192 271 M 40 126 166 TOTAL 213 524 737