A number of years ago the American Academy of Pediatrics recommended that tetracycline drugs not be used for children under the age of 8. Prior to this recommendation, a two-year study conducted in Tennessee had investigated the extent to which physicians prescribed this drug in the prior to years. In the study a random sample of 770 family practice physicians were characterized according to whether the county of their practice was urban, suburban, or rural. The numbers of doctors in each of these categories who did and did not prescribe tetracycline to at least one child under the age of 8 are given below.

Please use SPSS only for part (i) below.

(a) Identify the explanatory and the response variables in this case.

(b) For each county type, compute the percentage of physicians who prescribed tetracycline.

(c) Based on your percentages in part (b), explain the relationship, if any, between the two variables.

(d) Compute the expected counts for the six cells under the assumption that these two variables are independent.

(e) Compute the value for the test statistic associated with the Chi-square test for independence.

(f) Use the Vassar applet to compute the p-value associated with this test.

(g) Find the odds of prescribing tetracycline for physicians in each of the three county types.

(h) Find two odds ratio using urban counties as the reference. Put the odds for urban counties as the denominator in each case. Interpret your two values.

(i) You will find the raw data for this question in the data set tetracycline.sav. Use SPSS to obtain your values in parts (b), (d), (e), and (f).