Exercise 17.3: One and Two Tailed Tests

You are testing the hypothesis, using a random sample of court cases from an American state court system, that women are likely to receive shorter prison sentences for fraud convictions than are men. You calculate the mean length of sentence for men and women convicted of fraud in your sample. Now you want to know, using a t-test, if the difference between the means for men and women are statistically significant for the state court system as a whole. Would you employ a one tailed or two tailed t-test? Why?