Lying can have major economic and social implications. A double blind trial was conducted to study the effect of testosterone on lying. Among 91 healthy men, ni = 46 received 50 mg of testosterone on a skin patch while the control group, n2 = 45, received a placebo look alike patch. Subsequently, subjects performed a task where their payoff depended on the self-reported outcome of a roll of a die. Subjects received the number of Euro that they reported for their roll of the die unless a 6 appeared and then the payoff was 0. Clearly, they could increase their payoff by lying without the danger of being caught.
The data on the number of times a subject reported a 5, the most lucrative outcome, are
Placebo: 28 out of Testosterone: 16
45 timesout of 46 times
(a) Conduct a x2 test of homogeneity at level a = .05.
(b) Has the testosterone increased or decreased lying? Base your answer on a large sample confidence interval for the difference of two proportions for reporting 5.