Supplement 2 Asking Sensitive Questions in Surveys Kafu

Supplement 2: Asking Sensitive Questions in Surveys Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data Supplement 2 -1

“Sex can wait until we are married, say youngsters” 2007 -02 -12 HK Standard n Young people say sex before marriage is not widely acceptable, although it is perceived to be so, a survey has found. n The Hong Kong Sex Culture Society, a Christian group, polled 943 students from tertiary institutions and found that 51. 3 percent felt that pre-marital sexual intercourse is acceptable, but the remainder disagreed. n Nearly 30 percent of those interviewed said they would have a problem if their life-long partners were not virgins before marriage and another 41 percent said otherwise. n Despite the divergence of attitudes to sex, 411 students said kissing is the most commonly acceptable form of intimate behavior, while 27. 4 percent preferred sexual intercourse, followed by 13. 6 percent who opted for caressing. n According to the poll, students with a religious background are sexually more conservative and have stronger faith in life-long marriages. n More than half the students, or 448, who were mostly polled through Internet questionnaires and a small portion by street interviews, are religious, with 434 of them being Christians. Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 2

Examples of sensitive questions n n n n Have Have Ka-fu Wong © 2007 you you ever used illegal drugs? ever had an abortion? had more than one sexual partner in the past 6 months? ever driven a motor vehicle while intoxicated? ever shoplifted? cheated in exam in the past six months? ever attempted suicide? ever invaded tax? ECON 1003: Analysis of Economic Data 3

Asking Sensitive Questions in Surveys n How to frame a question so that the respondents have no incentive to lie? n Only if your answer to the survey question cannot be used to infer your true answer to the intended question. Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 4

The survey question n Spin a coin twice. Show no one the results. If the first spin resulted in a head, answer the question marked H. If the first spin resulted in a tail, answer the question marked T. n H: Did you have one or more sexual partners in the past 12 months? n T: Did the second spin of the coin result in a tail? The respondent has no incentive to lie because only the respondent knows which question he / she is responding to. Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 5

The calculations n The respondent will answer question H with probability q and T with probability (1 -q). If the coin is fair, q=0. 5. n Let the known probability of a YES to question T be r. If the coin is fair, r=0. 5. n Let p be the proportion of students who have one or more sexual partners in the past 12 months. That is, those who answer YES to question H. n So, the probability of answering YES to the question is Prob(YES) =q*p + (1 -q)*r n Hence, p = [ Prob(YES) –(1 -q)*r ] / q. Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 6

Did you pay tax? n According to a survey done by the National Bureau of Statistics of China in 2002 (covering about 700 residents of Beijing, Shanghai and Guangzhou): Did you pay tax according to you income, in time? I paid the full amount of my personal tax. 24. 7% I paid only part of my personal tax 33. 4% I should have paid, but I did not. 18. 2% I was not eligible to pay tax. 23. 0% Did not answer. 0. 7% Source: http: //www. stats. gov. cn/was 40/reldetail. jsp? docid=21970 薪酬系列��之二：个税��力度不� Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 7

References: n “Randomized Response” by Donald Bentley of Pomona College provide a good summary of the technical issues, as well as a list of classic references on the topic. (http: //www. dartmouth. edu/~chance/teaching_aids/RResponse. html) Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data 8

Supplement 2: Asking Sensitive Questions in Surveys - END - Ka-fu Wong © 2007 ECON 1003: Analysis of Economic Data Supplement 2 -9