| Objective: All kinds of sensitive issues are often encountered in today’s societysurveys. If we use direct questioning for sensitive questions survey, the respondentsmay refuse to answer or provide false answers to protect their privacy. Thus,authenticity of data was destroied and the extent of the impact was hard to estimate.Warner Model, the first randomized response technique, was proposed by Warner in1965to compensate for the defects of the traditional survey methods. Since then,several kinds of improved randomized response models were developed by statisticians.Before the study, researches of randomized response technique (RRT) fordichotomous and multichotomous sensitive questions were developed and studied withthe restriction that the sample was selected through simple random sampling. In theapplications of RRT, respondents were restricted to a small population coming fromsimple random sampling or the formulae for simple random sampling were abusedwhen the sample was selected by stratified sampling, cluster sampling or otherrelatively complicated sampling methods. The study on assessing the reliability andvalidity of the investigation on sensitive questions with RRT is seldom reported.In this regard, we select three RRT methods of Simmons model, the RRT model,and indirect RRT model and aim to explore the feasibility of the methods to investigatequalitative sensitive issues with the sample selected by stratified sampling, and toestimate the sensitive character of MSM in Beijing city. Meanwhile, the validity andreliability of the methods is assessed by the application example as well as simulativesampling by computer.Methods: The theory of RRT, stratified sampling techniques, Cochran’s classicsampling theories, total probability formulae, properties of the variance were applied tothree RRT models. Formulae were deduced for the estimator of the populationproportion and its estimated variance for three RRT models used for sensitive questionssurvey in stratified sampling. During July to September,2011,621persons were drawn from a sample of MSM in Beijing which51from urban,450from suburb and120fromsuburban. And it was successfully applied two RRT models to obtain their sensitivecharacters. In addition, Monte Carlo simulation by computer was used to estimate thevalidity and reliability of the methods above.Results:1. Sampling technology theory was combined with the theory of RRT in this study,and formulae were deduced for the estimator of the population proportion and itsestimated variance for three RRT models used for sensitive questions survey instratified sampling.2. The investigation methods and the statistical formula of two RRT models wereapplied to obtain the results of seven sensitive questions of MSM (men who have sexwith men) in Beijing. The results of seven sensitive questions are as follows:(1). The proportion of anal sex with condoms of MSM in urban, suburb, suburbanand the entire Beijing were64.71%,82.22%,81.94%and80.18%, respectively.(2) The proportion of condoms usage in the whole anal sex, which was classifiedinto never used, sometimes used, every time used and no anal sex in urban, suburb,suburban and the entire Beijing were2.94%,48.69%,45.43%,2.94%;5.93%,29.63%,56.29%,8.15%;8.33%,20.83%,68.06%,2.78%and6.57%,28.19%,59.83%,5.41%,respectively.(3) The proportion of average charge of commercial sex of MSM, which wasclassified into less than200RMB,200to399RMB,400to599RMB, more than600RMB and no commercial sex in urban, suburb, suburban and the entire Beijing were15.29%,3.53%,7.45%,7.45%,66.28%;2.22%,1.33%,0.44%,7.56%,88.45%;1.67%3.33%,8.33%,11.67%,75.00%and3.44%,2.38%,4.39%,9.2%,80.59%, respectively.(4). The proportion of results of HIV test of MSM, which were classified intopositive, negative, checked but not know the results, a uncertain results and not checkedin the last year in urban, suburb, suburban and the entire Beijing was3.53%,58.43%,3.53%,11.37%,23.14%;1.78%,88.00%,0.44%,1.78%,8.00%;5.00%,78.33%,5.00%,1.67%,10.00%and3.27%,80.85%,2.62%,2.79%,10.47%, respectively.(5). The proportion of results of STD test of MSM, which were classified intopositive, negative, checked but not know the results, a uncertain results and not checkedin the last year in urban, suburb, suburban and the entire Beijing was23.14%,54.51%,3.53%,3.53%,15.29%;12.00%,68.89%,4.89%,4.44%,9.78%;21.67%,60.00%, 1.67%,5.00%,11.66%and17.12%,63.72%,3.44%,4.57%,11.55%, respectively.(6). The proportion of way(up to ejaculation orgasm) of sexual behavior, whichwas anal sex, oral sex and other sexual behavior for MSM in urban, suburb, suburbanand the entire Beijing was67.65%,23.53%,8.82%;68.89%,20.55%,10.56%;68.75%,12.5%,18.75%and68.70%,17.64%,13.66%, respectively.(7). The proportion of condom breakage, no breakage, no condom use of MSM inurban, suburb, suburban and the entire Beijing was13.72%,67.65%,18.63%;12.12%,69.45%,18.33%;12.5%,66.67%,20.83%and12.5%,68.13%,19.37%, respectively.3. Monte Carlo simulation by computer was used to estimate the validity andreliability of the methods above. The outcomes of30simulations were compared withoverall parameters. The difference between Monte-Carlo simulative survey and overallparameters is not significant in statistical test (P>0.05). Hypothesis test betweenapplication outcomes and simulated outcomes showed no statistically significantdifference (P>0.05). So the survey methods and corresponding formulas have bettervalidity and reliability.Conclusion: The results of sampling survey using Monte Carlo simulation bycomputer suggested that the methods and corresponding formulas on sensitive questionssurvey with the sample selected by stratified sampling were effective and feasible. Thesurvey on MSM in Beijing showed that MSM were high-risk population for STD/AIDSin Beijing, token anal intercourse as their primary sexual behavior, with low proportionof condom use in the whole process of intercourse, high proportion of positive HIV orSTDs results, high proportion of condom breakage. This should be paid more attentionto by health sector. Health education and prevention and control efforts are required todevelop. |
PDF Full Text Request