The RRT Model For Inquiring Quantitative Sensitive Questions Under Cluster Sampling And Assessment Of Validity And Reliability Through Computer Simulation

Objective: Survey on sensitive question is universal and important for our healthwork and medical scientific research，especially for the prevention and cure of AcquiredImmune Deficiency Syndrome (AIDS) and Sexually Transmitted Disease (STD). It isnecessary for us to uncover the character of such sensitive questions so we can takeproperly measures to solve problems caused by sensitive questions. There is an urgentneed to develop reliable investigation methods and scientific formula to accuratelycalculate corresponding indicator. The current study design on sensitive questions islimited to simple random sampling or the improvement on randomized response device.Besides，we even can’t find the literature about assessment of validity and reliability. Atthe same time，according to the Epidemiological Fact Sheet on HIV and AIDS，China,2010. The major route of transmission of HIV is still by sex，and female sexworkers (FSMs）are the major HIV risk population. It is obviously important that weshould grasp the sensitive questions character of HIV high risk population.In order to acquire accurate data of epidemiological fact on HIV and AIDS，weneed to take properly methods to carry investigation on high risk population of AIDSsuch as FSWs and men who have sex with men (MSM)，and then we should calculatecorresponding indicators with scientific statistical formula. Our research studied theRRT model for quantitative sensitive questions combining cluster sampling method andderived relevant formulas. By using this method，we successfully investigated sensitivequestions about sexually transmitted disease (including HIV) among sex workers inXichang of Sichuan province. According to the results of practical survey，to assessvalidity and reliability of our method，we sampled in theoretical population repeatedlyby using computer simulation. So，on the one hand，we test and verify the accuracy offormulas derived by our study，on the other hand，we can provide data support for local health department to establish health policy.Method:In our study，Cluster sampling techniques，Cochran’s classic sampling theories，total probability formulae，properties of the variance were applied. We deducedformulas to calculate the estimator of population mean and corresponding varianceabout quantitative sensitive question RRT model under cluster sampling. Clustersampling methods were used in Xichang during May2011to July2011，19clusterswere selected and404female sex workers were investigated by using RRT model toobtain their sensitive characters. And then quantitative indexes were calculated by usingformulas we deduced. In addition，we used Visual Basic software to simulate the realsituation and repeatedly sampling100times in the same theoretical population. Then weobtained100samples and investigated everyone by using quantitative sensitive questionRRT model. At the same time we investigated same samples through Direct QuestionTechnique (DQT) and we took its average value as golden standard. Finally，to assessthe validity and reliability，we compared the results of quantitative sensitive questionRRT model with criterion by using correlation coefficient and Z test.Result:1. Sampling technology theory was combined with the theory of RRT in thisstudy，and formulae were deduced for the estimator of the population proportion and itsestimated variance for RRT model used for quantitative sensitive question survey incluster sampling.2. Cluster sampling methods were used in Xichang during May2011to July2011，404female sex workers were investigated by using RRT model to obtain their sensitivecharacters. And then quantitative indexes were calculated by using formulas wededuced.In this study，there are three quantitative sensitive questions were investigated andthe investigation results are as below.(1) a. Sensitive question1: selected a ball from the bag，multiply the number on theball with ten and then add the product with the age of sexual services for the first time.b. Results: The average age of sexual services for the first time for the FSWs inXichang is23.29and its95%confidence interval is (20.56，26.03).(2) a. Sensitive question2: selected a ball from the bag，multiply the number on the ball with ten and then add the product with the frequency of sexual services in the lastmonth.b. Results: The average frequency of sexual services in the last month for theFSWs in Xichang is39.74and its95%confidence interval is (33.65，45.82).(3) a. Sensitive question3: selected a ball from the bag，multiply the number on theball with ten and then add the product with the average charge of sexual services in thelast month.b. Results: The average charge of sexual services in the last month for the FSWs inXichang is215.12RMB and its95%confidence interval is （172.12，258.12) RMB.3. According to results of Xichang，we successfully simulated investigation of RRTmethod and DQT method using computer. We got the correlation coefficient betweenRRT method and DQT method is0.75584(p<0.01).Conclusion:1. The method and corresponding formulas on quantitative sensitivequestions survey with the sample selected by cluster sampling were feasible.2. Our research found that the average age of sexual services for the first time forthe FSWs in Xichang is23year old，which is almost the same as other cities of China.While the average frequency of sexual services per month for the FSWs in Xichang islarge which is almost40times per month，this phenomenon reflects that there is a largegroup of HIV high risk people who have sex behavior with FSWs. It is necessary todevelop health intervention measures to prevent the spreading of venereal disease orAIDS. The average charge of sexual services in the last month for the FSWs in Xichangis about215RMB. This is a low price warning us people in all classes，whether rich orpoor，are involved in buying sexual service. Thus，it may be a tough task to develophealth intervention measures on high risk population of HIV.3. According simulation results，the correlation coefficient r is0.75584，whichmeans the validity and reliability of our method are high. So we can draw theconclusion that our survey method and corresponding formulae are scientific，valid andreliable.