Confidence interval construction of risk difference in paired design
Risk difference is often used to compare the effectiveness of a disease before and after treatment or the implementation of a policy measure in paired design Aiming at the problem of confidence interval estimation of risk difference in paired design,this paper uses delta method,improved Wald method,inverse hyperbolic tangent transformation method,likelihood ratio test method and saddle point approximation method to construct confidence interval,and calculates the interval coverage and interval length through Monte Carlo simu-lation to compare the performance of these five methods.The results show that the interval performance of different methods is different Among them,the saddle point approximation method is significantly better than the other four methods;The performance of the improved Wald method is second only to the saddle point approximation method,and with the increase of the sample size,the performance gap between the improved Wald method and the inverse hyperbolic tangent transformation method gradually decreases;finally an example is given to verify the five interval estimation methods.
matching designsaddle point approximationMonte Carlo simulationcon-fidence interval
万方数据
维普
