Objective To analyz the survey data in world health survey and obtain reasonable estimation of health system performance for population.
Methods Sampling weights were used as adjustments, and the Taylor series linearization technique was used to estimate the population parameter and associated variances.
Results Of the ranks of health status indicators and health system responsiveness indicators, there were significant difference between the resutls of the common statistical method and of survey data analysis method for several sampled sites.
Results derived from survey data analysis method was used to construct the comprehensive performance evaluation for sampled areas.
Conclusion The multistage sample design introduced a complexity to the analysis of sample survey, and the survey data analysis method can compute approximations of point estimate and variance estimate of an population parameter, which reduced the bias associated with the deviation from independent and identical distribution(IID)assumption.