After each census, the governmental statistical departments of various countries issue items on the development and utilization of census data, in order to maximize the role of census data in social and economic development. The governmental statistical department often only provides item's successful bidders with the sample census micro-records. If the sample census micro-records are not representative of those of the population, the wrong analysis conclusions will be drawn. Few scholars or relevant personnel of governmental statistical departments pay attention to the representativeness of the sample census micro-records, which is reflected in the direct use of the census micro-records to draw statistical analysis conclusions. To solve this problem, the IPUMS database established by the Population Center of the Uni-versity of Minnesota is applied as the data source, the stratified double sampling method is adopted to construct the linear estimator, the ratio estimator and their sampling variance estimators of the number of men and women, the representativeness of 1% sample micro-records in Chongqing, China in 2000 is studies. The research results show that the linear estimator and the ratio estimator estimating the numbers of men and women obey or asymptotically obey the normal distribution, and the sample census micro-records are representative. Moreover, all the confidence intervals include the corresponding true value, and the estimated results are close to the true value, the estimation precision is high. This work helps scholars pay attention to and study the representativeness of the sample census micro-records, and test their representativeness before using sample census micro-records for statistical analysis. Also, it provides scientific basis for census data users to decide whether or to what extent to use census data.