In many surveys, some sample units do not answer some or all of the questions. This causes nonresponse. Bias and variance inflation are two important effects of nonresponse on survey statistics. Although increasing the sample size prevents variance inflation of estimates, it does not necessarily reduce statistical bias. Therefore, various methods are used to adjust for nonresponse bias. When the missingness structure is random, weighted adjustment is appropriate to compensate for the effect of statistical unit nonresponse. One of the weighting methods is the propensity score method. The weight assignment in the propensity score method is based on the estimated probability of response of the sample units. These estimates are obtained by fitting appropriate parametric models. In this article, the propensity score method and the resulting adjusted estimators are introduced. Then, the performance of the three adjusted propensity score estimators is compared. Finally, using the urban household income and expenditure survey dataset of the Statistical Center of Iran in 2011, the adjusted propensity score estimators are compared in terms of the square root of the mean square of the relative error and relative efficiency.