The alignment process is used to compare scores from two different tests with the same subjects. This study was conducted with the aim of finding the best method of data alignment using the logistic model. The data from the PhD exam in statistics in two consecutive years 2013 and 2014 were used. Only the questions from the specialized exams in statistics, which are 45 questions, were used for analysis. Question parameters and individuals' abilities were estimated based on a three-parameter model using MULTILOG software. In order to conduct this study, the mean-mean, mean-sigma, Haber and Stocking-Lord methods were implemented with the anchor test design, considering unequal groups, and the standard deviation of the errors was used to select the optimal method. The results of the analyses generally show that the methods based on the characteristic curve have more accurate results.