While the examining of the relations between the categorical variables, the observation values are collected on a cross classified table. In the situation that there is two categorical variables, then a two dimensional table is prepared as if one variable is in the row, and the other is in the column. With respect to this table, in order to examine the relations between the variables, usually applied method is chi-square analysis. When the variables is more than two, the analysis of variables'

more »

... ross classified tables as one within the other can not be done with the chi-square analysis. In this case, for the examining of the relation structures between the variables, log-linear analysis can be made use of. Log-linear analysis is a method that can analyze with the help of the models, both IX2 or ZX] dimensional tables to which chi-square can be applied, and three or more dimensional tables to which chi-square is insufficient. In this study, the results of the Entrance Examination to the Postgraduate Education (LES) carried out by The Student Selection and Placement Center (OSYM) are examined with the log-linear analysis by the classification on the three dimensional crosswise table. In the scope of the analysis, the factors, that is examined with their relation structures, are the variables of education level (bachelor's degree, master, doctor's degree), LES score (45 point and more or less 45 point) and the type ofpoint (qualitative, quantitative and equal weight). At the end of the analysis, the relation structure between the in question variables is revealed and the findings are interpreted. Chi-Square Independence Test, Log-Linear Models, Odds Ratio, LES Scores.