Investigation the effect of the item position on the person and item parameters

The aim of this study is to examine the effect of the same item in different positions among the booklets on the item difficulty parameter in different domains and item formats and to determine whether this effect differs across persons. In addition, it is to investigate the items that are more sensitive to the position effect and to examine some person and item characteristics that are thought to cause item position effects. The data set of the research is PISA 2015 Turkey sample. The analyzes of the study were carried out in the R program with the lme4 package, which is suitable for the analysis of Explanatory IRT models. According to the results, responding to the item in the later parts of the test increases the difficulty of the item, and this increase is more in the reading domain. In addition, item position effect differs among persons, and not every person is affected by the item position effect at the same level. When the relationship between ability and item position is examined, students with high ability levels are more affected by the item position effect in the domain of reading, and the opposite is true in the domain of mathematics. When the relationship between the item position and the item format is examined, in the reading domain, the item position effect between the open-ended and multiple-choice items is almost similar, but in the mathematics domain, open-ended items are more affected by the item position effect than the multiple-choice items. When the individual characteristics that may be related to the item position are examined, male students are more affected by the item position than female students in both domains. The most important variable in the relationship between persons' SES, anxiety and motivation levels and item position is SES, and students with high SES levels and average anxiety levels are less affected by the item position effect.