Evaluation of attitude items in PISA 2006 student questionnaire in terms of differential item functioning

Under the scope of this study, attitude items towards sciences were examined in PISA 2006 Student Survey is used to determine whether differential item functioning was observed in respect to gender and culture with ordinal logistic regression and poly- SIBTEST. Cultural differences and translational restrictions stand among the possible reasons of the observation of differential item functioning (DIF). Due to this reason, the research group of this study was determined as 5611 students from USA and 4942 students from Turkey whom participated in PISA aplication. In both methods, the number of items that showed DIF significantly and common is 7 in Turkey and 13 in the USA. In 30 items, with respect to culture, DIF was observed at a significant level through the analyses carried out in both countries utilizing both methods. When the above-mentioned two methods employed in the analyses, in terms of culture and gender, the data from USA and Turkey were compared, these two methods were found to be in compliance with each other, also with possible variations of their levels.