Validation of a scale of attitudes toward science across countries using rasch model: Findings from timss

Abstract

The purpose of the present study was to validate a scale of attitudes toward science based on items and data of TIMSS 2007. The present research intro-duces several implications and recommen-dations for science education researchers who are interested in validating and using instruments to measure attitudes toward science. This research examined, from Rasch measurement perspectives, to what extent the scale was valid across diverse participating countries from Rasch measurement perspectives. The data of the highest three achieving countries (Singa-pore, Taiwan, and Japan), the lowest three achieving countries (Qatar, Botswana, and Ghana), and three average achieving countries (Australia, Scotland, and Italy) were analyzed using WINSTEPS software. Several Rasch analyses (e.g, fit statistics, PCA, and person/item map reliability analysis) were performed to validate the scale. Although most of the item fit the Rasch model, a few items were found to be misfitting. Then, the misfitting items were eliminated and a second analysis for the fitting eight items was performed to check on the validity of the scale. The results pro-vided evidences that support the validity of the scale based on Rasch measurement.