A Contribution to the Evidence of the Scale Validity of the PISA Test. (Technical Report IEIA-INEE)

The international OECD PISA [Programme for International Assessment] 2006 test focused on the performance of Sciences of 15 years old students. The unsatisfactory results from Mexico were submitted to analysis, including multilevel models, to explain the origin of their deficiencies. It was clear that a differential functioning behavior or a design skewed against a specific country would produce serious validity problems. Besides the conceptual design of the test blueprint and the intrinsic quality of the items, an evidence of the quality of the test is provided by the Test Design Line, grounded on the distribution of the item difficulties calibrated with the Rasch model. The results of this approach show that the scale validity of the test is appropriate for Mexican students, an information that has to be known by practitioners and educational authorities. (Contains 3 figures.)