Open AccessAN INVESTIGATION OF ITEM BIAS 1
Author(s) -
Cleary T. Anne,
Hilton Thomas L.
Publication year - 1966
Publication title -
ets research bulletin series
Language(s) - English
Resource type - Journals
eISSN - 2333-8504
pISSN - 0424-6144
DOI - 10.1002/j.2333-8504.1966.tb00355.x
Subject(s) - variance (accounting) , statistics , mathematics , analysis of variance , section (typography) , psychology , sample (material) , socioeconomic status , test (biology) , demography , computer science , biology , population , paleontology , chemistry , accounting , chromatography , sociology , business , operating system
For this research, bias was defined as an item x group interaction. Grade 12 students in integrated high schools who had taken two forms of the PSAT were divided into two races and three socioeconomic levels within the races. Four analyses of variance were performed: one for the Verbal section and one for the Mathematical section of each of the two forms of the PSAT. Because of the large sample sizes used, the tested effects were expected to be and were significant. However, only a minimal percentage of the total variance was contributed by the item x group interactions. On this basis, it was concluded that, if PSAT scores are discriminatory, the discrimination is not largely attributable to particular items, but to the test as a whole.