NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED510539
Record Type: Non-Journal
Publication Date: 2005
Pages: 4
Abstractor: ERIC
Reference Count: N/A
ISBN: N/A
ISSN: N/A
Gender Fairness Using the ACT
ACT, Inc.
A criticism made against standardized tests is that they may be biased against females because males typically outscore females. On the ACT, males perform only slightly better than females. The population of students who take the ACT is self-selected--that is, the ACT has traditionally been taken primarily by those planning to attend college. For the past decade, 56 percent of ACT test takers have been female, a figure identical to the percentage of females attending college during the same period. One theory is that if all students, not just the college bound, were to take the ACT, there would be no score difference between males and females. To test this theory, eight years' worth of data from the states of Colorado and Illinois, both of which recently moved from traditional self-selected administrations of the ACT to a system in which all students take the test, were studied. There is clear evidence that gender differences on the ACT are a function of self-selection and that these differences disappear (or reverse slightly in favor of females) when all students are tested. One can conclude, therefore, that the differences seen between males and females on the ACT are a function of self-selection--that is, who decides to take the test--rather than inherent bias in what the test is measuring. (Contains 4 figures and 2 tables.)
ACT, Inc. 500 ACT Drive, P.O. Box 168, Iowa City, IA 52243-0168. Tel: 319-337-1270; Web site: http://www.act.org
Publication Type: Reports - Research
Education Level: Higher Education
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: ACT, Inc.
Identifiers - Location: Colorado; Illinois
Identifiers - Assessments and Surveys: School and College Ability Tests