NotesFAQContact Us
Search Tips
ERIC Number: ED120290
Record Type: RIE
Publication Date: 1975-Dec
Pages: 25
Abstractor: N/A
Reference Count: 0
Constructing Unbiased Achievement Tests.
Ebel, Robert L.
The approach characterizing this study of bias in tests of educational achievement is stated to be reflective and philosophical rather than experimental and statistical. Its concern is with operationally definable areas of learning and indicators of achievement rather than with the hypothetical latent traits that are presumed by some to influence or determine test performance. A brief discussion of emotionality in test bias, that also considers the seriousness of the problem, is followed by comments on bias in selection testing. The meaning of bias in achievement tests is discussed next together with other topics such as: bias in the test itself, the problem of reading difficulty and linguistic difference as a possible source of bias, non-standard English speakers, language specificity in achievement, test bias and test validity, constructing unbiased achievement tests, operational definition versus construct validation, appearance versus reality in test validation, the correctness of operational definitions, and a discussion of two propositions advanced to support the suggestion that criterion referenced tests minimize bias. Issues discussed throughout the paper are noted to suggest that bias in tests of educational achievement is quite improbable. The a priori assumption that an achievement test is essentially unbiased is considered more reasonable than the assumption that it is biased. (Author/AM)
Publication Type: Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: N/A
Sponsor: N/A
Authoring Institution: N/A
Note: Paper presented at the National Institute of Education Test Bias Conference (Annapolis, Maryland, December 2-4, 1975)