NotesFAQContact Us
Search Tips
PDF pending restoration PDF pending restoration
ERIC Number: ED169126
Record Type: Non-Journal
Publication Date: 1976-Oct
Pages: 11
Abstractor: N/A
Reference Count: N/A
Out-of-Level Testing. ESEA Title I Evaluation and Reporting System. Technical Paper No. 6.
Roberts, A. Oscar H.
For Title I evaluations, it may be appropriate to test out-of-level; that is, to override publisher's recommendations concerning the difficulty, length, and content appropriate for a particular grade. It is seldom necessary, however, to move more than one grade down. If the mean is substantially higher than the median, then some pupils will have encountered the floor of a test and an easier level of tests should have been chosen. The ceiling of most tests becomes a handicap when three-quarters of a group can answer the most difficult items correctly. In this case, the mean is substantially lower than the median and a more difficult test should have been chosen. In general, the level of a test is suitable when the raw score of the group is equal to or above a third of the maximum score, and somewhat less than three-quarters of the maximum. In norm-referenced evaluations out-of-level testing is possible with most standardized achievement tests because they provide tables for relating raw scores on out-of-level tests to in-level percentile norms. (CP)
Publication Type: Guides - Non-Classroom; Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: Office of Education (DHEW), Washington, DC.
Authoring Institution: RMC Research Corp., Mountain View, CA.