ERIC Number: ED259028
Record Type: Non-Journal
Publication Date: 1985-Apr
Large-Sample Pre-equating: How Accurate?
Livingston, Samuel A.
Four tests were pre-equated on the basis of item-test regressions computed from a large sample providing pretest data. The tests were two mathematics tests (math computation and elementary algebra) and two verbal tests (sentence sense and reading comprehension) from the New Jersey College Basic Skills Placement Test. Regular equipercentile equating results showed that pre-equating was highly accurate in three of the four tests. The fourth test (reading comprehension) showed a small but systematic inaccuracy in predicting the equating relationships--the difficulty level was underestimated, especially for students in the lower middle ability range. An investigation of this inaccuracy suggested that time limits may have affected the students' item responses. (Author/GDC)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers - Assessments and Surveys: New Jersey College Basic Skills Placement Test