Effects of Perceived Scoring Formula on Some Aspects of Test Performance
Abstract
The study examined the effects on test performance of systematic variations in the scoring formulas which examinees were told would be used in scoring their tests. Two equivalent 50-item vocabulary tests were constructed. Form A was administered without scoring formula specified to 420 pre-flight students. Three days later Form B was administered to the same groups subdivided six ways, i.e., same instructions as Form A; zero weight for wrongs; 1/4, 1, 2, or 4 points off for wrongs. Increases in the penalty for wrong responses were accompanied by consistent increases in the mean number of omitted items, but the mean number correct remained fairly stable over the various penalties. In general, interest correlations were largest when all items were attempted and smallest when random responses were substituted for omitted items. The scoring formula appropriate to the structure of the items, (R - w/4), was generally superior to the scoring formula appropriate to the penalty that examinees were told would be used in scoring the test.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 22, 1967
- Accession Number
- AD0662392
Entities
People
- Lawrence K. Waters
Organizations
- Naval Aerospace Medical Institute