ERIC Number: ED393883
Record Type: RIE
Publication Date: 1995
Reference Count: N/A
Refinement in Assessment Validation: Technicalities of Dealing with Low Correlations and Instructor Grading Variation.
Rasor, Richard A.; Barr, James
Issues and problems in assessment research are explored, with suggestions to help establish an acceptable correlation between student assessment scores and final grades. Topics include assessing instructor grading variation, instructor grade point average (GPA), success rates, and lack of linearity in grade scales. Solutions to these problems are offered in the form of a new four-point research grading scale and a new "contextual" student GPA based on cumulative grade average with the final grade in the target course removed from the calculation. The contextual GPA was used to equate student "skill" levels in course selections when identifying the degree of instructor grading variation. The techniques were applied to a sample of 6,077 students covering performance data in 26 courses. The magnitude of the resulting correlations suggests giving much greater emphasis to student cumulative college GPA as a multiple measure when establishing entrance "skill" levels deemed necessary for success in general courses having no specific course prerequisite. Practical suggestions are included for identifying the true correlation between assessment test scores and grades given the problem of instructor grading variation. (Contains three tables and six figures.) (Author/SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Note: Portions of paper presented at the Annual Research Conference of the RP Group Granlibakken, 1993.