The effects of selected student errors on the consistency of teacher grading in mathematics

• Main Author: Maxwell, Patricia Jean
• Format: Thesis
• Language: English
• Published: Memorial University of Newfoundland 1982
• Subjects: Newfoundland
• Online Access: https://research.library.mun.ca/4404/; https://research.library.mun.ca/4404/1/Maxwell_PatriciaJean.pdf; https://research.library.mun.ca/4404/3/Maxwell_PatriciaJean.pdf

The purpose of this study was to examine the consistency of teachers in the identification, classification, and grading of student solutions to grade 10 Algebra word problems. Also investigated were the effects on assigned grades, of error location, error type, and whether or not a correct numerical solution was obtained. -- The sample group was randomly selected from the population of all teachers of grade 10 Academic Mathematics in the Province of Newfoundland for the school year 1981-82. Of the 100 schools selected, 69 responded before the deadline amounting to a total of 90 teachers. -- The results indicated that teachers were very consistent in the location of student errors. In fact, the, errors were successfully located 99.4% of the time. They were slightly less consistent in the classification of these errors. Teachers used both general and specific classification terminologies to describe the student's errors and in a few instances simply referred to the error as reflecting careless or sloppy work. The grading of these errors showed much greater variability. Teacher's grades were most consistent when errors were made near the beginning of solutions and least consistent when errors were made near the end. The standard deviations of the scores of individual items ranged from a low of 1.07 to a high of 1.97. Grading practices were found to vary also in the level of severity individuals maintained. The amount the grades were distributed throughout the grading scale and the relative judgments of importance that were attributed to the various error types. -- It was also found that teachers graded property errors more severely than computational errors. Furthermore, errors made near the beginning of solutions appeared to be graded more severely than equivalent errors made near the end of a solution. No conclusive relationship was apparent between the grades teachers assign and whether a correct or incorrect numerical solution was obtained.