ESSAY TEST SCORING: INTERACTION OF RELEVANT VARIABLES

Abstract

In studies of essay tests, a single independent variable, such as penmanship, is often observed and conclusions are made about the relevance of that variable in scoring tests. But in this study it is hypothesized that the readers of an essay respond to a variable in terms of its context with other variables, that is, readers may be more tolerant of poor penmanship for males than females, especially if expectations for the writer are involved. In this investigation sex, race, reader expectation, and quality of handwriting were crossed to study their interaction effects. A single student essay was contrived. Each of 80 readers was given the essay plus a student's record card, which identified the student as a high or low achiever (expectancy), black or white, male or female. The essay was written in either good or poor handwriting. A reader read only one essay under one combination of conditions. The results showed complex interactions of expectations, writing, and sex within race.